RESEARCH
HIGHLIGHTS HIGHLIGHT ADVISORS CEZMI AKDIS SWISS INSTITUTE OF ALLERGY AND ASTHMA RESEARCH, SWITZERLAND BRUCE BEUTLER SCRIPPS RESEARCH INSTITUTE, USA PETER CRESSWELL YALE UNIVERSITY, USA JAMES DI SANTO PASTEUR INSTITUTE, FRANCE GARY KORETZKY UNIVERSITY OF PENNSYLVANIA, USA CHARLES MACKAY GARVAN INSTITUTE OF MEDICAL RESEARCH, AUSTRALIA CORNELIS J. M. MELIEF LEIDEN UNIVERSITY MEDICAL CENTRE, THE NETHERLANDS MICHEL NUSSENZWEIG THE ROCKEFELLER UNIVERSITY, USA ALAN SHER NATIONAL INSTITUTE OF ALLERGY AND INFECTIOUS DISEASE, USA ANDREAS STRASSER THE WALTER AND ELIZA HALL INSTITUTE, AUSTRALIA MEGAN SYKES HARVARD MEDICAL SCHOOL, USA ERIC VIVIER CENTRE D’IMMUNOLOGIE DE MARSEILLE-LUMINY, FRANCE MATTHIAS VON HERRATH LA JOLLA INSTITUTE FOR ALLERGY AND IMMUNOLOGY, USA.
834
H A E M ATO P O I E S I S
Keeping HSCs under control Although in certain situations haematopoietic stem cells (HSCs) can undergo extensive proliferation, they remain largely quiescent in normal adults. And whereas many nuclear factors that promote the proliferation of HSCs have been characterized, the factors that hold HSC proliferation in check have been difficult to identify. But now, two new reports indicate that the transcriptional repressor GFI1 (growth-factor independent 1) restricts HSC proliferation. GFI1 has previously been shown to promote T-cell proliferation. So, because GFI1 is expressed by HSCs, both groups set out to investigate whether it also promotes proliferation of these cells. Surprisingly, Hock et al. observed that, compared with wildtype litter-mate controls, there were at least as many, if not more, phenotypic HSCs in the bone marrow of GFI1deficient mice. By contrast, Zeng et al. observed that the number of phenotypic HSCs was decreased in the bone marrow of GFI1-deficient mice. However, both groups showed that HSC function was compromised; when lethally irradiated mice were transplanted with a mixture of GFI1deficient bone marrow and wild-type bone marrow, GFI1-deficient HSCs were outcompeted as the mice aged, and haematopoietic cells derived from the GFI1-deficient bone marrow were undetectable. Together with the observation that bone marrow from recipients of GFI1-deficient bone marrow was unable to reconstitute a secondary lethally irradiated recipient
in a serial-transplantation assay, these results establish that GFI1 is crucial for HSC self-renewal and engraftment function. Further analysis indicated that a larger proportion of GFI1-deficient bone-marrow HSCs were in the proliferative stages of the cell cycle, indicating that GFI1 functions to restrain HSC proliferation. Consistent with this proposed role for GFI1 in controlling proliferation and cell-cycle progression of HSCs, the level of mRNA encoding the G1 checkpoint regulator p21 (also known as CIP1 or WAF1), which is required to maintain HSCs in G0, was found to be decreased in GFI1-deficient HSCs. These studies identify GFI1 as a negative regulator of HSC prolifera-
| NOVEMBER 2004 | VOLUME 4
tion, and both groups suggest it is probable that the excessive proliferation of GFI1-deficient HSCs results in exhaustion and the observed loss of self-renewal function. This role as a factor restricting HSC proliferation is in contrast to the observation that GFI1 promotes T-cell proliferation, highlighting that transcription-factor function is context dependent and can be celltype specific. Karen Honey References and links ORIGINAL RESEARCH PAPERS Hock, H. et al. Gfi-1 restricts proliferation and preserves functional integrity of haematopoeitc stem cells. Nature 431, 1002–1007 (2004) | Zeng, H., Yücel, R., Kosan, C., Klein-Hitpass, L. & Möröy, T. Transcription factor Gfi-1 regulates self-renewal and engraftment of haematopoietic stem cells. EMBO J. 23 September 2004 (doi:10.1038/sj.emboj.7600419).
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
RESEARCH HIGHLIGHTS
IN BRIEF
DENDRITIC CELLS
Not the end of the road What happens to a dendritic cell (DC) when it has completed its journey from the peripheral tissues to the secondary lymphoid organs, where it presents antigen to naive T cells? Previous in vitro studies have shown that these mature DCs succumb to activationinduced apoptosis. But new work published in Nature Immunology indicates that the path of DC involvement might be longer than previously thought: mature DCs can differentiate further to a regulatory phenotype under the influence of the lymphoid microenvironment. Zhang et al. cultured mature DCs on monolayers of endothelial-like splenic stromal cells (ESSCs) to mimic the secondary lymphoid microenvironment. These DCs continued to proliferate for several months, whereas mature DCs cultured alone die after ∼1 week. The ESSCs also induced a distinct DC phenotype involving downregulation of expression of MHC class II molecules and increased expression of some co-stimulatory molecules, such as CD40 and CD80. These DCs produced more interleukin-10 (IL-10) and nitric oxide but less transforming growth factor-β (TGF-β) and IL-12 than mature DCs. The induction of mature DC proliferation was shown to require direct cell contact with ESSCs, because fixed ESSCs, but not ESSC culture supernatants, could reproduce the effect. The addition of blocking antibodies specific for fibronectin inhibited ESSC-induced proliferation of mature DCs, indicating one pathway by which this cell contact might be mediated. By contrast, the further differentiation of mature DCs was shown to require both cell contact and ESSC-derived soluble factors. A blocking antibody specific for TGF-β inhibited the further differentiation of mature DCs induced by ESSC supernatant, indicating that this is a crucial soluble factor. Ovalbumin (OVA)-pulsed ESSCdifferentiated DCs did not promote the proliferation of OVA-specific
CD4+ T cells and could inhibit T-cell proliferation induced by OVA-pulsed mature DCs. However, the ESSCdifferentiated DCs did induce T-cell activation, as indicated by cell-surface marker expression and increased production of IL-2 and interferon-γ. The inhibitory function of these DCs on T-cell proliferation was shown to depend on the production of soluble factors and not on cell–cell contact, and a selective inhibitor of nitricoxide synthase reversed the inhibition of T-cell proliferation. Also, T cells activated by ESSC-differentiated DCs did not function as regulatory cells when added to mature DC and T-cell co-cultures. Taken together, these results indicate that the increased production of nitric oxide by ESSCdifferentiated DCs directly inhibits T-cell proliferation, without the generation of regulatory T cells that has been shown for other DC subtypes. Finally, the authors looked for an in vivo counterpart to this new DC phenotype observed in vitro. The main distinguishing characteristic in vitro was decreased expression of MHC class II molecules, so they sorted splenic CD11c+ DCs into negative, low and high populations on the basis of MHC class II expression. The DCs with low-level expression of MHC class II had a similar cell-surface phenotype to ESSC-differentiated DCs and also produced high levels of IL-10 and nitric oxide. Their function was also similar in that they inhibited the proliferation of T cells induced by DCs that expressed high levels of MHC class II. So, the authors suggest that, rather than being the end of the road for mature DCs, the lymphoid microenvironment can induce a regulatory DC phenotype that might have a role in maintaining immune homeostasis by exerting negative feedback at late stages of a response. Kirsty Minton References and links ORIGINAL RESEARCH PAPER Zhang, M. et al.
Splenic stroma drives mature dendritic cells to differentiate into regulatory dendritic cells. Nature Immunol. 5, 1124–1133 (2004).
N AT U R A L K I L L E R T C E L L S
A subset of liver NK T cells is activated during Leishmania donovani infection by CD1d-bound lipophosphoglycan. Amprey, J. L. et al. J. Exp. Med. 200, 895–904 (2004).
This is the first paper to show that presentation of a pathogenderived glycolipid antigen by CD1d to a subset of natural killer T (NKT) cells has a key role in resistance to Leishmania donovani infection of mice. CD1d-deficient mice have increased susceptibility to various microbial pathogens, and the authors show that they also have increased susceptibility to L. donovani. In wild-type mice, L. donovani infection induces rapid production of interferon-γ (IFN-γ) by CD1d-restricted NKT cells in the liver. In vitro binding assays showed that L. donovani lipophosphoglycan (LPG) and other related glycolipids bind strongly to CD1d, and LPG-treated dendritic cells could stimulate IFN-γ production by hepatic lymphocytes, indicating that glycolipid-specific responses are likely to contribute to resistance to leishmaniasis. T CELLS
Unique gene expression program of human germinal center T helper cells. Kim, C. H. et al. Blood 104, 1952–1960 (2004).
Germinal-centre T helper (GC-TH) cells are specifically localized in the germinal centres of lymphoid follicles to help B cells produce antibodies. They are characterized by expression of CXCR5 and CD57, and are non-polarized in terms of cytokine production. Now, Eugene Butcher and colleagues have used cDNA-microarray analysis to find out more about the differences between GC-TH cells and other T-cell subsets. In particular, GC-TH cells are unique in their expression of the chemokine CXCL13, which is crucial for B-cell entry to lymphoid follicles but was previously thought to be produced by non-T cells, such as follicular dendritic cells. They also found marked differences in the expression of transcription factors between GC-TH cells and other T-cell subsets, which supports the theory that GC-TH cells are a novel functional lineage. ASTH MA AN D ALLE RGY
Defining a link with asthma in mice congenitally deficient in eosinophils. Lee, J. J. et al. Science 305, 1773–1776 (2004).
A critical role for eosinophils in allergic airway remodeling. Humbles, A. A. et al. Science 305, 1776–1779 (2004).
Two studies in Science have used different transgenic mice that lack eosinophils to look at how these cells are involved in the pathogenesis of asthma. Both studies reveal an essential role for eosinophils, but they differ in terms of the stage that is affected. Lee et al. show that eosinophils are required for mucus accumulation and airway hyper-responsiveness. By contrast, Humbles et al. show that eosinophils are not required for this acute lung dysfunction but are necessary for the collagen deposition and smooth-muscle hyperplasia that lead to chronic airway remodelling. Regardless of these differences, both studies indicate that eosinophils would be a good therapeutic target for human asthma.
NATURE REVIEWS | IMMUNOLOGY
VOLUME 4 | NOVEMBER 2004 | 8 3 5 ©2004 Nature Publishing Group
RESEARCH HIGHLIGHTS
IN THE NEWS Flu vaccine shortages “The flu is not a severe cold: it can be a serious illness, and three to four thousand deaths [in the United Kingdom] are linked to flu every year”, according to Chief Medical Officer Sir Liam Donaldson (Department of Health). “If you suffer from a chronic illness like asthma or diabetes, or are 65 years or older, you are particularly at risk from flu”, he states. A similar warning is given in the United States, although vaccination is recommended for a wider group of people, including pregnant women and individuals of 50 years or older (New York State Office for Aging). However, the Medicines and Healthcare Products Regulatory Agency (UK) has recently suspended the manufacturing licence of the influenza-vaccine maker Chiron Corporation because of concerns about the way the vaccine is manufactured. This decision is likely to have a particularly severe impact in the United States, where Chiron is responsible for nearly 50% of the influenza vaccines distributed, and Dr Anthony Fauci (Director of the National Institute of Allergy and Infectious Disease) has said that there will be “significant shortages” (The New York Times). In the United Kingdom, Chiron supplies a smaller proportion of the influenza vaccines, and a Department of Health spokesperson said, “We are confident that we have sufficient vaccines for this winter’s campaign” (BBC News). As a result of the vaccine shortage, the Centers for Disease Control and Prevention (USA) is planning to teach people how to protect themselves through hygiene and ‘cough etiquette’. It has said that you should avoid touching your eyes, nose or mouth and that if you do get flu, stay at home so that you don’t infect others (The New York Times).
M U C O S A L I M M U N O LO G Y
IL-15 triggers destruction The destruction of the intestinal epithelium by interleukin-15 (IL-15)-activated intraepithelial cytotoxic T lymphocytes (CTLs) is one of the main pathological events of coeliac disease — a T-cellmediated disease of the small intestine induced by wheat gliadin. Two papers that were recently published in Immunity now provide insight into the mechanisms of this cytotoxicity: the triggering of NKG2D (naturalkiller (NK) group 2, member D) on the cell-surface of intraepithelial CTLs by its ligand MICA (MHC-class-Ipolypeptide-related sequence A) or MICB present on the cell-surface of intestinal epithelial cells (IECs) initiates signalling pathways that lead to the killing of IECs, and this is coordinated by IL-15. CTLs have been shown to express NK-cell receptors such as NKG2D, and signals transduced by these receptors can modulate the cytotoxicity of CTLs. So, Meresse et al. set out to investigate whether NKG2D expression by intraepithelial CTLs was important in coeliac disease. Intraepithelial CTLs isolated from biopsies of patients with active coeliac disease expressed markedly higher levels of NKG2D than the same cells from healthy individuals. In contrast to peripheralblood-derived CTL clones (which cannot mediate targetcell lysis after stimulation through NKG2D alone), ligation of NKG2D alone was sufficient to induce lysis of target cells by intraepithelial CTLs from patients with active coeliac disease. Further analysis indicated that intraepithelial CTLs from healthy individuals could acquire this phenotype — high levels of NKG2D expression and the ability to lyse target cells after ligation of NKG2D alone — when cultured with IL-15. However, both exposure to IL-15 and recent stimulation through the T-cell receptor was required to enable peripheralblood CD8+ memory T cells to elicit NKG2D-mediated lysis of target cells. These observations indicate that the high levels of IL-15 and the constant exposure to antigen in the intestine of patients with active coeliac disease are probably responsible for the NKG2D-mediated cytolytic phenotype of the intraepithelial CTLs isolated from these individuals. IL-15 was shown to increase not only the expression of NKG2D but also the expression of signal-transduction adaptor molecule DAP10, further priming the cells to be responsive to NKG2D signalling. In the intestine of patients with active coeliac disease such signals could be provided by the NKG2D ligands MICA and/or MICB, the expression of which were found to be upregulated on the cell-surface of IECs from such individuals. Similarly, in the second study, Hüe et al. observed that, compared with cells from healthy individuals, the level of expression of MICA was markedly increased on the cell-surface of IECs from patients with active coeliac disease. Furthermore, the level of MICA expression correlated with the severity of disease. In cultures of biopsies taken from coeliac patients on gluten-free
diets (who are therefore free of active disease), wheat gliadin was shown to induce the expression of MICA by IECs. The gliadin peptide p31–49 (known to cause damage of the small intestine by itself) could also induce MICA expression by IECs, and this was abolished in the presence of IL-15-neutralizing antibody. In contrast to the report of Meresse et al., Hüe et al. observed that intraepithelial CTLs from both healthy individuals and those with active coeliac disease expressed similar levels of NKG2D. Furthermore, when stimulated through NKG2D alone, intraepithelial CTL cell lines derived from patients with active coeliac disease could induce lysis of target cells only in certain situations — for example, if cells were recently exposed to high concentrations of IL-15. However, in all cases, NKG2D ligation markedly amplified CD3-mediated lysis of target cells, indicating a role for NKG2D signalling in the regulation of intraepithelial CTL cytotoxicity. The results obtained from these studies indicate that NKG2D–MICA/MICB interactions are crucial for mediating destruction of IECs in active coeliac disease and that IL-15 regulates the frequency of this interaction. Further studies are required to confirm the precise relationship between IL-15, NKG2D and MICA/MICB in coeliac disease, but both groups suggest that targeting this axis could provide a novel therapy for coeliac disease. Karen Honey References and links ORIGINAL RESEARCH PAPERS Meresse, B. et al. Coordinated induction
by IL-15 of a TCR-independent NKG2D signaling pathway converts CTL into lymphokine-activated killer cells in celiac disease. Immunity 21, 357–366 (2004) | Hüe, S. et al. A direct role for NKG2D/MICA interaction in villous atrophy during celiac disease. Immunity 21, 367–377 (2004).
Karen Honey
836
| NOVEMBER 2004 | VOLUME 4
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
HIV
Getting to the bottom of CD4+ T-cell loss Previous studies of HIV pathogenesis have largely ignored events in the intestines of HIV-infected patients and have mostly concentrated on events in the blood, perhaps owing to the difficulty in obtaining intestinal lymphoid-tissue samples. However, two groups now report that the gastrointestinal tract has the most marked depletion of CD4+ T cells, and this occurs rapidly and at all stages of HIV infection, regardless of administration of highly active antiretroviral therapy (HAART). Given that the gastrointestinal tract and other lymphoid tissues harbour most of the body’s CD4+ T cells and that, in the gut, a large number of these cells are activated and express the HIV co-receptor CC-chemokine receptor 5 (CCR5), the intestinal CD4+ T cells are potentially highly susceptible to infection with HIV. So, the authors of both studies set out to examine CD4+ T-cell depletion in the intestine of patients with HIV. Danny Douek’s group compared T-cell depletion in intestinal tissue, lymph nodes and blood from untreated HIV-infected patients (at all stages of disease) and from HIVuninfected individuals. As expected, the frequency of CD4+ T cells was significantly lower in each compartment in HIV-infected patients compared with HIV-uninfected individuals; however, the greatest loss of CD4+ T cells was seen in the gastrointestinal tract, even in acute infection. This preferential depletion of intestinal CD4+ T cells was confirmed by histological studies, showing that lymphoid aggregates, which are abundant in normal intestinal-tissue samples, were largely absent in samples from patients with HIV. Further analysis revealed that the CD4+ T-cell depletion in the gut was specific for those cells expressing CCR5 and the activation marker Ki67, which is consistent with the observation that HIV
preferentially replicates in and causes the death of activated CD4+ T cells. It is well known that chronic infection with HIV results in a state of general immune activation. Accordingly, Douek’s group observed that effector memory T cells accumulated abnormally in the lymph nodes of HIVinfected patients. Moreover, this increased T-cell activation was associated with collagen deposition in the lymph nodes, a marker of inflammation associated with chronic immune activation. This led the authors to suggest that disrupted lymphoid architecture might disturb normal lymphoid-tissue homeostasis, which might then impair CD4+ T-cell reconstitution of the gut following depletion by infection with HIV. Martin Markowitz’s group studied T-cell depletion in patients with acute or early HIV infection and compared untreated patients with those who had received HAART. Consistent with observations by Douek’s group, in primary HIV infection, they found that CD4+ T-cell depletion was most marked in the gastrointestinal tract and that it occurred before changes observed in the blood. Specifically, deletion mainly occurred in the effector sites (the lamina propria) as opposed to the inductive sites (Peyer’s patches and lymphoid follicles) of the gut mucosa. To examine whether HAART would allow the reconstitution of CD4+ T cells in the gut, they studied patients who had initiated HAART during primary infection. However, although CD4+ T-cell numbers in the blood of these patients were mostly restored following HAART, reconstitution in the gastrointestinal tract remained incomplete despite up to 5 years of fully suppressive therapy. Both of these studies highlight the need for further investigation of the mucosal compartment, given its crucial role in HIV infection, replication and persistence. Lucy Bird References and links ORIGINAL RESEARCH PAPERS Brenchley, J. M.
et al. CD4+ T cell depletion during all stages of HIV disease occurs predominantly in the gastrointestinal tract. J. Exp. Med. 200, 749–759 (2004) | Mehandru, S. et al. Primary HIV-1 infection is associated with preferential depletion of CD4+ T lymphocytes from effector sites in the gastrointestinal tract. J. Exp. Med. 200, 761–770 (2004).
C Y TO K I N E S
Cytokine receptors take sugar Many proteins are modified in the endoplasmic reticulum through glycosylation of asparagine residues in a process known as N-glycosylation. N-glycans are further modified in the Golgi by enzymes, such as mannosidases and glycosyltransferases, to generate complex-type N-glycans. New research published in Science now shows that mannoside acetylglucosaminyltransferase 5 (MGAT5) modifies cytokine-receptor N-glycans and that this regulates cytokine-mediated signalling. MGAT5-mediated N-glycan modifications generate ligands for the galectin (GAL) family of lectins. It has previously been shown that GAL3 — which binds N-glycans modified by MGAT5 — can oligomerize and form a molecular lattice of galectins and glycoproteins at the cell surface and that GAL3 binding of MGAT5-modified N-linked glycans on the T-cell receptor opposes antigen-induced clustering. Partridge et al. sought to investigate whether cytokine-receptor function was modulated by MGAT5 modification of N-linked glycans. MGAT5-deficient epithelial tumour cell lines derived from mammary tissue were less responsive to several cytokines, including epidermal growth factor (EGF) and transforming growth factor-β (TGF-β), than MGAT5-sufficient cells, and the number of N-glycans per cytokine receptor correlated with the decrease in sensitivity. The loss in EGF sensitivity was associated with a decrease in the level of cell-surface expression of the EGF receptor (EGFR), together with a decrease in the association of the EGFR with GAL3, a decrease in EGF-dependent activation of ERK (extracellular signal-regulated kinase) and an accumulation of the EGFR in the early endosomes. Similar data were obtained for the TGF-β receptor II. Taken together, these data are consistent with the hypothesis that MGAT5-mediated modification of N-linked glycans on these cytokine receptors allows the formation of molecular lattices that oppose receptor endocytosis. The binding of TGF-β to receptors on the surface of lipopolysaccharide-elicited peritoneal macrophages was decreased in the absence of MGAT5, as was TGF-β-dependent activation of SMAD2 and/or SMAD3. Furthermore, a functional role for MGAT5-mediated modification of N-glycans on cell-surface receptors was indicated by the observations that MGAT5-deficient macrophages were impaired in their ability to phagocytose latex beads and that leukocyte recruitment was delayed in both a model of skin inflammation and after intraperitoneal injection of thioglycollate. This study shows that MGAT5-mediated modification of N-glycans on cytokine receptors regulates their cell-surface expression levels by balancing expression on the cell surface with loss through endocytosis and that this is an important mechanism controlling cellular cytokine responsiveness. Karen Honey References and links ORIGINAL RESEARCH PAPER Partridge, E. A. et al. Regulation of cytokine receptors
by Golgi N-glycan processing and endocytosis. Science 306, 120–124 (2004).
NATURE REVIEWS | IMMUNOLOGY
VOLUME 4 | NOVEMBER 2004 | 8 3 7 ©2004 Nature Publishing Group
RESEARCH HIGHLIGHTS
A U TO I M M U N I T Y
Gene therapy for diabetes
Type 1 diabetes mellitus (T1DM) occurs as a result of T-cell-mediated destruction of the insulin-producing β-cells in the pancreas. Non-obese diabetic (NOD) mice, which spontaneously develop diabetes, are a useful model system of this disease. In this study, the authors show that expression of diabetes-resistant MHC class II alleles by NOD mice, using retroviral transduction of autologous bone-marrow cells, is sufficient to prevent the development of diabetes. In humans, the development of T1DM is associated with the inheritance of particular MHC class II alleles that lack a charged amino acid at position 57 of the β-chain. In NOD mice, the single MHC class II allele present is I–Ag7, which also lacks a charged residue at position 57. It is thought that lack of a charged residue at this position prevents the formation of a salt bridge between the α- and β-chains of the MHC class II molecule, which could affect the ability of these molecules to mediate negative selection of autoreactive T cells. Previous studies using transgenic mice and allogeneic bone-marrow chimeras have shown that it is possible to prevent diabetes in NOD mice, but it has not been possible to determine the preventative mechanism.
T- C E L L M E M O R Y
Memories are made of this… Infection with the parasite Leishmania causes considerable morbidity and mortality, against which there is no effective human vaccine. Both humans and mice can resolve primary infections and become resistant to further infection, but some parasites persist and might contribute to long-term protection by maintaining the presence of effector T cells. If persistent antigen is required for long-term protection, then developing a non-live vaccine against leishmaniasis will be difficult. In this study, Phillip Scott and colleagues characterized the CD4+ T-cell response during infection and showed that protective central memory T (TCM) cells, which are not dependent on the presence of parasites, can develop in infected mice. Memory T cells are a heterogeneous population thought to contain two distinct subsets — effector memory T (TEFF) cells, which migrate to tissues and produce cytokines, and TCM cells, which circulate through the lymph nodes. Scott and colleagues investigated the development of CD4+ memory T cells during infection of mice with Leishmania major. CD4+ T cells were labelled
838
with CFSE (5,6-carboxyfluorescein diacetate succinimidyl ester) — which allows proliferative responses to be analysed by flow cytometry — and then transferred into naive recipients. After parasitic challenge infection of the recipient mice, some of these donor T cells migrated to the draining lymph nodes (dLNs) and proliferated, indicating that immune mice contain a TCM-cell population. During proliferation, most cells downregulated their expression of the lymphnode homing molecule CD62L, indicating that these cells had differentiated into TEFF cells. To confirm that TCM cells present in the donor T-cell population from immune mice could differentiate into TEFF cells and mediate protection, CD4+CD62Lhi T cells — which do not produce interferon-γ (IFN-γ) when stimulated with leishmanial antigens — were purified and used in transfer experiments. Again, the donor T cells proliferated in the dLN; they also developed the capacity to produce IFN-γ and could be detected at the site of infection within two weeks. Importantly, the mice that received these TCM cells were protected from challenge infection.
| NOVEMBER 2004 | VOLUME 4
Retroviral constructs with green fluorescent protein fused to the cytoplasmic tails of diabetesresistant I–A β-chains (which have a charged residue at position 57) were shown to be expressed at the cell surface of MHC-class-IIpositive cells and could pair with endogenous α-chains.Young NOD mice were lethally irradiated and transplanted with bone marrow retrovirally transduced with the I–Aβ constructs, and their blood-glucose levels were monitored each week. All of the mice were protected from diabetes, compared with four of six mice that received bone marrow transfected with a control construct. Even when diabetes was aggressively induced using cyclophosphamide, the mice that received I–Aβ-transduced bone marrow were resistant to the development of diabetes for up to 46 weeks after transfer, and the level of insulitis was reduced. To address the mechanism of resistance, the authors used enzyme-linked immunosorbent spot (ELISPOT) assays to look at the frequency of T cells that produced cytokines in response to peptide 206–220 from glutamic-acid decarboxylase, an immunodominant self-antigen peptide in NOD mice. No cytokine production was detectable in NOD mice that received
To address the role of persistent antigen in the maintenance of memory T cells, a mutant L. major strain that is unable to persist in mice was used. Mice infected with this mutant L. major had no detectable parasites by 15 weeks after infection. At 25 weeks after infection, no TEFF cells were detectable, as shown by the lack of both an IFN-γ response to leishmanial antigens in vitro and a delayed-type hypersensitivity response in vivo. However, TCM cells were detectable at 25 weeks; CFSE-labelled CD4+ T cells from these mice, when transferred to naive recipients that were subsequently challenged with L. major, could migrate to and proliferate in the dLN. Furthermore, 25 weeks after the initial infection with mutant L. major, mice were protected against infection with virulent L. major, showing that the TCM cells conferred protection. This study shows that both TEFF and TCM cells contribute to immunity to L. major infection, but TCM cells can be maintained in the absence of parasites and confer protection. Targeting TCM cells could therefore be the basis for the development of a successful non-live vaccine against leishmaniasis. Elaine Bell References and links ORIGINAL RESEARCH PAPER Zaph, C., Uzonna, J.,
Beverley, S. M. & Scott, P. Central memory T cells mediate long-term immunity to Leishmania major in the absence of persistent parasites. Nature Med. 10, 1104–1110 (2004).
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
I-Aβ-transduced bone marrow, indicating that self-reactive T cells had been functionally inactivated or eliminated. To determine whether this was owing to deletion of self-reactive T cells in the thymus or to peripheral tolerance, the authors used I–Ag7 tetramers loaded with a peptide known to stimulate pancreatic-isletreactive T-cell clones. The frequency of CD4+ T cells labelled with the tetramer was significantly reduced among thymocytes from NOD mice reconstituted with I–Aβ-transduced cells compared with NOD mice that received cells transduced with the control construct, supporting the idea that I–Aβ mediated the removal of self-reactive T cells by negative selection. This study offers the prospect that T1DM could be prevented by providing susceptible individuals with protective MHC class II alleles, using autologous bone marrow. This approach would be preferable to the use of allogeneic cells, because graft-versus-host disease would be avoided and it might be possible to use milder conditioning regimens before transplantation. Elaine Bell References and links ORIGINAL RESEARCH PAPER Tian, C. et al. Prevention of type 1 diabetes by gene therapy. J. Clin. Invest. 114, 969–978 (2004).
B - C E L L D E V E LO P M E N T
BCL-6 recruits new team member Delegation is the key to success, and just as every boss needs a team to ensure that the work gets done, this paper in Cell describes an important member of the team of cofactors that control B-cell differentiation in association with the master regulator BCL-6 (B-cell lymphoma 6). The transcriptional repressor BCL-6 is expressed by germinalcentre B cells and functions to antagonize plasma-cell differentiation controlled by BLIMP1 (B-lymphocyte-induced maturation protein 1). Now, this study shows that MTA3, a cell-type-specific subunit of the Mi-2/NuRD co-repressor complex, associates with BCL-6 to inhibit the terminal differentiation of B cells to plasma cells. Immunohistochemical analysis of human lymph-node and tonsil sections showed that MTA3 is highly expressed by a population of germinal-centre B cells that also express BCL-6, and the co-expression of MTA3 and BCL-6 was observed in B-cell lines but not in plasma-cell lines. Furthermore, MTA3 and BCL-6 could be co-precipitated from a B-cell line together with other subunits of Mi-2/NuRD, indicating that BCL-6 stably interacts with an MTA3-containing Mi-2/NuRD co-repressor complex. Protein– protein interaction assays showed that the central region of BCL-6 interacts directly with the carboxyl terminus of MTA3. A GAL4 tethering assay was then used to test the functional effects of the BCL-6– MTA3 interaction. BCL-6 fused to the DNA-binding domain of GAL4 represses transcription of a luciferase reporter construct containing GAL4-binding sites in HeLa cells, which constitutively express high levels of MTA3. But when RNA interference was used to block the
NATURE REVIEWS | IMMUNOLOGY
expression of MTA3, transcriptional repression of the luciferase reporter mediated by BCL-6 was reduced. MTA3 was also shown to be required for BCL-6dependent transcriptional repression in a more physiological setting; the depletion of MTA3 protein from B-cell lines, using RNA interference, resulted in the expression of plasma-cell-specific proteins, such as BLIMP1, that are known to be repressed by BCL-6. Using adenoviral vectors expressing BCL-6 and/or MTA3 to infect plasma-cell lines, the authors showed that marked repression of plasma-cell-specific transcripts and upregulation of B-cellspecific transcripts only occurred when both BCL-6 and MTA3 were co-expressed. This reprogramming of the plasma-cell transcriptional pattern to a B-cell pattern was accompanied by cell-surface expression of B-cell markers, such as CD19, CD20 and HLA-DR, and decreased cytoplasmic staining for the κ-light chain of immunoglobulins. The evidence therefore points to a model in which BCL-6 recruits MTA3 to form a complex that prevents the differentiation of germinal-centre B cells to plasma cells until appropriate signals are received. Additional experiments showed that acetylation of the central domain of BCL-6 prevents interaction with MTA3 and abolishes repressive function, which indicates that such post-translational modification of BCL-6 might be one way in which signals for plasma-cell differentiation are transduced. Kirsty Minton References and links ORIGINAL RESEARCH PAPER Fujita, N. et al. MTA3 and the
Mi-2/NuRD complex regulate cell fate during B lymphocyte differentiation. Cell 119, 75–86 (2004).
VOLUME 4 | NOVEMBER 2004 | 8 3 9 ©2004 Nature Publishing Group
RESEARCH HIGHLIGHTS
IN BRIEF
B - C E L L D E V E LO P M E N T
Recipe for a B cell Cell-fate specification of multipotent haematopoietic stem cells (HSCs) is determined by unique expression patterns of combinations of transcription factors, and for B cells, the five most important factors are PU.1, Ikaros, E2A, EBF (early B-cell factor) and PAX5 (paired box protein 5). Using a series of genetic knockout and reconstitution studies, Harinder Singh and colleagues have begun to order PU.1, EBF and PAX5 in a hierarchical regulatory network that directs an HSC through key developmental stages on the road to becoming a B cell. First, they showed that PU.1 lies upstream of EBF in the B-cell developmental pathway and that EBF can bypass the requirement for PU.1 in early B-cell development. PU.1 is essential for both lymphoid and myeloid development and induces expression of the interleukin-7 receptor (IL-7R) by lymphoid precursors, but when PU.1-deficient fetal liver haematopoietic progenitors (FLPs) were transduced with the gene encoding EBF, IL-7-dependent proliferating cells were rapidly generated that exclusively expressed the B-cell lineage marker CD19 and not the myeloid marker MAC1. The Ebftransduced cells expressed PAX5 and underwent VDJ recombination of the immunoglobulin heavy-chain loci, which are hallmarks of specified pro-B cells. Singh and colleagues identified a putative PU.1binding site in Ebf that is conserved in mouse and human genes and showed that this is functional in vitro and in vivo, indicating that PU.1 directly regulates the expression of EBF. By contrast, PU.1-deficient FLPs could not be rescued by transduction of Pax5, indicating that the function of PAX5, as well as its expression, depends on EBF. However, there was a fivefold reduction in the number of PU.1deficient FLPs that could be rescued with EBF compared with PU.1. PU.1-deficient fetal
840
T- C E L L A C T I V AT I O N
livers contained significantly fewer progenitor cells expressing the cytokine receptor FLT3 (fms-related tyrosine kinase 3), which is known to promote lymphoid versus myeloid cell fate, than did wild-type fetal livers, and FLPs expressing FLT3 were shown to have increased B-cell potential compared with those lacking FLT3. So, PU.1 is required for the generation of FLT3+IL-7R+ lymphoid progenitors, on which EBF acts, together with induced PAX5, to specify B-cell fate. Deficiency of EBF did not affect the percentage of FLT3+IL-7R+ cells, which clearly distinguishes the functions of PU.1 from those of EBF and PAX5 in B-cell development. On the basis of these results, the authors propose a hierarchical regulatory network to describe the generation of B-cell precursors from HSCs. Kirsty Minton References and links ORIGINAL RESEARCH PAPER Medina, K. L. et al. Assembling a gene regulatory network for specification of the B cell fate. Dev. Cell 7, 607–617 (2004).
A role for the immunological synapse in lineage commitment of CD4 lymphocytes. Maldonado, R. A. et al. Nature 431, 527–532 (2004).
The immunological synapse forms at the T cell–antigen-presenting cell interface and is important for optimal T-cell activation. In this study, activation of naive T cells with peptide-loaded dendritic cells (DCs) induced T-cell receptor and interferon-γ receptor (IFN-γR) co-localization at the T cell–DC interface. IFN-γR synapse polarization occurred after activation of naive T cells from C57BL/6 mice (which are TH1-response prone), but polarization was only partial for naive T cells from BALB/c mice (which are TH2-response prone). Furthermore, IL-4 inhibited IFN-γR synapse polarization, so the authors suggest that IL-4 levels may determine T-cell lineage commitment by regulating IFN-γR migration to the synapse. I M M U N OT H E R A P Y
Targeted deletion of T-cell clones using α-emitting suicide MHC tetramers. Yuan, R. R. et al. Blood 104, 2397–2402 (2004).
Fluorescently labelled peptide–MHC class I tetramers are used to detect peptide-specific CD8+ T cells. Yuan et al. generated tetramers composed of three peptide-bound MHC class I molecules and a radioisotope. Multimers containing the γ-emitting isotope 111In and the LMP1 peptide from the Epstein–Barr virus latency membrane protein were used to show that radiolabelled tetramers specifically bind their cognate CD8+ T cells. Using LMP1 peptide–MHC multimers labelled with 225Ac (an α-emitting radioisotope), the authors were able to show specific killing of LMP1-specific CD8+ T cells. Specific killing of CD8+ T cells that recognize the listeriolysin-derived peptide LLO91–99 from Listeria monocytogenes was also shown using 225Ac-labelled LLO91–99–MHC multimers. Further studies using ‘suicide’ tetramers in vivo could allow the generation of T-cell-selective therapies for autoimmunity. SIGNALLING
Human Tribbles, a protein family controlling mitogen-activated protein kinase cascades. Kiss-Toth, E. et al. J. Biol. Chem. 279, 42703–42708 (2004).
Using an assay to identify proteins that mediate inflammatory cytokine signalling, Kiss-Toth et al. found a human homologue of the Drosophila melanogaster gene tribbles. Database searching identified two other homologues. Overexpression of either TRB1 or TRB3 inhibited activation of the transcription factor AP1 mediated by the mitogen-activated protein kinase kinase kinase (MAPKKK) MEKK1. Further analysis indicated that TRB3 acts downstream of MEKK1 and that TRB1 associates with the MAPKKs MEK1 and MAPKK4, whereas TRB3 interacts with MEK1 and MAPKK7. Downstream of this, low levels of TRB3 had distinct effects on the three sets of effector MAPKs: it increased both JNK and ERK activity but inhibited p38 activity. By contrast, high levels of TRB3 inhibited JNK, ERK and p38 activity. So, this study indicates that TRBs regulate MAPK activity by binding to MAPKKs.
| NOVEMBER 2004 | VOLUME 4
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS REGULATORY T CELLS: FRIEND OR FOE IN IMMUNITY TO INFECTION? Kingston H. G. Mills Abstract | Homeostasis in the immune system depends on a balance between the responses that control infection and tumour growth and the reciprocal responses that prevent inflammation and autoimmune diseases. It is now recognized that regulatory T cells have a crucial role in suppressing immune responses to self-antigens and in preventing autoimmune diseases. Evidence is also emerging that regulatory T cells control immune responses to bacteria, viruses, parasites and fungi. This article explores the possibility that regulatory T cells can be both beneficial to the host, through limiting the immunopathology associated with anti-pathogen immune responses, and beneficial to the pathogen, through subversion of the protective immune responses of the host. CHRONIC INFECTIONS
Infections that persist for a long time, often indefinitely, and might not be cleared following the development of anti-pathogen immune responses. These include infection with HIV, hepatitis C virus and many parasites.
Immune Regulation Research Group, Department of Biochemistry, Trinity College, Dublin 2, Ireland. e-mail:
[email protected] doi:10.1038/nri1485
Protection against infection is fundamental to the survival of all animals and is mediated by the immune system, which has evolved both innate and adaptive mechanisms to deal with invading microorganisms. The effector mechanisms used by the host to control infection include production of pro-inflammatory cytokines and chemokines, recruitment of inflammatory cells to the site of infection and activation of cytotoxic T lymphocytes (CTLs) and natural killer (NK) cells, which lyse infected host cells (FIG. 1). Although these responses help to eliminate or slow the spread of the pathogen, if they are not tightly controlled, they can result in severe inflammation and collateral tissue damage1. A further potential for damage arises because the cells and molecules of the immune system that respond to pathogen-derived antigens can also respond to self-antigens, and if this reactivity is uncontrolled, it can result in autoimmune disease2. Inflammation and the immune response to pathogens are regulated by various host suppressor mechanisms, including the production of antiinflammatory cytokines by cells of the innate immune system in response to conserved pathogen-derived products3,4. However, recent evidence indicates that the adaptive immune system might also help to control infection-induced immunopathology through the generation of antigen-specific regulatory T cells (FIG. 1).
NATURE REVIEWS | IMMUNOLOGY
Therefore, regulatory T cells might have a host-protective role in immunity to infection. It is also possible that pathogens can exploit regulatory T cells to subvert the protective immune responses of the host. Although the infections caused by many pathogens are self-limiting in immunocompetent hosts, other pathogens can persist and cause CHRONIC INFECTIONS. In infections such as those caused by HIV, hepatitis C virus (HCV) and many parasites, the pathogen persists because the appropriate immune response required for pathogen elimination either fails to develop or is suppressed. Furthermore, there is evidence that the incidence of atopic diseases (such as allergy and asthma) and autoimmune diseases is lower in individuals infected with helminth parasites or exposed to microbial products as children 5,6. It seems that many, and possibly all, pathogens that cause PERSISTENT INFECTIONS or chronic infections have evolved strategies to subvert the immune responses of the host. These strategies include evasion of humoral and cellular immunity by ANTIGENIC VARIATION, interference with antigen processing or presentation, and subversion of phagocytosis and killing by cells of the innate immune system7. However, of most relevance to this discussion is a common immune-subversion strategy used by many pathogens that involves increasing the production of anti-inflammatory or immunosuppressive
VOLUME 4 | NOVEMBER 2004 | 8 4 1 ©2004 Nature Publishing Group
REVIEWS
PERSISTENT INFECTIONS
Non-lethal infections (such as infection with Bordetella pertussis) that are not cleared immediately (lasting for weeks or months rather than days) and are usually associated with the delayed development or suppression of anti-pathogen immune responses. In persistent viral infections, virus production occurs in a cell that is not killed by the virus (non-lytic); this includes chronic, latent and transforming infections.
responses, which normally function to control or terminate the protective effector immune responses of the host. This can be achieved in the following ways: through the production of molecules with homology to human cytokines, such as viral interleukin-10 (IL-10); through the direct induction of host immunosuppressive cytokines, such as IL-10 and transforming growth factor-β (TGF-β), which are produced by innate immune cells in response to pathogen-derived molecules3,8; or indirectly through the generation of regulatory T cells.
Our understanding of the role of regulatory T cells in immune homeostasis is far from complete, and there are several important unanswered questions. How do regulatory mechanisms control the development of autoimmunity but allow the same type of immune response to mediate protection against infection? What is the advantage to the host of inducing microorganism-activated regulatory T cells that suppress the immune responses that facilitate pathogen elimination? However, our knowledge has increased in the past few years, and several studies, mainly in mouse models of
• TH1-type pro-inflammatory response • Inflammation • Tissue damage • (Autoimmunity)
• TH2-type response • Inflammation • Tissue damage • (Allergic reactions)
ANTIGENIC VARIATION
Changes in the composition, structure or amino-acid sequence of antigenic components of pathogens recognized by T or B cells, which allow the microorganism to escape recognition by the adaptive immune response.
Pathogen invasion
TLR
Macrophage
Dendritic cell PRR
Pathogen
CD80/CD86
MHC class II
CD28 IgA
IL-1β, TNF and chemokines
TCR
Naive T cell
IgG
Neutrophil recruitment
B cell IFN-γ IgE IL-4, IL-5 and IL-6
TH2 cell
TH1 cell
CD8+ T cell
Eosinophil recruitment Epithelial cell
Cell lysis Regulatory T cell
Figure 1 | Immunity to infection. Innate immune effector cells, including macrophages, dendritic cells (DCs), neutrophils and natural killer (NK) cells (not shown), together with various protein components of the complement system, provide the first line of defence against invading microorganisms. Binding of conserved pathogen-derived molecules to pattern-recognition receptors (PRRs), such as Toll-like receptors (TLRs), on the cell surface of macrophages and DCs activates the production of pro-inflammatory cytokines and chemokines, which help to attract other effector cells to the site of infection. Pathogenactivated DCs present pathogen-derived antigens to T cells and promote the differentiation of naive T cells to various subtypes of effector CD4+ and CD8+ T cell. CD4+ T helper 1 (TH1) cells secrete interferon-γ (IFN-γ), which activates the anti-microbial activity of macrophages and helps B-cell production of IgG2a antibodies, whereas TH2 cells provide help for B-cell production of IgG1, IgA and IgE. CD8+ T cells lyse host cells infected with viruses, intracellular bacteria or parasites. Many of these responses can cause host tissue damage — for example, excessive inflammation from uncontrolled pro-inflammatory cytokine and chemokine production by innate immune cells and TH1 cells, eosinophilia and allergic reactions from uncontrolled TH2-cell responses, and killing of host cells by CD8+ cytotoxic T lymphocytes (CTLs) and NK cells. In normal individuals, regulatory T cells (both natural regulatory T cells circulating in the periphery and those induced by infection) help to control these effector functions and the associated damage to host tissues. IL, interleukin; TCR, T-cell receptor; TNF, tumour-necrosis factor.
842
| NOVEMBER 2004 | VOLUME 4
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS
ATHYMIC
Mice that lack a thymus and are therefore deficient in T cells. NUDE
A mutation in mice that causes both hairlessness and defective formation of the thymus, which results in a lack of mature T cells.
infection with bacteria, viruses, parasites or fungi, have shown that regulatory T cells specific for pathogenderived antigens are induced during infection. Furthermore, studies involving depletion or transfer of CD4+CD25+ regulatory T cells (TReg cells) have provided evidence that natural TReg cells can influence the immune response to pathogens and the outcome of infectious disease. This article reviews the recent evidence for pathogen-specific regulatory T cells and their role in infection, focusing on the protective role of these cells in immunity to pathogens, as a means of limiting infection-induced immunopathology, as well as the exploitation of regulatory T cells by pathogens, as an immune-subversion mechanism to prolong pathogen survival in the host. The biology of regulatory T cells
Natural regulatory T cell
Natural and inducible regulatory T cells. It is now firmly established that there are both natural (or constitutive) and inducible (or adaptive) populations of regulatory T cells (FIG. 2), which probably have complementary and overlapping functions in the control of immune responses. However, the lineage relationship, if any, between these subtypes remains to be defined. The failure to identify definitive cell-surface markers for either population has compromised advances in the field and has led to some confusion about the precise nature of
CD4+ CD25+ FOXP3+ TReg cell Antigen (foreign or self)
Naive CD8+ CD25– T cell
CD8+ regulatory T cell
Antigen (foreign or self)
Naive CD4+ CD25– T cell
TH3 cell
Inducible regulatory T cells
Thymus
TR1 cell
Figure 2 | Natural and inducible regulatory T cells. Natural regulatory T cells express the cell-surface marker CD25 and the transcriptional repressor FOXP3 (forkhead box P3). These cells mature and migrate from the thymus and constitute 5–10% of peripheral T cells in normal mice. Other populations of antigen-specific regulatory T cells can be induced from naive CD4+CD25– or CD8+CD25– T cells in the periphery under the influence of semi-mature dendritic cells, interleukin-10 (IL-10), transforming growth factor-β (TGF-β) and possibly interferon-α (IFN-α). The inducible populations of regulatory T cells include distinct subtypes of CD4+ T cell: T regulatory 1 (TR1) cells, which secrete high levels of IL-10, no IL-4 and no or low levels of IFN-γ; and T helper 3 (TH3) cells, which secrete high levels of TGF-β. Although CD8+ T cells are normally associated with cytotoxic T-lymphocyte function and IFN-γ production, these cells or a subtype of these cells can secrete IL-10 and have been called CD8+ regulatory T cells.
NATURE REVIEWS | IMMUNOLOGY
the cells being studied in different laboratories. It seems that natural self-antigen-reactive CD4+CD25+ TReg cells develop in the thymus and then enter peripheral tissues, where they suppress the activation of other self-reactive T cells9,10. By contrast, IL-10- or TGF-β-secreting regulatory T cells, which are known as T regulatory 1 (TR1) or T helper 3 (TH3) cells respectively, are generated from naive T cells in the periphery after encounter with antigen presented by dendritic cells (DCs) that have an activation status distinct from those DCs that promote the differentiation of TH1 or TH2 cells. In addition to these well-defined populations of CD4+ regulatory T cells, there is also evidence for an immunosuppressive function of CD8+ regulatory T cells that secrete either IL-10 or TGF-β11,12. Furthermore, antigen-activated CD8+ γδ T cells can prevent insulin-dependent diabetes in mice13, and IL-10- and TGF-β-producing regulatory γδ T cells can suppress the anti-tumour activity of CTLs and NK cells14. In addition, natural killer T (NKT) cells, which co-express NK-cell and T-cell markers, can secrete regulatory cytokines, including IL-10 (REF. 15). Therefore, NKT and γδ T cells can also be categorized as regulatory T cells. In this review, ‘TReg cells’ denotes only CD4+CD25+ regulatory T cells and ‘T R1 cells’ denotes T regulatory 1 cells. When referring to other types of regulatory T cell or regulatory T cells in general, no abbreviations are used. Natural CD4+CD25+ TReg cells were first defined in 1995 by Sakaguchi and colleagues16, who showed that the transfer into ATHYMIC NUDE mice of lymphoid-cell populations from which CD4+ T cells expressing the α-chain of the IL-2 receptor (IL-2Rα; also known as CD25) had been removed caused spontaneous development of various T-cell-mediated autoimmune diseases. Furthermore, reconstitution with CD4+CD25+ T cells prevented the development of autoimmunity. This discovery, together with the work of Powrie and colleagues on a CD45RBlow population of T cells17, challenged traditional theories about clonal deletion being the only mechanism of self-tolerance and provided convincing evidence that self-antigen-reactive T cells that cause autoimmune diseases can be controlled through active suppression by natural TReg cells. CD4+CD25+ TReg cells, which constitute 5–10% of peripheral T cells in mice, are continuously produced in the thymus as a functionally mature T-cell population that includes cells with immunosuppressive activity. However, CD25 is not a definitive marker of natural regulatory T cells; CD25 is an activation marker for T cells and is therefore also expressed by effector TH1 and TH2 cells, and suppressive function has also been documented for CD25– T cells. These observations led to attempts to find alternative markers for regulatory T cells. Putative markers for regulatory T cells include cell-surface expression of CD38, CD62L, CD103 or glucocorticoid-induced tumournecrosis factor (TNF) receptor (GITR), or low levels of cell-surface CD45RB expression or intracellular expression of the transcriptional repressor FOXP3 (forkhead box P3)17–19. FOXP3 seems to be the most promising marker of natural regulatory T cells, and recent studies have shown that transfection of CD4+CD25– T cells
VOLUME 4 | NOVEMBER 2004 | 8 4 3 ©2004 Nature Publishing Group
REVIEWS
ANERGIC
A state of unresponsiveness by T or B cells to antigens. After stimulation, anergic T cells cannot produce interleukin-2 or proliferate, even in the presence of co-stimulatory signals.
with Foxp3 confers them with intracellular regulatory activity20. T-cell receptor (TCR) engagement seems to be necessary for optimal suppressive activity, and it has been assumed that circulating CD4+CD25+ TReg cells are activated by the recognition of self-antigens in vivo9. However, evidence that natural regulatory T cells are antigen specific is still limited. A unique cytokine-production profile, rather than the expression of cell-surface markers, has been used to define at least two populations of inducible regulatory T cells. Although it had been recognized for some time that T cells with suppressive or ANERGIC activity could be generated in vivo in certain situations — for example, in oral tolerance 21,22 or during infection with certain pathogens, such as rabies virus23, Brugia malayi 24 and Mycobacterium tuberculosis 25 — it was not until the mid-1990s that these cells were given a unified nomenclature. Weiner and colleagues showed that the induction of oral tolerance and the prevention of TH1-cell-mediated autoimmune diseases by feeding self-antigens were associated with the generation of TGF-β-secreting T cells in the gut26. These T cells, which were distinct from TH2 cells in that they produced large amounts of TGF-β and varying amounts of IL-4 and IL-10, were named TH3 cells. In 1997, Groux et al. showed that repeated in vitro stimulation of T cells isolated from ovalbumin-specific TCR-transgenic mice with their cognate antigen in the presence of IL-10 resulted in the expansion of a population of regulatory T cells that produced large amounts of IL-10 and could suppress TH1-cell responses and TH1-cell-mediated autoimmune diseases27; they called these cells TR1 cells. More recently, it has been shown that antigen-specific TR1 cells can be generated in vivo during certain infections and that IL-10 might be a differentiation factor rather than a growth factor for these cells28. Because TH2 cells secrete the immunosuppressive or antiinflammatory cytokines IL-10 and IL-4, these cells might also have regulatory function, as well as effector function, but they are distinguished from T H3 and TR1 cells by the production of large amounts of IL-4 and smaller amounts of IL-10, as well as a lack of TGF-β production. Targets of suppressor activity. Immunity to intracellular pathogens is mediated by CD4+ TH1 cells and CD8+ CTLs, whereas immunity to extracellular pathogens is mediated by antibodies and TH2 cells. Innate immune responses also have a protective role early in infection and instruct the adaptive immune response (FIG. 1). Each of these effector mechanisms can be suppressed by natural and inducible regulatory T cells. It has been shown that T R1 cells and CD4+CD25+ TReg cells can suppress the proliferation of and cytokine production by naive CD4 +CD25– T cells or antigen-specific TH1 or TH2 cells in vitro28–33. There is more-limited evidence that regulatory T cells can suppress pathogen-specific T cells in vivo; existing evidence includes TR1-cell-mediated suppression of interferon-γ (IFN-γ) production by TH1 cells in response to Bordetella pertussis 28 and CD4+CD25+
844
| NOVEMBER 2004 | VOLUME 4
TReg-cell-mediated suppression of CD4+CD25– T cells responding to Leishmania major 29. More recently, it has been shown that CD25+ T cells can suppress the activation of CD8+ T cells in vitro 34, as well as secondary CD8+ T-cell responses to Listeria monocytogenes 35 and herpes simplex virus (HSV)36 in vivo. Finally, there is evidence that regulatory T cells can suppress the recruitment and activation of innate immune cells induced by Helicobacter hepaticus that leads to inflammatory pathology in the colon37. Therefore, the targets of suppressor activity by regulatory T cells are immune responses that confer protection against infection with microorganisms and also responses that can cause collateral damage to host tissue during infection. Mechanisms of suppression. The mechanism of the suppressive function of natural and inducible regulatory T cells is still debated, but in different model systems, suppressive activity has been shown to be mediated either through secretion of immunosuppressive cytokines or through cell–cell contact (FIG. 3). Many studies have shown that the suppression mediated by TR1 or TH3 cells can be reversed using antibodies specific for IL-10 or TGF-β. IL-10 inhibits the production of TNF and IL-12 by DCs and macrophages, whereas TGF-β inhibits TH1-cell responses through its effects on expression of the transcription factor T-bet and the IL-12R38–40. It has been reported that the production of TGF-β by regulatory T cells induces IL-10 production by TH1 cells through SMAD4-induced activation of the IL-10 promoter41. This indicates that there might be interdependent, as well as distinct, roles for IL-10 and TGF-β in the immunosuppressive function of inducible regulatory T cells. Cytokine-mediated suppression might also operate at the level of the antigen-presenting cell, because IL-10 and regulatory T cells can inhibit the expression of MHC class II and co-stimulatory molecules by DCs40,42. The suppressive mechanisms of CD4+CD25+ TReg cells are not clear, but there is evidence that cell–cell contact is required and that expression of the inhibitory costimulatory molecule CTLA4 (CTL antigen 4) might be involved43. However there is also conflicting evidence concerning roles for IL-10 and secreted or cell-surface TGF-β29,43,44. Finally, it has also been proposed that regulatory T cells might inhibit pathogenic effector T-cell responses by competing for shared resources in the normal immune system45. So, although the mechanisms of suppression by TR1 and TH3 cells seem to be mediated mainly by cytokines, CD4+CD25+ TReg cells might use many and as-yet-unidentified mechanisms to mediate suppression. Pathogen-specific regulatory T cells
Studies involving cell depletion and transfer, as well as cytokine-knockout or -inhibition experiments, have provided considerable indirect evidence of a role for inducible (TABLE 1) and natural (TABLE 2) regulatory T cells during infection. However, there are still only a small number of reports showing that regulatory T cells are specific for pathogen-derived antigens.
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS
Natural regulatory T cells
Inducible regulatory T cells
TH3 cell
TReg cell (CD4+ CD25+FOXP3+)
TR1 cell
CD8+ regulatory T cell
?
IL-10 and/or TGF-β
TCR
↓ MHC and co-stimulatory molecules ↓ APC function ↓ Inflammatory cytokines
MHC class II
Dendritic cell
CTLA4
CD80/ CD86 CD28
Cell–cell contact
CD25– cell
↓ Proliferation
TH1 cell
TH2 cell
↓ Proliferation ↓ IFN-γ
↓ Proliferation ↓ IL-4
CD8+ cell ↓ CTL activity ↓ IFN-γ
Figure 3 | Targets of regulatory T cells and mechanisms of suppression. CD4+CD25+FOXP3+ (forkhead box P3) natural regulatory T cells (TReg cells) inhibit the proliferation of CD25– T cells. The mechanism of suppression seems to be multifactorial and includes cell–cell contact. CD4+CD25+ TReg cells express cytotoxic T-lymphocyte antigen 4 (CTLA4), which interacts with CD80 and/or CD86 on the surface of antigen-presenting cells (APCs) such as dendritic cells (DCs), and this interaction delivers a negative signal for T-cell activation. There is also some evidence that secreted or cell-surface transforming growth factor-β (TGF-β) or secreted interleukin-10 (IL-10) might have a role in suppression mediated by natural regulatory T cells. Natural killer T (NKT) cells (not shown) and inducible populations of regulatory T cells, which include T regulatory 1 (TR1) cells, T helper 3 (TH3) cells and CD8+ regulatory T cells, secrete IL-10 and/or TGF-β. These immunosuppressive cytokines inhibit the proliferation of and cytokine production by effector T cells, including TH1 cells, TH2 cells and CD8+ cytotoxic T lymphocytes (CTLs), either directly or through their inhibitory influence on the maturation and activation of DCs or other APCs. IFN-γ, interferon-γ; TCR, T-cell receptor.
TR1 cells and TH3 cells. Although many studies have shown that pathogens, in particular those that cause chronic infections or are associated with immunosuppression, induce production of the regulatory cytokines IL-10 and TGF-β, the cellular source of these cytokines has not always been defined. In some cases, it has been shown that innate immune cells, usually macrophages or more rarely DCs, are the source, whereas in other studies, it has been shown that these cytokines are produced by T cells8. However, the distinction between IL-10 or TGF-β production by TH2 cells versus regulatory T cells has not always been made. The definitive demonstration of antigen-specific regulatory T cells depends on the generation of antigen-specific T-cell clones or on careful ex vivo intracellular cytokine staining of antigen-stimulated T cells, showing high levels of IL-10 production, no IL-4 production and low (human) or no (mouse) IFN-γ production. The first definitive observations of inducible antigenspecific TR1-cell clones generated during infection were made using mice infected with B. pertussis 28 and
NATURE REVIEWS | IMMUNOLOGY
humans infected with HCV46 or the nematode parasite Onchocerca volvulus 47,48. The study involving B. pertussis 28 showed direct evidence of suppression of TH1 cells by TR1-cell clones specific for bacterial antigens, and the human studies46–48 showed indirect evidence of a role for TR1 cells through increased IFN-γ production in the presence of IL-10-specific antibodies. The studies with B. pertussis28 and of virus-specific CD8+ regulatory T cells in chronic HCV infection49 indicate that antigen-specific regulatory T cells are recruited to the site of infection in mucosal tissues. More recently, antigen-specific TR1 cells have been described in several other chronic infections, including infection with Epstein–Barr virus (EBV)50, M. tuberculosis 25,51,52, HIV53 and murine leukaemia virus, which is a mouse model of AIDS54. It has also been shown that IL-10-producing regulatory T cells can be induced in vitro by DCs stimulated with phosphatidylserine isolated from Schistosoma mansoni 55. Although the problems of cultivating and cloning antigen-specific regulatory T cells in vitro have hampered advances in this area, it is tempting to speculate
VOLUME 4 | NOVEMBER 2004 | 8 4 5 ©2004 Nature Publishing Group
REVIEWS that regulatory T cells are induced during infection with most, if not all, pathogens, in particular those that cause persistent or chronic infections. Natural regulatory T cells. Most studies of CD4+CD25+ TReg cells in infection have shown a role for these cells in controlling anti-pathogen immunity, but few studies have shown that they are specific for pathogen-derived antigens (TABLE 2). CD4+CD25+ TReg cells specific for pathogen-derived antigens have been shown to accumulate at the site of infection in the dermis soon after infection with L. major and to suppress IFN-γ production and the ability of effector T cells to eliminate the parasite from the host29. CD4+CD45RBlow regulatory T cells from mice infected with H. hepaticus prevent the development of intestinal inflammation induced by the transfer of CD4+ T cells from IL-10-deficient mice to recombinationactivating gene (RAG)-deficient mice56. The observations that CD4+CD45RBlow regulatory T cells from H. hepaticusinfected wild-type mice inhibit IFN-γ production by
T cells from IL-10-deficient mice and produce IL-10 after exposure to H. hepaticus-derived antigens in vitro indicate that these regulatory T cells, rather than being endogenous, might be a memory population resulting from previous exposure to bacterial antigens. Protective role of regulatory T cells in infection
There is convincing evidence of a protective role for regulatory T cells against autoimmune diseases, allograft rejection and allergy, in these situations, they suppress potentially pathogenic immune responses mediated by effector TH1 cells, TH2 cells or CTLs2,16,27,57–59. As these effector T-cell responses also have important roles in protection against pathogens, it might seem counterintuitive that regulatory T cells could have a protective role in infection. However, in many infectious diseases, immune responses to the pathogen can result in collateral damage to host tissues, and immunoregulatory mechanisms, including the induction of regulatory T cells, are essential to control this immunopathology.
Table 1 | Pathogen-induced regulatory T cells and their role in infection Pathogen
Cell type
Antigen T-cell specific?* clones
Cytokine secreted
Responses suppressed
Manipulation of regulatory cells
Effect on immune response, immunopathology and pathogen load
References
Friend virus
Mouse TR1 cell
ND
ND
IL-10
CD8+ T-cell IFN-γ production
Depletion with GITR-specific antibody
Increases IFN-γ-secreting CD8+ T cells and reduces viral load
84
Murine leukaemia virus
Mouse TR1 CD4+CD25+ cell
ND
ND
IL-10
ND
CD25+ T-cell depletion
Prevents spleen pathology and disease progression but has no effect on viral load
54
HCV
Human TR1cell
Yes
Yes
IL-10
PBMC IFN-γ production
ND
ND
46,60
HCV
Human CD8+ T cell
Yes
ND
IL-10
Antigen-specific PBMC proliferation
ND
ND
49
EBV
Human TR1 cell
Yes
IL-10
T-cell proliferation ND and IFN-γ production to recall antigen
ND
50
HIV
Human CD8+ T cell
Yes
ND
TGF-β
Vaccinia-virusspecific CD8+ T-cell IFN-γ production
ND
11
Bordetella pertussis
Mouse TR1 CD4+CD25+ cell
Yes
Yes
IL-10 +/– TGF-β
Antigen-specific Cell transfer TH1-cell proliferation and IFN-γ production in vitro and in vivo
Suppresses TH1-cell response and increases bacterial load
28
Bordetella pertussis
Mouse TR1 cell
Yes
Yes
IL-10
TH1-cell IFN-γ production
Defective TR1 cells in TLR4-deficient mice
Increases TH1-cell response, lung inflammation and bacterial load
67
Helicobacter hepaticus
Mouse CD45RBlow T cell
Yes
ND
IL-10
Antigen-specific IL-10-deficient CD4+ T-cell IFN-γ production
CD45RBlow-cell transfer
Prevents colitis in IL-10deficient mice
56
Yes
ND
IL-10
Allogeneic CD4+ T-cell proliferation
ND
ND
25,51
Yes Human TH3 and/or TR1 cell
Yes
IL-10 +/– TGF-β
PBMC proliferation
ND
ND
47,48
Viruses
ND
Bacteria
Mycobacterium Human tuberculosis TR1 cell Parasites Onchocerca volvulus
*Demonstration that regulatory cells respond to the pathogen or pathogen-derived antigens in vitro. EBV, Epstein–Barr virus; GITR, glucocorticoid-induced tumour-necrosis factor receptor-related protein; HCV, hepatitis C virus; IFN-γ, interferon-γ; IL-10, interleukin-10; ND, not determined; PBMC, peripheral-blood mononuclear cell; TGF-β, transforming growth factor-β; TH, T helper; TLR4, Toll-like receptor 4; TR1, T regulatory 1.
846
| NOVEMBER 2004 | VOLUME 4
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS
MIXED CRYOGLOBULINAEMIA
Cryoglobulins are antibodies that precipitate at cold temperatures and dissolve on warming. Mixed cryoglobulinaemia is a B-cell proliferative disorder that is characterized by polyclonal B-cell activation and autoantibody production. Patients with mixed cryoglobulinaemia have circulating cryoproteins and inflammation of small blood vessels, with inflammatory changes prominent in the skin (vasculitis), and might have renal and neurological involvement.
Viruses. IL-10-producing CD4+ and CD8+ regulatory T cells have been shown in HCV infection, and there is indirect evidence that both CD4+ and CD8+ T-cell populations can inhibit HCV-specific T cells in chronically infected individuals46,49,60. However, it has been suggested that HCV-specific CTLs that home to the liver produce IL-10 and help to reduce liver inflammation49. Furthermore, HCV-infected patients with reduced numbers of CD4+CD25+ T cells often develop an autoimmune syndrome, known as MIXED CRYOGLOBULINAEMIA, which is characterized by B-cell proliferation and autoantibody production 61. So, although regulatory T cells can prevent viral clearance, they also prevent immunopathology and the development of autoimmunity.
More-direct evidence of a role for regulatory T cells in preventing immunopathology has come from studies using mouse models of viral infection. Infection of mice with Theiler’s virus induces a DEMYELINATING disease mediated by CD4+ T cells, and the transfer of virus-specific CD8+ regulatory T cells has been shown to prevent inflammation and the pathogenic effects of the CD4+ T cells12. In footpad infection of mice with HSV, removal of CD25+ T cells increases the virus-specific CD8+ T-cell response and improves viral clearance36. However, TH1-cell responses and the severity of T-cell-mediated lesions in the cornea of HSV-infected mice were increased if mice were depleted of CD25+ T cells before infection62. CD4+CD25+ TReg cells therefore seem to reduce the severity of immune-mediated inflammatory
Table 2 | Natural CD4+CD25+ TReg cells and their role in infection Pathogen
Species
Antigen specific?*
Cytokine secreted
Responses suppressed‡
Manipulation of regulatory T cells
Effect on immune response, immunopathology and pathogen load
References
Herpes simplex virus
Mouse
ND
IL-10
Antigen-specific CD4+ T-cell IFN-γ production
In vivo depletion
Increases TH1-cell response, CD4+ T-cell infiltration and stromal keratitis
62
HIV
Human
ND
ND
Antigen-specific CD4+ and CD8+ T-cell proliferation and cytokine production
In vitro depletion
Increases HIV-specific CD8+ T-cell IFN-γ production
81,82
Helicobacter hepaticus
Mouse
ND
ND
Innate immune responses in vivo
Transfer
Prevents H. hepaticus-induced intestinal inflammation (reversed by IL-10- or TGF-β-specific antibody) but has no effect on bacterial colonization
37
Helicobacter pylori
Mouse
ND
ND
ND
In vivo depletion
Increases CD4+ T-cell IFN-γ production and gastritis but decreases bacterial load
68
Helicobacter pylori
Human
ND
ND
Antigen-specific CD25low T-cell proliferation and IFN-γ production
In vitro depletion
Increases antigen-specific T-cell proliferation
30
Schistosoma mansoni
Mouse
ND
IL-10
Naive T-cell proliferation
Transfer
Reduces liver damage and increases survival
77
Leishmania major
Mouse
Yes
IL-10
CD25– T-cell proliferation and IFN-γ production in vitro and in vivo
Transfer to IL-10deficient or wildtype mice
Increases non-healing skin lesions and parasite load
29
Leishmania major
Mouse
ND
ND
ND
Depletion from splenocytes before transfer to SCID mice
Increases antigen-specific lymph node IFN-γ and IL-4, severity of colon lesions and parasite load
78
Plasmodium yoelii
Mouse
ND
ND
ND
In vivo depletion
Increases splenocyte proliferation and decreases parasite load
91
Plasmodium berghei
Mouse
ND
ND
ND
In vivo depletion
Decreases parasite load
92
Pneumocystis carinii
Mouse
ND
ND
ND
Transfer
Increases pathogen load
69
Candida albicans
Mouse
ND
IL-10, IL-4 and TGF-β
ND
Transfer
Decreases IFN-γ-secreting cells in vivo and decreases fungal load
70
Viruses
Bacteria
Parasites
Fungi
*Demonstration that regulatory cells respond to the pathogen or pathogen-derived antigens in vitro. None of these studies showed antigen-specific CD4+CD25+ regulatory T (TReg) cells. ‡In vitro unless otherwise stated. IFN-γ, interferon-γ; IL, interleukin; ND, not determined; SCID, severe combined immunodeficient; TGF-β, transforming growth factor-β; TH1, T helper 1.
NATURE REVIEWS | IMMUNOLOGY
VOLUME 4 | NOVEMBER 2004 | 8 4 7 ©2004 Nature Publishing Group
REVIEWS
DEMYELINATING
Causing damage to the myelin sheath surrounding nerves in the brain and spinal cord, which affects the function of the nerves involved. Demyelination occurs in multiple sclerosis, a chronic disease of the nervous system affecting young and middleaged adults, and in experimental autoimmune encephalomyelitis, which is a mouse model of multiple sclerosis. SECONDARY INFECTION
An infection in a host already infected with another pathogen, often caused by opportunistic pathogens in immunodeficient or immunosuppressed hosts. TOLL-LIKE RECEPTOR
(TLR). A member of a family of receptors that recognize pathogen-associated molecular patterns. TLRs recognize conserved molecular patterns that are common to large groups of microorganisms and/or viruses.
848
lesions by preventing the induction of pathogenic CD4+ T cells and by limiting the migration of these cells to inflammatory sites. Therefore, in chronic viral infections, regulatory T cells might be beneficial to the host by maintaining a balance between efficient effectors and memory responses, but with a low level of inflammation that causes minimal damage to the host. Bacteria. Indirect evidence of a protective role for IL-10producing regulatory T cells in host defence against bacteria-induced immune-mediated pathology has come from studies showing that disease severity is increased in IL-10-deficient mice. IL-10-deficient mice succumb more readily to primary and SECONDARY INFECTION with L. monocytogenes than control mice; the IL-10-deficient mice have an increased number of cells in the inflammatory infiltrate, increased production of pro-inflammatory cytokines in the brain and increased severity of brain lesions63. Peritonitis and mortality from infection with Escherichia coli is increased in IL-10-deficient mice, despite accelerated clearance of the bacteria compared with wild-type animals64. Colonization of the gastric mucosa by Helicobacter pylori is reduced in IL-10deficient mice, but the severity of chronic active gastritis is significantly greater than in wild-type mice65. Similarly, IL-10-deficient mice infected with H. hepaticus develop severe inflammation that is associated with IL-12 production and TH1-cell responses66. The IL-10 that helps to limit inflammation during bacterial infection might, in part, be derived from innate immune cells. However, it has been shown that the induction of IL-10 production by macrophages and DCs in response to certain pathogen-derived molecules facilitates the induction of TR1 cells, thereby amplifying the effect of IL-10 produced by innate immune cells 28. Indeed, studies with mice that lack functional TOLL-LIKE RECEPTOR 4 (TLR4-defective mice) have indicated that IL-10 produced by both innate immune cells and TR1 cells might help to limit inflammatory pathology in the lungs induced by B. pertussis infection. B. pertussis stimulates IL-10 production by DCs and macrophages and generates TR1 cells in the respiratory tract of infected mice. However, TLR4-defective mice have reduced IL-10 production by DCs and macrophages and do not generate TR1 cells when infected with B. pertussis67. These mice have significantly greater cellular infiltrates, lung damage and bacterial loads than wild-type mice, which has led to the hypothesis that the induction of TR1 cells helps to limit inflammatory pathology and thereby improves pathogen elimination by preventing damage to the ciliated epithelial cells required for removal of the bacteria from the lungs. Direct evidence of a role for CD4+CD25+ TReg cells in the prevention of intestinal inflammation has come from the demonstration that infection with H. hepaticus induces a population of CD4+CD45RBlow regulatory T cells that inhibit the development of colitis in IL-10-deficient mice56. Removal of CD25+ T cells from the lymph-node cells used to reconstitute athymic mice before infection with H. pylori reduced the bacterial load but increased the severity of gastritis 68.
| NOVEMBER 2004 | VOLUME 4
Furthermore, transfer of CD4+CD25+ TReg cells from normal mice to Rag2 –/– mice prevented intestinal inflammation induced by infection with H. hepaticus, through IL-10- and TGF-β-dependent mechanisms37. The regulatory T cells did not affect bacterial colonization in the gut; instead, the protective effect of the regulatory T cells seemed to be mediated by suppressing T-cell-dependent and innate inflammatory responses, including the recruitment of neutrophils and macrophages and the activation of NK cells in the intestine37. Fungi. Pneumocystis carinii causes pneumonia in immunocompromised individuals. In a mouse model, transfer of CD4+CD25– T cells to Rag2–/– mice infected with P. carinii reduced the pathogen load, but these mice developed severe lung inflammation and a fatal wasting disease. Co-transfer of CD4+CD25+ T cells prevented lung inflammation and the development of disease induced by CD4+CD25– T cells but increased the pathogen load69. A similar situation has been reported for infection with Candida albicans. TH1 cells mediate protection against C. albicans, and in the absence of CD4+CD25+ T cells, which are not induced in CD86- or CD28-deficient mice or in situations in which IL-10mediated signalling is deficient, the fungal growth is reduced, but inflammatory pathology is increased70. Furthermore, transfer of IL-10- and TGF-β-secreting CD4+CD25+ T cells decreased inflammation in CD86deficient mice. In a separate study, an absence of TLR2 was associated with impaired IL-10 production, fewer CD4+CD25+ TReg cells and more inflammatory infiltrate, but lower pathogen load in C. albicans-infected mice71. So, although regulatory T cells might compromise fungal clearance, they can also be beneficial to the host by limiting infection-induced pathology. Parasites. In human malaria, polymorphisms in the TNF promoter have been associated with disease severity; among children with severe malaria, those with the TNF-308A allele had lower plasma levels of IL-10 than of TNF72. Furthermore, higher ratios of IL-10 to TNF in children with mild malaria compared with those who have severe malaria indicate a role for IL-10 in controlling the excessive inflammatory activities of TNF73. Although these studies do not provide direct evidence that IL-10 is produced by parasite-specific regulatory T cells, it has been suggested that regulatory T cells might contribute to the anti-inflammatory cytokine pool that controls TNF-mediated inflammation in malaria1. These observations are complemented by studies in mice showing that infection with Plasmodium chabaudi chabaudi is more severe in IL-10-deficient mice than in wild-type mice and that this is associated with increased inflammation, including increased production of TNF, IFN-γ and IL-12 (REF. 74). Treatment of these infected, IL-10-deficient mice with TNF-specific antibodies increases survival74. Furthermore, treatment of infected mice with a neutralizing antibody specific for TGF-β exacerbated infection with Plasmodium berghei and P. chabaudi chabaudi, and treatment with recombinant
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS
SEVERE COMBINED IMMUNODEFICIENT
(SCID). Mice with this immunesystem defect do not have B or T cells and therefore can accept tumour cells from another species without rejection.
TGF-β slowed the rate of parasite growth and increased survival75. CD4+CD25+ T cells and CD8+ T cells from malaria-infected mice secreted high levels of TGF-β in response to parasite-derived antigens in vitro76, which indicates that antigen-specific CD8+ regulatory T cells might help to control malaria-induced inflammation. IL-10-producing CD4+CD25+ T cells are induced in mice during infection with S. mansoni, and these regulatory T cells (as well as the production of IL-10 by innate immune cells) help to protect the host from the hepatocyte damage induced by S. mansoni eggs and to prevent death from the infection through immune-mediated pathology77. Similarly, depletion of CD4+CD25+ T cells increased the parasite load, severity of colon lesions and colitis in L. major-infected SEVERE COMBINED IMMUNODEFICIENT (SCID) mice adoptively transferred with splenocytes78. Although infection of IL-10-deficient mice with L. major is associated with increased parasite-specific immune responses and pathogen clearance from the host, this sterilizing cure results in a loss of immune memory and therefore of resistance to re-infection by the same parasite29. Therefore, regulatory T cells seem to control the immune response sufficiently to contain but not eradicate the infection, thereby suppressing potentially pathogenic T-cell effector responses but allowing the maintenance of T-cell memory. Pathogen immune evasion
Although regulatory T cells are beneficial to the host by preventing immunopathology and enabling the development of immune memory, they can also be beneficial to the pathogen, enabling it to establish a chronic infection. Many pathogens have evolved strategies that facilitate their persistence, largely through their ability to evade or subvert the host immune response. One strategy is to induce a state of immunosuppression, either through direct interference with host immune effector mechanisms or through the production of immunosuppressive cytokines. Many viruses produce antagonists of pro-inflammatory cytokines or their receptors, or molecules that are homologous to host IL-10 or TGF-β, or they stimulate the production of anti-inflammatory cytokines by host macrophages or other innate immune cells3,8. It has recently been recognized that parasite-induced immunosuppression can also be extended to the induction of T cells with suppressor activity, including natural and inducible regulatory T cells. Viruses. Most patients infected with HCV remain persistently infected despite the induction of HCV-specific antibodies and T-cell responses. Many chronically infected patients remain disease free for decades; others go on to develop cirrhosis of the liver and, in certain cases, hepatocarcinoma. It has recently been shown that patients with chronic HCV infection have circulating HCV-specific CD4+ TR1 cells46 and CD8+ regulatory T cells49. These regulatory T cells seem to be capable of inhibiting anti-viral immunity, because the addition of a neutralizing IL-10-specific antibody significantly increased HCV-specific IFN-γ production by T cells
NATURE REVIEWS | IMMUNOLOGY
in vitro60. Furthermore, there is a higher frequency of CD4+CD25+ TReg cells in patients with chronic infections than in those that have cleared the infection79. These CD4+CD25+ TReg cells could suppress HCV-specific CTL responses, indicating that natural TReg cells could also contribute to chronic infection by suppressing protective immune responses. This is consistent with findings in HSV-infected mice, in which CD4+CD25+ TReg cells suppress virus-specific CD8+ T-cell responses and delay viral clearance36. Retroviruses, such as HIV, usually persist for the lifetime of the infected host and escape immunity by antigenic variation. The immunodeficiency syndrome in the later stages of AIDS is the direct result of a reduction in the number of CD4+ T cells. However, even before the numbers of CD4+ T cells start to decline, immune responses to HIV and unrelated pathogens are suppressed. One explanation for this is the switch from TH1- to TH2-cell-dominated responses that has been observed to occur during disease progression80. Alternatively, activation of regulatory T cells that inhibit TH1-cell and CTL responses in vivo might explain the immunosuppression that occurs during retroviral infection before depletion of CD4+ T cells. Individuals with progressive or active HIV replication have a high frequency of IL-10-producing CD4+ T cells; these cells include those that produce IL-10 constitutively and those that only produce it after stimulation with the HIV protein Gag (group-specific antigen)53. Furthermore, CD4+CD25+ T cells from HIV-infected individuals suppress the proliferation of and cytokine production by CD8+ and CD4+ T cells in response to HIV antigens81, and depletion of CD4+CD25+ TReg cells from peripheral-blood mononuclear cells (PBMCs) increases the frequency of CD8+ and CD4+ T cells that secrete IFN-γ in response to HIV81,82 and cytomegalovirus82 antigens. In addition, HIV antigens induce TGF-β-secreting CD8+ regulatory T cells that inhibit IFN-γ secretion by CD8+ T cells specific for vaccinia virus. So, regulatory T cells specific for HIV antigens might contribute to general immunosuppression during retroviral infection11, and this conclusion is supported by studies using animal models of retroviral infection. Infection of cats with feline immunodeficiency virus is associated with the activation of CD4+CD25+CTLA4+ TReg cells that inhibit the proliferation of and IL-2 production by CD4+CD25– T cells from normal cats31. Furthermore, ablation of IL-10-secreting regulatory T cells in mice prevented the progression of mouse AIDS (an immunodeficiency syndrome induced by murine leukaemia virus)54. Persistent infection of mice with the Friend retrovirus is associated with a decreased ability to develop anti-tumour immune responses83. IL-10-producing CD4+ T cells from mice persistently infected with Friend virus suppress IFN-γ production by CD8+ T cells84. Similarly, in humans infected with EBV, TR1 cells are induced that are specific for LMP1 (latent membrane protein 1) of EBV, and these cells inhibit TH1-cell responses to other EBV proteins, which might facilitate viral persistence and promote the induction of EBV-associated tumours50.
VOLUME 4 | NOVEMBER 2004 | 8 4 9 ©2004 Nature Publishing Group
REVIEWS These findings indicate that, in some cases, virus-specific regulatory T cells not only prevent pathogen elimination but also can promote a generalized state of immune suppression in vivo such that the host is more susceptible to secondary infections with other pathogens or has reduced resistance to tumours. Bacteria. A proportion of individuals who are infected with M. tuberculosis do not have positive skin-test responses to mycobacterial purified protein derivative (PPD), and this absence of delayed-type hypersensitivity responses to mycobacterial antigens is associated with a poorer clinical outcome. T cells from patients with positive PPD skin tests proliferated and secreted IFN-γ and IL-10 in response to PPD, whereas T cells from non-responding patients produced IL-10 but not IFN-γ 25,51. Furthermore, IL-10-specific antibodies increased PPD-specific IFN-γ production by T cells from non-responding patients25. This indicates that TR1 cells that suppress TH1-cell responses to PPD through IL-10 production mediate T-cell suppression in patients with tuberculosis. In addition, recent studies using mice transgenic for the gene encoding IL-10 show that reactivation of chronic M. tuberculosis infection and suppression of protective TH1-cell responses is strongly influenced by the expression of IL-10 during the latent phase of infection85. Furthermore, cell-mediated immunity to Mycobacterium bovis bacillus CalmetteGuérin (BCG) is increased in IL-10-deficient mice and these mice eliminate the bacteria faster than wildtype mice86. Collectively, these findings indicate that IL-10-producing cells, probably TR1 cells as well as innate immune cells, contribute to the chronic state of mycobacterial infections. Similarly, during infection with Yersinia enterocolitica, the V antigen of this pathogen stimulates IL-10 production by macrophages, which suppresses production of the host-protective cytokine TNF, and IL-10-deficient mice are highly resistant to infection with Y. enterocolitica 87. Although Y. enterocolitica-specific regulatory T cells have not yet been documented, it is probable that the bacteriatriggered IL-10 production by macrophages facilitates suppression of protective immunity either directly or indirectly through the induction of regulatory T cells. This conclusion is supported by studies with B. pertussis, in which two bacterial virulence factors — filamentous haemagglutinin (FHA) and adenylate cyclase toxin (CyaA), which stimulate IL-10 production by macrophages and DCs — direct the induction of TR1 cells in vivo 28,88. Suppression of local TH1-cell responses early in infection with B. pertussis 89 seems to result from the induction of regulatory T cells, which can be detected in the lung during acute infection28, even before the appearance of TH1 cells (P. McGuirk and K.H.G.M., unpublished observations). Furthermore, co-transfer of B. pertussis-specific TR1-cell clones with TH1 cells from convalescent mice to naive mice before infection with B. pertussis suppressed TH1-cell responses and exacerbated infection28. Therefore, it seems that the persistence of infection with certain bacteria might be associated with the induction of IL-10-producing CD4+ TR1 cells 850
| NOVEMBER 2004 | VOLUME 4
and that this might be a strategy used by the bacteria to evade host immune responses. Indirect evidence of a suppressive role for CD4+ regulatory T cells in the development of memory CD8+ T cells has also been provided by studies using mice infected with L. monocytogenes, in which removal of CD4+ T cells increases the generation of CD8+ T-cell responses and increases protection against infection induced by immunization with killed L. monocytogenes 90. Parasites. Infection with malaria parasites is persistent and is associated with suppressed immune responses both to the parasite and to unrelated antigens. In a mouse model of malaria, depletion of CD4+CD25+ TReg cells protected mice against lethal infection with Plasmodium yoelii 91 and reduced the parasite load in naive and immunized mice infected with P. berghei 92. CD25+ T-cell depletion also reversed the defect in the proliferative response of splenocytes from P. yoelii-infected mice to parasitized erythrocytes91. Furthermore, treatment of P. yoelii-infected mice with antibodies specific for TGF-β and IL-10 reduced parasitaemia and increased survival76. Similarly, infection of IL-10-deficient mice with L. major results in more rapid clearance of the infection93. Although susceptibility to L. major infection in BALB/c mice has been associated with IL-4 production and TH2-cell polarization, recent evidence indicates that IL-10 production by regulatory T cells might have an important role in persistence of L. major infection in these mice94. IL-10-deficient mice clear L. major infection more rapidly than wild-type mice. Furthermore, CD4+CD25+ TReg cells specific for L. major antigens have been shown to accumulate rapidly at the site of infection in the dermis. These regulatory T cells suppress the ability of effector T cells to eliminate the parasite from the host29. Therefore, pathogens have evolved strategies for persistence by the subversion of host-protective immune responses through activation of anti-inflammatory cytokine production by innate immune cells and through activation of natural and inducible regulatory T cells. Pathogen-activated DCs induce regulatory T cells
During infection, the differentiation of naive T cells to distinct effector CD4+ T-cell subtypes is controlled by DCs and regulatory cytokines produced by innate immune cells (FIG. 4). Following the binding of conserved pathogen-derived molecules to pattern-recognition receptors, such as TLRs, on the surface of immature DCs at the site of infection, the DCs mature and migrate to the lymph nodes, where they present antigen to naive T cells. The differentiation of naive T cells to TH1 cells is promoted by the production of IL-12, IL-23 and IL-27, whereas differentiation to TH2 cells is promoted by IL-4 and IL-6. Although there is some evidence that immature DCs can selectively activate regulatory T cells95, it seems that T cells induced by immature DCs are anergic rather than regulatory and that the DCs that direct the induction of regulatory T cells have an intermediate phenotype, including increased expression of MHC class II
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS molecules and CD86 but low levels of expression of CD40 and intercellular adhesion molecule 1 (ICAM1)32,88,96. This is supported by studies showing that DCs lacking surface expression of CD40 can suppress a primed immune response and induce IL-10secreting CD4+ regulatory T cells96. Furthermore, the interaction between ICAM1 and leukocyte functionassociated antigen 1 (LFA1) is thought to promote the induction of T H1 cells independently of IL-12. So, DCs in which CD40 and ICAM1 expression is suppressed but CD80 and CD86 expression is increased might promote the induction of TR1 cells but block the differentiation of TH1 cells 8. The cytokine environment that promotes the differentiation of regulatory T cells is also distinct from that which drives TH1- and TH2-cell differentiation.
TH1-cell-promoting pathogen molecules (such as the TLR ligands LPS, CpG-containing DNA and viral RNA)
Pathogen molecules that inhibit IL-12 production and increase IL-10 production have been shown to promote the induction of TR1 cells in vitro and in vivo8. FHA from B. pertussis induces IL-10 production and inhibits IL-12 production by DCs and macrophages, and FHA-stimulated DCs promote the clonal expansion of IL-10-secreting T cells from naive T cells in vitro28. Furthermore, incubation of FHA-stimulated DCs with an IL-10-specific antibody prevents the induction of TR1 cells, which indicates that IL-10 is a differentiation factor for TR1 cells. Cholera toxin and B. pertussis CyaA also inhibit IL-12 production and CD40 expression by DCs and synergize with TLR ligands in activating IL-10 production by DCs and macrophages32,88. DCs activated by these pathogen-derived molecules induce TR1 and TH2 cells.
TH2-cell-promoting pathogen molecules (helminth products, yeast hyphae, cholera toxin, LT and TLR2 ligands)
TR1/TH3-cell-promoting pathogen molecules (FHA, CyaA, cholera toxin, NS4, and S. mansoni phosphatidylserine)
Innate immune cell
TLR Macrophage
CD11b–CD18 Semi-mature DC (CD40–CD80+)
Mature DC (CD40+CD80+CD86+)
Mature DC (CD86+OX40L+) PRR
IL-12 and IL-27
CD40
MHC class II
CD40L
TCR
IL-10 and TGF-β
CD28 Regulatory T cell
TH1 cell
IL-1β and TNF
IFN-γ
• TH1-type responses • Pro-inflammatory responses • Immunity to intracellular pathogens • (Autoimmune diseases)
IL-4 and IL-6
CD80/ CD86
IL-10 and/or TGF-β • Immune regulation • Anti-inflammatory responses
OX40L OX40
TH2 cell
IL-4, IL-5, IL-6, IL-10 and IL-13 • TH2-type responses • Immunity to extracellular pathogens • (Allergy)
Figure 4 | Role of pathogen-derived molecules in promoting the induction of regulatory T cells versus TH1 and TH2 cells. Pathogens produce a range of conserved molecules that interact with pattern-recognition receptors (PRRs), such as Toll-like receptors (TLRs), on the surface of innate immune cells, including macrophages and dendritic cells (DCs). Most TLR ligands, including lipopolysaccharide (LPS), CpG-containing DNA and viral RNA, activate DC maturation (that is, upregulate cell-surface expression of CD40, CD80 and CD86) and production of interleukin-12 (IL-12) and IL-27, leading to T helper 1 (TH1)-cell induction. Distinct families of pathogen-derived molecules — including filamentous haemagglutinin (FHA) and adenylate cyclase toxin (CyaA) from Bordetella pertussis, cholera toxin, hepatitis C virus non-structural protein 4 (NS4) and Schistosoma mansoni-specific phosphatidylserine — interact with PRRs, including CD11b–CD18, ganglioside GM1 or TLR2 on the surface of DCs. This stimulates IL-10 production and inhibits IL-12 production by macrophages and DCs and activate DCs to a semi-mature or intermediate phenotype, which promotes the induction of T regulatory 1 (TR1) and/or TH3 cells. Finally, yeast hyphae, cholera toxin, CyaA, Escherichia coli heat-labile enterotoxin (LT) and products of helminth parasites, which stimulate the production of IL-4 and/or IL-6 by DCs or other innate immune cells, promote the induction of TH2 cells. IL-10 and transforming growth factor-β (TGF-β) produced by regulatory T cells and by innate immune cells inhibit the activation of TH1 cells, which mediate immunity to intracellular pathogens, and the activation of TH2 cells, which mediate immunity to extracellular pathogens. However, these immunosuppressive cytokines also prevent innate inflammatory responses, autoimmunity and allergy, which are mediated by pathogenic TH1 and TH2 cells, respectively. CD40L, CD40 ligand; IFN-γ, interferon-γ; OX40L, OX40 ligand; TCR, T-cell receptor; TNF, tumour-necrosis factor.
NATURE REVIEWS | IMMUNOLOGY
VOLUME 4 | NOVEMBER 2004 | 8 5 1 ©2004 Nature Publishing Group
REVIEWS
Endothelial cell
a
Pathogen persistence Pathogen
TH1/2 TCR MHC class II Naive T cell
CD28 CD80/CD86 Semi-mature DC Effector T cells IL-10 or TGF-β Regulatory T cell
Thymus
Natural regulatory T cell
Natural and inducible regulatory T cells Mature DC
b
Pathogen clearance with immunopathology
IL-4 or IL-12
Pathogen clearance with limited immunopathology and memory development
c
IL-4 or IL-12
IL-10 or TGF-β
Figure 5 | Protective immunity versus immunopathology depends on a balance between regulatory and effector T cells. I propose a model in which certain pathogens in different individuals can induce three different responses. a | Pathogens can stimulate potent pathogen-specific regulatory T-cell responses, through the selective induction of interleukin-10 (IL-10) or transforming growth factor-β (TGF-β) production by innate immune cells, which together with natural CD4+CD25+ regulatory T (TReg) cells can inhibit the generation and function of effector T cells and prevent clearance of the microorganism. This immuneevasion strategy is used by many pathogens that cause chronic infections. b | Pathogens can induce T helper 1 (TH1)-cell-biased or TH2-cell-biased immune responses in certain individuals, by activating innate IL-12 or IL-4 production, respectively. In cases in which the number of regulatory T cells is limiting, as a result of either a defect in natural CD4+CD25+ TReg cells or the limited induction of inducible regulatory T cells — T regulatory 1 (TR1) cells and/or TH3 cells — by the pathogen, these effector cells can mediate clearance of the microorganism. However, the absence of control by an appropriate complement of regulatory T cells allows these effector T cells to cause pathological damage to host tissue. c | Pathogens can induce a balanced number of regulatory and effector T cells, possibly by stimulating both IL-10/TGF-β and IL-4/IL-12 production by innate immune cells, thereby allowing effector T-cell-mediated clearance of the microorganism, together with control of the inflammatory response by regulatory T cells, which limits damage to host tissue. DC, dendritic cell; TCR, T-cell receptor.
852
| NOVEMBER 2004 | VOLUME 4
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS
ALLOREACTIVE
Responding to antigens that are distinct between members of the same species, such as MHC molecules or blood-group antigens.
Lactobacillus paracasei inhibits the proliferation of and the TH1 and TH2 cytokine production by ALLOREACTIVE T cells, but it increases IL-10 and TGF-β production, indicating that these bacteria probably promote the induction of regulatory T cells97. C. albicans hyphae induce IL-4 and IL-10 production by DCs, and hyphaepulsed DCs can induce the clonal expansion of CD4+CD25+ T cells in vivo. Furthermore, an antibody specific for the IL-10R prevented the clonal expansion of CD4+CD25+ T cells in C. albicans-infected mice70. It seems, however, that the source of innate IL-10 that promotes the induction of TR1 cells is not confined to DCs. Non-structural protein 4 from HCV stimulates IL-10 production by monocytes (but not DCs), which in turn activates DCs to induce TR1 cells at the expense of TH1 cells60. EBV infection is associated with the induction of TR1-type cells specific for LMP1. EBV produces a viral homologue of mammalian IL-10, which is expressed, together with LMP1, during the lytic cycle. Because LMP1-specific IL-10-secreting T cells are induced in EBV-infected individuals, viral IL-10 might help to promote the differentiation of these TR1 cells in vivo50. Therefore, the production of immunoregulatory cytokines — IL-10 and TGF-β, and possibly IFN-α — by innate immune cells in response to certain pathogenderived products, together with the suppression of IL-12 production and the selective activation of co-stimulatory molecule expression by DCs, might have an important influence on the induction of regulatory T cells during infection. Finally, it has also been proposed that regulatory T cells might respond directly to physiological or pathogenic ligand interaction with TLR4 (REF. 98) or CD46 (REF. 99) expressed on the surface of T cells. Conclusions and therapeutic prospects
The study of regulatory T cells in the context of infection has shown that these cells form an essential component of the protective armoury of the host immune system through their ability to limit immunopathology and allow the development of immunological memory. However, regulatory T cells can also be detrimental to the host because they can be exploited by pathogens to facilitate pathogen persistence by suppressing anti-pathogen protective immune responses. This review has provided evidence from many studies that regulatory T cells can be both beneficial and detrimental to the host in response to the same pathogen. An explanation for these apparently contradictory findings might lie in the balance between regulatory and effector cells in different individuals, disease settings and experimental systems. The different outcomes of infection — persistence, resolution with excessive collateral damage, or resolution with limited immunopathology and development of immune memory — might be influenced by the ratio of regulatory T cells to effector T cells (FIG. 5). This hypothesis is supported by a recent report showing that re-infection of mice with L. major at a secondary site increased the number of regulatory T cells, resulting in disease reactivation at the primary site of infection; the equilibrium between effector and regulatory T cells controlled the efficiency of recall immune responses and disease reactivation100.
NATURE REVIEWS | IMMUNOLOGY
These studies have also helped to increase our understanding of the role of regulatory T cells in immune homeostasis and how they could be manipulated for the treatment of human diseases. In a normal healthy individual, the immune system must be capable of preventing the development of autoimmune diseases by suppressing immune responses to self-antigens. It must also be able to mount immune responses that control infections with a range of pathogenic organisms. Immune homeostasis is achieved through a careful balance between effector and suppressor responses, possibly through an appropriate frequency of TH1 cells, TH2 cells and CTLs versus natural and induced regulatory T cells (FIG. 5). The development of autoimmunity and allergy might, in part, arise from a deficit in regulatory T cells, whereas the development of cancer and chronic infections might be associated with an excess of these cells. Therefore, manipulation of this balance has opened up new approaches to therapy for a range of human diseases. The clonal expansion of regulatory T cells using strategies traditionally associated with the induction of tolerance has had some success in reducing symptoms of autoimmune disease in animal models26,101. However, this approach requires further development before it can be routinely applied to humans. Studies of animal tumour models have shown that altering the ratio of regulatory T cells to TH1 cells and CTLs can affect tumour survival; removal of CD4+CD25+ T cells increases anti-tumour immunity102, and therapy with pathogen-derived molecules that promote TH1-cell and CTL responses versus TR1-cell responses results in reduced versus increased tumour survival, respectively (A. Jarnicki, J. Lysaght, S. Todryk and K.H.G.M., unpublished observations). In mouse models of infectious disease, there is some evidence that removal of CD4+CD25+ regulatory T cells can help to resolve infection36,91. However, the application of this approach to humans will not be straightforward or without risk, and there are many unanswered questions. Will it be possible to deplete CD4+CD25+ TReg cells in vivo using monoclonal antibodies? Will the transient removal of CD4+CD25+ TReg cells be sufficient for resolution of infection and, if not, will the longer-term removal of these cells increase the risk of developing autoimmunity? Inhibition of pathogen-induced TR1 or TH3 cells or the cytokines they secrete is an alternative approach for the treatment of chronic infections. Studies with PBMCs from patients infected with HCV60 or with M. tuberculosis 25 have shown that antigen-specific IFN-γ secretion can be increased in vitro by the addition of IL-10-specific antibodies. Targeting IL-10 and TGF-β has the advantage of inhibiting the innate cytokines that induce regulatory T cells, as well as the products of these cells that mediate suppression. However, this is also not without risk because antiinflammatory cytokines can favour T cells that mediate pathogen clearance, but if uncontrolled, the same T cells can contribute to immunopathology. Therefore, the key to success with immunotherapeutic approaches will be to elicit the correct balance of effector/pathogenic and regulatory T cells.
VOLUME 4 | NOVEMBER 2004 | 8 5 3 ©2004 Nature Publishing Group
REVIEWS 1.
2.
3.
4. 5.
6.
7.
8.
9.
10. 11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
854
Artavanis-Tsakonas, K., Tongren, J. E. & Riley, E. M. The war between the malaria parasite and the immune system: immunity, immunoregulation and immunopathology. Clin. Exp. Immunol. 133, 145–152 (2003). von Herrath, M. G. & Harrison, L. C. Antigen-induced regulatory T cells in autoimmunity. Nature Rev. Immunol. 3, 223–232 (2003). Redpath, S., Ghazal, P. & Gascoigne, N. R. Hijacking and exploitation of IL-10 by intracellular pathogens. Trends Microbiol. 9, 86–92 (2001). Reed, S. G. TGF-β in infections and infectious diseases. Microbes Infect. 1, 1313–1325 (1999). Yazdanbakhsh, M., Kremsner, P. G. & van Ree, R. Allergy, parasites, and the hygiene hypothesis. Science 296, 490–494 (2002). Braun-Fahrlander, C. et al. Environmental exposure to endotoxin and its relation to asthma in school-age children. N. Engl. J. Med. 347, 869–877 (2002). Mills, K. H. G. & Boyd, A. in Topley and Wilsons’s Microbiology and Microbial Infections 10th edn (eds Kaufmann, S. H. & Steward, M.) (Edward Arnold Publishers Ltd, London, in press). McGuirk, P. & Mills, K. H. G. Pathogen-specific regulatory T cells provoke a shift in the TH1/TH2 paradigm in immunity to infectious diseases. Trends Immunol. 23, 450–455 (2002). Cozzo, C., Larkin, J. & Caton, A. J. Self-peptides drive the peripheral expansion of CD4+CD25+ regulatory T cells. J. Immunol. 171, 5678–5682 (2003). Bluestone, J. A. & Abbas, A. K. Natural versus adaptive regulatory T cells. Nature Rev. Immunol. 3, 253–257 (2003). Garba, M. L., Pilcher, C. D., Bingham, A. L., Eron, J. & Frelinger, J. A. HIV antigens can induce TGF-β1-producing immunoregulatory CD8+ T cells. J. Immunol. 168, 2247–2254 (2002). Haynes, L. M., Vanderlugt, C. L., Dal Canto, M. C., Melvold, R. W. & Miller, S. D. CD8+ T cells from Theiler’s virus-resistant BALB/cByJ mice downregulate pathogenic virus-specific CD4+ T cells. J. Neuroimmunol. 106, 43–52 (2000). Harrison, L. C., Dempsey-Collier, M., Kramer, D. R. & Takahashi, K. Aerosol insulin induces regulatory CD8 γδ T cells that prevent murine insulin-dependent diabetes. J. Exp. Med. 184, 2167–2174 (1996). Seo, N., Tokura, Y., Takigawa, M. & Egawa, K. Depletion of IL-10- and TGF-β-producing regulatory γδ T cells by administering a daunomycin-conjugated specific monoclonal antibody in early tumor lesions augments the activity of CTLs and NK cells. J. Immunol. 163, 242–249 (1999). Sonoda, K. H. et al. NK T cell-derived IL-10 is essential for the differentiation of antigen-specific T regulatory cells in systemic tolerance. J. Immunol. 166, 42–50 (2001). Sakaguchi, S., Sakaguchi, N., Asano, M., Itoh, M. & Toda, M. Immunologic self-tolerance maintained by activated T cells expressing IL-2 receptor α-chains (CD25). Breakdown of a single mechanism of self-tolerance causes various autoimmune diseases. J. Immunol. 155, 1151–1164 (1995). The first description of CD25 as a marker for TReg cells and the demonstration that these cells can suppress immune responses that mediate autoimmune diseases. Powrie, F., Carlino, J., Leach, M. W., Mauze, S. & Coffman, R. L. A critical role for transforming growth factor-β but not interleukin 4 in the suppression of T helper type 1mediated colitis by CD45RBlowCD4+ T cells. J. Exp. Med. 183, 2669–2674 (1996). Shimizu, J., Yamazaki, S., Takahashi, T., Ishida, Y. & Sakaguchi, S. Stimulation of CD25+CD4+ regulatory T cells through GITR breaks immunological self-tolerance. Nature Immunol. 3, 135–142 (2002). Fontenot, J. D., Gavin, M. A. & Rudensky, A. Y. Foxp3 programs the development and function of CD4+CD25+ regulatory T cells. Nature Immunol. 4, 330–336 (2003). Hori, S., Nomura, T. & Sakaguchi, S. Control of regulatory T cell development by the transcription factor Foxp3. Science 299, 1057–1061 (2003). Richman, L. K., Chiller, J. M., Brown, W. R., Hanson, D. G. & Vaz, N. M. Enterically induced immunologic tolerance. I. Induction of suppressor T lymphocytes by intragastric administration of soluble proteins. J. Immunol. 121, 2429–2434 (1978). Miller, A., Lider, O., Roberts, A. B., Sporn, M. B. & Weiner, H. L. Suppressor T cells generated by oral tolerization to myelin basic protein suppress both in vitro and in vivo immune responses by the release of transforming growth factor-β after antigen-specific triggering. Proc. Natl Acad. Sci. USA 89, 421–425 (1992). Hirai, K. et al. Suppression of cell-mediated immunity by street rabies virus infection. Microbiol. Immunol. 36, 1277–1290 (1992).
24. Mahanty, S. et al. High levels of spontaneous and parasite antigen-driven interleukin-10 production are associated with antigen-specific hyporesponsiveness in human lymphatic filariasis. J. Infect. Dis. 173, 769–773 (1996). 25. Boussiotis, V. A. et al. IL-10-producing T cells suppress immune responses in anergic tuberculosis patients. J. Clin. Invest. 105, 1317–1325 (2000). 26. Chen, Y., Kuchroo, V. K., Inobe, J., Hafler, D. A. & Weiner, H. L. Regulatory T cell clones induced by oral tolerance: suppression of autoimmune encephalomyelitis. Science 265, 1237–1240 (1994). 27. Groux, H. et al. A CD4+ T-cell subset inhibits antigen-specific T-cell responses and prevents colitis. Nature 389, 737–742 (1997). The first description of TR1-cell clones and their ability to prevent inflammatory disease. 28. McGuirk, P., McCann, C. & Mills, K. H. G. Pathogen-specific T regulatory 1 cells induced in the respiratory tract by a bacterial molecule that stimulates interleukin 10 production by dendritic cells: a novel strategy for evasion of protective T helper type 1 responses by Bordetella pertussis. J. Exp. Med. 195, 221–231 (2002). The first demonstration of TR1 cells specific for a bacterial antigen, their cloning from a mucosal site of infection and a role for IL-10 produced by DCs in their induction. 29. Belkaid, Y., Piccirillo, C. A., Mendez, S., Shevach, E. M. & Sacks, D. L. CD4+CD25+ regulatory T cells control Leishmania major persistence and immunity. Nature 420, 502–507 (2002). The first description of a role for CD4+CD25+ TReg cells in protective immunity to infection and in the development of immunological memory to a parasite. 30. Lundgren, A., Suri-Payer, E., Enarsson, K., Svennerholm, A. M. & Lundin, B. S. Helicobacter pylorispecific CD4+CD25high regulatory T cells suppress memory T-cell responses to H. pylori in infected individuals. Infect. Immun. 71, 1755–1762 (2003). 31. Vahlenkamp, T. W., Tompkins, M. B. & Tompkins, W. A. Feline immunodeficiency virus infection phenotypically and functionally activates immunosuppressive CD4+CD25+ T regulatory cells. J. Immunol. 172, 4752–4761 (2004). 32. Lavelle, E. C. et al. Cholera toxin promotes the induction of regulatory T cells specific for bystander antigens by modulating dendritic cell activation. J. Immunol. 171, 2384–2392 (2003). 33. Cottrez, F., Hurst, S. D., Coffman, R. L. & Groux, H. T regulatory cells 1 inhibit a TH2-specific response in vivo. J. Immunol. 165, 4848–4853 (2000). 34. Piccirillo, C. A. & Shevach, E. M. Control of CD8+ T cell activation by CD4+CD25+ immunoregulatory cells. J. Immunol. 167, 1137–1140 (2001). 35. Kursar, M. et al. Regulatory CD4+CD25+ T cells restrict memory CD8+ T cell responses. J. Exp. Med. 196, 1585–1592 (2002). 36. Suvas, S., Kumaraguru, U., Pack, C. D., Lee, S. & Rouse, B. T. CD4+CD25+ T cells regulate virus-specific primary and memory CD8+ T cell responses. J. Exp. Med. 198, 889–901 (2003). 37. Maloy, K. J. et al. CD4+CD25+ TR cells suppress innate immune pathology through cytokine-dependent mechanisms. J. Exp. Med. 197, 111–119 (2003). This paper describes the role of CD45RBhi TReg cells in the prevention of bacteria-induced inflammation in the intestine. 38. Kitani, A., Chua, K., Nakamura, K. & Strober, W. Activated self-MHC-reactive T cells have the cytokine phenotype of TH3/T regulatory cell 1 T cells. J. Immunol. 165, 691–702 (2000). 39. Gorelik, L., Constant, S. & Flavell, R. A. Mechanism of transforming growth factor β-induced inhibition of T helper type 1 differentiation. J. Exp. Med. 195, 1499–1505 (2002). 40. Moore, K. W., de Waal Malefyt, R., Coffman, R. L. & O’Garra, A. Interleukin-10 and the interleukin-10 receptor. Annu. Rev. Immunol. 19, 683–765 (2001). 41. Kitani, A. et al. Transforming growth factor (TGF)-β1producing regulatory T cells induce Smad-mediated interleukin 10 secretion that facilitates coordinated immunoregulatory activity and amelioration of TGF-β1mediated fibrosis. J. Exp. Med. 198, 1179–1188 (2003). 42. Grundstrom, S., Cederbom, L., Sundstedt, A., Scheipers, P. & Ivars, F. Superantigen-induced regulatory T cells display different suppressive functions in the presence or absence of natural CD4+CD25+ regulatory T cells in vivo. J. Immunol. 170, 5008–5017 (2003). 43. Read, S., Malmstrom, V. & Powrie, F. Cytotoxic T lymphocyteassociated antigen 4 plays an essential role in the function of CD25+CD4+ regulatory cells that control intestinal inflammation. J. Exp. Med. 192, 295–302 (2000).
| NOVEMBER 2004 | VOLUME 4
44. Nakamura, K., Kitani, A. & Strober, W. Cell contactdependent immunosuppression by CD4+CD25+ regulatory T cells is mediated by cell surface-bound transforming growth factor-β. J. Exp. Med. 194, 629–644 (2001). 45. Barthlott, T., Kassiotis, G. & Stockinger, B. T cell regulation as a side effect of homeostasis and competition. J. Exp. Med. 197, 451–460 (2003). 46. MacDonald, A. J. et al. CD4 T helper type 1 and regulatory T cells induced against the same epitopes on the core protein in hepatitis C virus-infected persons. J. Infect. Dis. 185, 720–727 (2002). 47. Doetze, A. et al. Antigen-specific cellular hyporesponsiveness in a chronic human helminth infection is mediated by TH3/TR1type cytokines IL-10 and transforming growth factor-β but not by a TH1 to TH2 shift. Int. Immunol. 12, 623–630 (2000). The first demonstration of parasite-specific TR1 and TH3 cells at the clonal level in humans. 48. Satoguina, J. et al. Antigen-specific T regulatory-1 cells are associated with immunosuppression in a chronic helminth infection (onchocerciasis). Microbes Infect. 4, 1291–1300 (2002). 49. Accapezzato, D. et al. Hepatic expansion of a virus-specific regulatory CD8+ T cell population in chronic hepatitis C virus infection. J. Clin. Invest. 113, 963–972 (2004). 50. Marshall, N. A., Vickers, M. A. & Barker, R. N. Regulatory T cells secreting IL-10 dominate the immune response to EBV latent membrane protein 1. J. Immunol. 170, 6183–6189 (2003). 51. Delgado, J. C. et al. Antigen-specific and persistent tuberculin anergy in a cohort of pulmonary tuberculosis patients from rural Cambodia. Proc. Natl Acad. Sci. USA 99, 7576–7581 (2002). 52. Gerosa, F. et al. CD4+ T cell clones producing both interferon-γ and interleukin-10 predominate in bronchoalveolar lavages of active pulmonary tuberculosis patients. Clin. Immunol. 92, 224–234 (1999). 53. Ostrowski, M. A. et al. Quantitative and qualitative assessment of human immunodeficiency virus type 1 (HIV-1)-specific CD4+ T cell immunity to gag in HIV-1infected individuals with differential disease progression: reciprocal interferon-γ and interleukin-10 responses. J. Infect. Dis. 184, 1268–1278 (2001). 54. Beilharz, M. W. et al. Timed ablation of regulatory CD4+ T cells can prevent murine AIDS progression. J. Immunol. 172, 4917–4925 (2004). 55. van der Kleij, D. et al. A novel host–parasite lipid cross-talk. Schistosomal lyso-phosphatidylserine activates Toll-like receptor 2 and affects immune polarization. J. Biol. Chem. 277, 48122–48129 (2002). 56. Kullberg, M. C. et al. Bacteria-triggered CD4+ T regulatory cells suppress Helicobacter hepaticus-induced colitis. J. Exp. Med. 196, 505–515 (2002). 57. Graca, L., Le Moine, A., Cobbold, S. P. & Waldmann, H. Dominant transplantation tolerance. Curr. Opin. Immunol. 15, 499–506 (2003). 58. Chen, C., Lee, W. H., Yun, P., Snow, P. & Liu, C. P. Induction of autoantigen-specific TH2 and TR1 regulatory T cells and modulation of autoimmune diabetes. J. Immunol. 171, 733–744 (2003). 59. Akdis, M. et al. Immune responses in healthy and allergic individuals are characterized by a fine balance between allergen-specific T regulatory 1 and T helper 2 cells. J. Exp. Med. 199, 1567–1575 (2004). 60. Brady, M. T., MacDonald, A. J., Rowan, A. G. & Mills, K. H. Hepatitis C virus non-structural protein 4 suppresses TH1 responses by stimulating IL-10 production from monocytes. Eur. J. Immunol. 33, 3448–3457 (2003). 61. Boyer, O. et al. CD4+CD25+ regulatory T-cell deficiency in patients with hepatitis C-mixed cryoglobulinemia vasculitis. Blood 103, 3428–3430 (2004). 62. Suvas, S., Azkur, A. K., Kim, B. S., Kumaraguru, U. & Rouse, B. T. CD4+CD25+ regulatory T cells control the severity of viral immunoinflammatory lesions. J. Immunol. 172, 4123–4132 (2004). This study shows that CD25+ TReg cells suppress CD4+ T-cell responses in vivo and limit inflammation during corneal infection with HSV. 63. Deckert, M. et al. Endogenous interleukin-10 is required for prevention of a hyperinflammatory intracerebral immune response in Listeria monocytogenes meningoencephalitis. Infect. Immun. 69, 4561–4571 (2001). 64. Sewnath, M. E. et al. IL-10-deficient mice demonstrate multiple organ failure and increased mortality during Escherichia coli peritonitis despite an accelerated bacterial clearance. J. Immunol. 166, 6323–6331 (2001). 65. Chen, W., Shu, D. & Chadwick, V. S. Helicobacter pylori infection: mechanism of colonization and functional dyspepsia. Reduced colonization of gastric mucosa by Helicobacter pylori in mice deficient in interleukin-10. J. Gastroenterol. Hepatol. 16, 377–383 (2001).
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS 66. Kullberg, M. C. et al. Helicobacter hepaticus-induced colitis in interleukin-10-deficient mice: cytokine requirements for the induction and maintenance of intestinal inflammation. Infect. Immun. 69, 4232–4241 (2001). 67. Higgins, S. C. et al. Toll-like receptor 4-mediated innate IL-10 activates antigen-specific regulatory T cells and confers resistance to Bordetella pertussis by inhibiting inflammatory pathology. J. Immunol. 171, 3119–3127 (2003). This paper describes a role for TLR4 in the induction of IL-10 production by innate immune cells and regulatory T cells and its role in preventing bacteriainduced lung inflammation. 68. Raghavan, S., Fredriksson, M., Svennerholm, A. M., Holmgren, J. & Suri-Payer, E. Absence of CD4+CD25+ regulatory T cells is associated with a loss of regulation leading to increased pathology in Helicobacter pyloriinfected mice. Clin. Exp. Immunol. 132, 393–400 (2003). 69. Hori, S., Carvalho, T. L. & Demengeot, J. CD25+CD4+ regulatory T cells suppress CD4+ T cell-mediated pulmonary hyperinflammation driven by Pneumocystis carinii in immunodeficient mice. Eur. J. Immunol. 32, 1282–1291 (2002). This paper shows that regulatory T cells can increase the pathogen load but protect the host by limiting lung inflammation and the wasting syndrome induced by P. carinii. 70. Montagnoli, C. et al. B7/CD28-dependent CD4+CD25+ regulatory T cells are essential components of the memoryprotective immunity to Candida albicans. J. Immunol. 169, 6298–6308 (2002). 71. Netea, M. G. et al. Toll-like receptor 2 suppresses immunity against Candida albicans through induction of IL-10 and regulatory T cells. J. Immunol. 172, 3712–3718 (2004). 72. May, J., Lell, B., Luty, A. J., Meyer, C. G. & Kremsner, P. G. Plasma interleukin-10: tumor necrosis factor (TNF)-α ratio is associated with TNF promoter variants and predicts malarial complications. J. Infect. Dis. 182, 1570–1573 (2000). 73. Othoro, C. et al. A low interleukin-10 tumor necrosis factor-α ratio is associated with malaria anemia in children residing in a holoendemic malaria region in western Kenya. J. Infect. Dis. 179, 279–282 (1999). 74. Li, C., Corraliza, I. & Langhorne, J. A defect in interleukin-10 leads to enhanced malarial disease in Plasmodium chabaudi chabaudi infection in mice. Infect. Immun. 67, 4435–4442 (1999). 75. Omer, F. M. & Riley, E. M. Transforming growth factor-β production is inversely correlated with severity of murine malaria infection. J. Exp. Med. 188, 39–48 (1998). 76. Omer, F. M., de Souza, J. B. & Riley, E. M. Differential induction of TGF-β regulates proinflammatory cytokine production and determines the outcome of lethal and nonlethal Plasmodium yoelii infections. J. Immunol. 171, 5430–5436 (2003). 77. Hesse, M. et al. The pathogenesis of schistosomiasis is controlled by cooperating IL-10-producing innate effector and regulatory T cells. J. Immunol. 172, 3157–3166 (2004). 78. Xu, D. et al. CD4+CD25+ regulatory T cells suppress differentiation and functions of TH1 and TH2 cells, Leishmania major infection and colitis in mice. J. Immunol. 170, 394–399 (2003). 79. Sugimoto, K. et al. Suppression of HCV-specific T cells without differential hierarchy demonstrated ex vivo in persistent HCV infection. Hepatology 38, 1437–1448 (2003).
80. Clerici, M. & Shearer, G. M. The TH1–TH2 hypothesis of HIV infection: new insights. Immunol. Today 15, 575–581 (1994). 81. Kinter, A. L. et al. CD25+CD4+ regulatory T cells from the peripheral blood of asymptomatic HIV-infected individuals regulate CD4+ and CD8+ HIV-specific T cell immune responses in vitro and are associated with favorable clinical markers of disease status. J. Exp. Med. 200, 331–343 (2004). 82. Aandahl, E. M., Michaelsson, J., Moretto, W. J., Hecht, F. M. & Nixon, D. F. Human CD4+CD25+ regulatory T cells control T-cell responses to human immunodeficiency virus and cytomegalovirus antigens. J. Virol. 78, 2454–2459 (2004). 83. Iwashiro, M. et al. Immunosuppression by CD4+ regulatory T cells induced by chronic retroviral infection. Proc. Natl Acad. Sci. USA 98, 9226–9230 (2001). 84. Dittmer, U. et al. Functional impairment of CD8+ T cells by regulatory T cells during persistent retroviral infection. Immunity 20, 293–303 (2004). This study describes a role for IL-10-secreting CD4+ regulatory T cells in suppressing the protective effect of CD8+ CTLs in a viral infection. 85. Turner, J. et al. In vivo IL-10 production reactivates chronic pulmonary tuberculosis in C57BL/6 mice. J. Immunol. 169, 6343–6351 (2002). 86. Jacobs, M., Brown, N., Allie, N., Gulert, R. & Ryffel, B. Increased resistance to mycobacterial infection in the absence of interleukin-10. Immunology 100, 494–501 (2000). 87. Sing, A., Roggenkamp, A., Geiger, A. M. & Heesemann, J. Yersinia enterocolitica evasion of the host innate immune response by V antigen-induced IL-10 production of macrophages is abrogated in IL-10-deficient mice. J. Immunol. 168, 1315–1321 (2002). 88. Ross, P. J., Lavelle, E. C., Mills, K. H. G. & Boyd, A. P. Adenylate cyclase toxin from Bordetella pertussis synergises with lipopolysaccaride to promote innate IL-10 production and enhance the induction of TH2 and regulatory T cells. Infect. Immun. 72, 1568–1579 (2003). 89. McGuirk, P., Mahon, B. P., Griffin, F. & Mills, K. H. G. Compartmentalization of T cell responses following respiratory infection with Bordetella pertussis: hyporesponsiveness of lung T cells is associated with modulated expression of the co-stimulatory molecule CD28. Eur. J. Immunol. 28, 153–163 (1998). 90. Kursar, M., Kohler, A., Kaufmann, S. H. & Mittrucker, H. W. Depletion of CD4+ T cells during immunization with nonviable Listeria monocytogenes causes enhanced CD8+ T cell-mediated protection against listeriosis. J. Immunol. 172, 3167–3172 (2004). 91. Hisaeda, H. et al. Escape of malaria parasites from host immunity requires CD4+ CD25+ regulatory T cells. Nature Med. 10, 29–30 (2004). This study shows that depletion of natural regulatory T cells allows the host immune response to clear a parasite infection. 92. Long, T. T., Nakazawa, S., Onizuka, S., Huaman, M. C. & Kanbara, H. Influence of CD4+CD25+ T cells on Plasmodium berghei NK65 infection in BALB/c mice. Int. J. Parasitol. 33, 175–183 (2003). 93. Noben-Trauth, N., Lira, R., Nagase, H., Paul, W. E. & Sacks, D. L. The relative contribution of IL-4 receptor signaling and IL-10 to susceptibility to Leishmania major. J. Immunol. 170, 5152–5158 (2003).
NATURE REVIEWS | IMMUNOLOGY
94. Sacks, D. & Noben-Trauth, N. The immunology of susceptibility and resistance to Leishmania major in mice. Nature Rev. Immunol. 2, 845–858 (2002). 95. Jonuleit, H., Schmitt, E., Schuler, G., Knop, J. & Enk, A. H. Induction of interleukin 10-producing, nonproliferating CD4+ T cells with regulatory properties by repetitive stimulation with allogeneic immature human dendritic cells. J. Exp. Med. 192, 1213–1222 (2000). 96. Martin, E., O’Sullivan, B., Low, P. & Thomas, R. Antigenspecific suppression of a primed immune response by dendritic cells mediated by regulatory T cells secreting interleukin-10. Immunity 18, 155–167 (2003). 97. von der Weid, T., Bulliard, C. & Schiffrin, E. J. Induction by a lactic acid bacterium of a population of CD4+ T cells with low proliferative capacity that produce transforming growth factor-β and interleukin-10. Clin. Diagn. Lab. Immunol. 8, 695–701 (2001). 98. Caramalho, I. et al. Regulatory T cells selectively express Toll-like receptors and are activated by lipopolysaccharide. J. Exp. Med. 197, 403–411 (2003). 99. Kemper, C. et al. Activation of human CD4+ cells with CD3 and CD46 induces a T-regulatory cell 1 phenotype. Nature 421, 388–392 (2003). 100. Mendez, S., Reckling, S. K., Piccirillo, C. A., Sacks, D. & Belkaid, Y. Role for CD4+CD25+ regulatory T cells in reactivation of persistent leishmaniasis and control of concomitant immunity. J. Exp. Med. 200, 201–210 (2004). 101. Bynoe, M. S., Evans, J. T., Viret, C. & Janeway, C. A. Jr. Epicutaneous immunization with autoantigenic peptides induces T suppressor cells that prevent experimental allergic encephalomyelitis. Immunity 19, 317–328 (2003). 102. Golgher, D., Jones, E., Powrie, F., Elliott, T. & Gallimore, A. Depletion of CD25+ regulatory cells uncovers immune responses to shared murine tumor rejection antigens. Eur. J. Immunol. 32, 3267–3275 (2002).
Acknowledgements I acknowledge support from Science Foundation Ireland, The Irish Health Research Board and Enterprise Ireland, and I am grateful to P. McGuirk and E. Lavelle for helpful discussions.
Competing interests statement The author declares competing financial interests: see Web version for details.
Online links DATABASES The following terms in this article are linked online to: Entrez Gene: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=gene CD4 | CD25 | CD28 | CD38 | CD40 | CD45 | CD46 | CD62L | CD86 | CD103 | CTLA4 | FOXP3 | GITR | ICAM1 | IFN-α | IFN-γ | IL-4 | IL-6 | IL-10 | IL-10R | IL-12 | IL-12R | IL-23 | IL-27 | LFA1 | RAG | SMAD4 | T-bet | TGF-β | TLR2 | TLR4 | TNF Infectious Disease Information: http://www.cdc.gov/ncidod/diseases/index.htm Bordetella pertussis | Brugia malayi | Candida albicans | EBV | Escherichia coli | HCV | HIV | Leishmania major | Listeria monocytogenes | malaria | Mycobacterium tuberculosis | Onchocerca volvulus | Pneumocystis carinii | rabies virus | Schistosoma mansoni | Yersinia enterocolitica Access to this interactive links box is free online.
VOLUME 4 | NOVEMBER 2004 | 8 5 5 ©2004 Nature Publishing Group
REVIEWS
STRATEGIES TO ENHANCE T-CELL RECONSTITUTION IN IMMUNOCOMPROMISED PATIENTS Marcel R. M. van den Brink*, Önder Alpdogan* and Richard L. Boyd‡ Abstract | Immune deficiency, together with its associated risks such as infections, is becoming an increasingly important clinical problem owing to the ageing of the general population and the increasing number of patients with HIV/AIDS, malignancies (especially those treated with intensive chemotherapy or radiotherapy) or transplants (of either solid organs or haematopoietic stem cells). Of all immune cells, T cells are the most often affected, leading to a prolonged deficiency of T cells, which has important clinical consequences. Accordingly, strategies to improve the recovery and function of T cells, as we discuss here, should have a direct impact on reducing the morbidity and mortality of many patients and should increase the efficacy of therapeutic and prophylactic vaccinations against microbial pathogens or tumours.
*Departments of Medicine and Immunology, Box 111Kettering 406D, Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, New York, New York 10021, USA. ‡ Department of Pathology and Immunology, Central and Eastern Clinical School, Monash University, Clayton, Victoria 3181, Australia. Correspondence to M.R.M.B. e-mail:
[email protected] doi:10.1038/nri1484
856
Immune deficiency — a decrease in the number or function of immune cells — leads to a significant increase in the incidence and severity of infections, the occurrence and relapse of cancers, and the failure of immunotherapies, including vaccination. Apart from genetic causes (such as SEVERE COMBINED IMMUNODEFICIENCY, SCID) and autoimmune diseases, immunodeficiency is commonly associated with ageing but can also arise directly as a result of infections that target the immune system — most notably, infection with HIV — and as a consequence of common cancer treatments, such as myeloablative chemotherapy and radiation. Immunodeficiency also occurs through treatment with drugs that are frequently used to prevent rejection of foreign cell, tissue or organ transplants. In these situations, T cells are more suppressed than other immune cells and are slower to recover. T-cell deficiency is more pronounced in adults as a result of the markedly reduced function of the thymus, which undergoes atrophy early in life, particularly from the onset of puberty, lapsing to less than 10% of its maximum size by the early 20s1. In this review, we discuss strategies that could enhance the reconstitution of T cells after an ALLOGENEIC haematopoietic stem cell (HSC) transplant (HSCT) — which is associated with a severe deficiency in Tcells, as
| NOVEMBER 2004 | VOLUME 4
discussed later and in BOX 1 — but these strategies have equal potential to overcome T-cell deficiencies in general. The approaches include those that target lymphoid progenitors and promote thymopoiesis, and those that use ex vivo culture systems, hormones, growth factors, cytokines or co-stimulatory molecules (FIG. 1). We have focused on strategies that are currently being studied in clinical trials or have realistic potential for clinical use in the foreseeable future. Furthermore, in patients with T-cell deficiencies, adoptive T-cell therapy has resulted in the successful treatment of Epstein–Barr-virus-induced lymphoproliferative disease and lymphoma, as well as infection with cytomegalovirus (CMV), and it is now being studied as a promising therapeutic strategy for the prophylaxis and treatment of a variety of infections and malignancies. However, the use of adoptive cell therapy to treat specific infections or tumours is beyond the scope of this article and is reviewed in REFS 2,3. Allogeneic HSC transplantation
Allogeneic HSC transplantation is a potentially curative therapy for a variety of life-threatening diseases of lymphohaematopoietic cells and tissues, including malignancies and diseases characterized by defective
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS
Box 1 | Immune reconstitution after allogeneic haematopoietic stem-cell transplantation
SEVERE COMBINED IMMUNODEFICIENCY
(SCID). Humans or mice with this rare genetic disorder lack functional T and B cells owing to a mutation in a gene that is involved in T-cell and/or B-cell development; consequently, they suffer from recurrent infections. Several forms of SCID have been described, including mutations in the common cytokinereceptor γ-chain of several interleukin receptors, Janus activated kinase 3 (JAK3) and adenosine deaminase. ALLOGENEIC
Allogeneic tissues or cells are genetically different from the host and can elicit an immune response when transplanted, resulting in rejection or graftversus-host disease. GRAFT-VERSUS-HOST DISEASE
(GVHD). Tissue damage in a recipient of allogeneic transplanted tissue (usually bone marrow) that results from the activity of donor cytotoxic T cells that recognize the tissue of the recipient as foreign. GVHD varies markedly in severity, but it can be life threatening in severe cases and, in particular, affects the intestines, liver and skin. CONGENIC
An animal strain that is genetically identical to another strain except for one or more allelic differences that do not result in an antigen that can elicit an immunological response when tissue is transferred or transplanted from one strain to another.
The post-transplant reconstitution of T cells in recipients of an allogeneic haematopoietic stem-cell transplant (HSCT) involves the clonal expansion of mature donor lymphocytes (both alloreactive and non-alloreactive) that were infused with the allograft, as well as residual host lymphocytes resistant to the conditioning regimen, and thymic-dependent or possibly thymic-independent de novo-generated donor lymphocytes (FIG. 2). In young hosts, the thymus can support de novo generation of T cells from donor haematopoietic precursor cells. However, the contribution of the thymus to post-transplant T-cell recovery depends on the following: the extent of thymic damage from conditioning of the recipient with chemotherapy and/or radiotherapy, the degree of age-associated thymic involution, the engraftment of donor-derived haematopoietic precursor cells, the occurrence of graft-versus-host disease (GVHD) and the extent of drug-induced immunosuppression from antibiotics or prophylaxis for GVHD. Studies using thymectomized mice that have received an HSCT have shown de novo generation of T cells (commonly of non-classical T cells, such as CD4–CD8– or CD8αα+ T cells) in extrathymic sites, including the mucosa-associated lymphoid tissues in the gut115,116; however, their contribution to T-cell recovery in euthymic recipients of an HSCT was limited. Studies of patients who received a T-cell-depleted HSCT have indicated that the deficient T-cell immunity in the first year after transplantation is not only due to an insufficient number of T cells but also due to an insufficient T-cell repertoire117. The restoration of the repertoire probably depends on the repopulation of thymus-derived naive T cells. Strategies to restore thymic function are therefore a principal aim of immune reconstitution. In addition to the importance of intact thymic precursors (BOX 2) and thymic function for post-transplant T-cell reconstitution, de novo-generated T cells (thymic emigrants), as well as mature T cells that are transferred in the allograft, require extrathymic support for their survival and proliferation. The factors that determine the homeostasis of peripheral T cells are beginning to be defined. For example, the process of homeostatic proliferation (BOX 3), which is particularly relevant in lymphopenic states and involves the cytokines interleukin-7 (IL-7) and IL-15, has been recognized as an important contributor to post-transplant T-cell reconstitution84,92,93,106. In addition, survival signals have a crucial role in the reconstitution and homeostasis of T cells after transplantation. Adult recipients of an HSCT have a 5–10-fold increase in the number of apoptotic peripheral T cells compared with healthy controls, and this is associated with GVHD, HLA disparity between the donor and the host, time after transplant, and the expression of FAS (CD95) and B-cell lymphoma 2 (BCL-2)118,119.
lymphohaematopoiesis. Myeloablative and nonmyeloablative conditioning regimens, which consist of radiation, chemotherapy and/or immunosuppressive drugs to enable engraftment of the donor HSCs, specifically target T cells to prevent graft rejection by T cells of the host and GRAFT-VERSUS-HOST DISEASE (GVHD) by T cells of the donor. In contrast to the relatively early recovery of innate immunity (mediated by myeloid and natural killer (NK) cells), all recipients (but especially adult recipients) of an allogeneic HSCT have post-transplant deficiencies in their reconstitution of B cells and T cells, and these can exceed the period of lymphocytopaenia as a result of delays in the recovery of function4,5. Although children commonly recover T-cell-based immunity within 6 months of an HSCT following chemotherapy, adults can require years and, even then, rarely re-establish a fully competent T-cell repertoire4,6. This prolonged post-transplant lymphoid deficiency (in particular of the T-cell lineage) is associated with an increased risk of infections4,7, relapse of malignancy8 and development of secondary malignancies9, and it reduces the efficacy of immunotherapeutic strategies, such as vaccination against microorganisms or tumours. The risk of opportunistic infections in the post-transplant period directly correlates with the recovery of T cells (especially CD4+ T cells)4,5,7. Particularly in the first weeks after transplantation, the lymphoid system of recipients of an allogeneic HSCT contains shifting populations of donor and host T cells, which include the following: de novo-generated donor T cells, which originate from donor haematopoietic
NATURE REVIEWS | IMMUNOLOGY
precursors and are produced in the thymus or possibly in extrathymic sites such as the intestinal mucosa; non-alloreactive T cells, which derive from mature donor T cells in the allograft and can undergo homeostatic proliferation in the lymphopenic host; alloreactive T cells, which are transferred in the allograft and cause GVHD; and residual host T cells, which have survived the conditioning regimen and can reject the allograft (BOX 1; FIG. 2). Therefore, any strategy that affects the T cells of transplant recipients will need to be tested for its potential risks of enhancing GVHD or graft rejection and its potential benefits of promoting thymopoiesis, homeostatic proliferation and T-cell survival. Adoptive transfer of lymphoid progenitor cells
The recent identification of common lymphoid progenitors (CLPs) in adult bone marrow (BOX 2; FIG. 3) has allowed the development of adoptive transfer of donor lymphoid precursors to recipients of an allogeneic HSCT as a strategy to expedite and enhance de novo generation of donor T cells and to promote recovery of T cells. This strategy operates under the assumption that the addition of committed lymphoid precursors to the graft will result in an accelerated (but transient) recovery of the thymus before lymphoid precursors derived from the pluripotent donor HSCs begin their continuing repopulation of the thymus. Adoptive transfer of committed progenitors has been successfully applied in mouse models to enhance myeloid reconstitution through the CONGENIC transplantation of common myeloid progenitors (CMPs)
VOLUME 4 | NOVEMBER 2004 | 8 5 7 ©2004 Nature Publishing Group
REVIEWS
Thymopoiesis • Thymic graft • TEC graft • IL-7 • IL-12
• Sex-steroid inhibition • Growth hormone and IGF1 • KGF
Peripheral T cells • IL-7 • IL-15 • Superagonistic CD28specific antibodies • Growth hormone and IGF1 • Oncostatin M
Bone marrow • Lymphoid precursors • Notch ligand • IL-7 • Sex-steroid inhibition Ex vivo culture • Notch ligand • Thymic organoid • Thymic organ culture
Mucosal T cells • IL-7 • IL-15 • KGF
Figure 1 | Strategies to enhance T-cell reconstitution after allogeneic bone-marrow transplantation. After transplantation, various immunostimulatory strategies are used to improve reconstitution of T cells. The regions of the body and/or cell populations that are targeted by these strategies are depicted, and the strategies — which include the administration of various cells, tissues or factors, and the inhibition of sex steroids — are indicated. These strategies function to increase the number of lymphoid precursors in the bone marrow, stimulate thymopoiesis, and/or enhance the production of peripheral T cells and/or mucosal T cells. Some strategies promote T-cell reconstitution by affecting more than one of these areas. IGF1, insulinlike growth factor 1; IL, interleukin; KGF, keratinocyte growth factor; TEC, thymic epithelial cell.
DIGEORGE SYNDROME
A syndrome characterized by cardiac malformations, facial anomalies and hypoplasia of the parathyroid gland and thymus. Most cases are the result of a deletion of the chromosomal region 22q11.2. Mice deficient in the homeobox A3 protein (HOXA3) develop a phenotype similar to patients with DiGeorge syndrome. CENTRAL TOLERANCE
Lack of self-responsiveness that occurs as lymphoid cells develop. It is associated with the deletion of autoreactive clones. For T cells, this occurs in the thymus. MIXED LYMPHOCYTE REACTION
A tissue-culture technique that is used for the in vitro testing of the proliferative response of T cells from one individual to lymphocytes from another individual. NUDE MICE
Mice with a mutation in the forkhead box N1 gene (Foxn1), which results in hairlessness, defective formation of the thymus and a lack of mature T cells.
858
and granulocyte/monocyte progenitors (GMPs) in recipients of an HSCT, which resulted in increased protection against infection with Aspergillus fumigatus or Pseudomonas aeruginosa10. The addition of CLPs to the allograft increased the protection of lethally irradiated recipients of an HSCT against infection with murine CMV (MCMV), which requires a complex immune response involving CD4+ and CD8+ T cells, NK cells and B cells 11. However, accelerated reconstitution was mainly achieved in the B-cell lineage, and only minor effects on thymic cellularity and peripheral T-cell reconstitution were observed. Moreover, thymectomized recipients of an HSCT plus CLPs also had increased protection against infection with MCMV, which might be explained by an increase in maturation of extrathymic CD8+ (but not CD4+) T cells, as well as early reconstitution of NK cells. Adoptive transfer of CLPs did not cause GVHD in allogeneic recipients, indicating that CLP-derived T cells with alloreactive potential were adequately deleted or functionally downregulated in the thymus and/or periphery. The relatively modest effects of the transfer of CLPs on T-cell reconstitution probably result from the commitment of most CLPs to the B-cell lineage. Therefore, adoptive-transfer studies using precursors that preferentially commit to the T-cell lineage, such as the CD62L+ thymic precursor (BOX 2), are expected to result in stronger effects on T-cell recovery. Enhancing thymopoiesis
Thymic grafts. The thymus is the main and most efficient (if not the only) site of T-cell lymphopoiesis (FIG. 3), and this has resulted in several studies using transplantation of thymic grafts to enhance T-cell recovery. Thymus transplants have been used in both animal models and children with SCID or DIGEORGE SYNDROME to
| NOVEMBER 2004 | VOLUME 4
establish functional lymphocyte responses or as a method for the induction of CENTRAL TOLERANCE following HSC or solid-organ transplantation. Early human studies in which human fetal thymic tissue or cultured thymic epithelium were transplanted into recipients of an allogeneic HSCT were unsuccessful, but these might have been compromised by the immunosuppressive therapy that these patients received to prevent GVHD12. More-recent studies, using transplantation of cultured thymic fragments (in the absence of an HSCT) to patients with DiGeorge syndrome, have been more successful13,14. Thymocyte maturation and a normal thymic microenvironment were observed, and T-cell responses to mitogens and antigens were evident13,14. The T-cell receptor (TCR) repertoire of these patients was initially oligoclonal and then progressed to being polyclonal, allowing for adequate immune responses to a wider variety of antigens14. MIXED LYMPHOCYTE REACTIONS showed tolerance to donor antigens, and recipients had normal antibody titres after immunization with tetanus toxoid or Pneumovax (the polyvalent pneumococcal vaccine)14. These results in human patients were supported by similar results obtained using a mouse model of DiGeorge syndrome (NUDE MICE). Nude mice transplanted with cultured thymic fragments (both congenic and allogeneic) survived and showed relatively normal lymphopoiesis15. Antibody responses normalized, and T-cell proliferation and cytotoxicity increased from 10% of the level of wild-type mice to 100%. Third-party skin grafts were rejected, and second-party grafts (from the thymic donor) were accepted or rejected slowly in both nude mice and mice subjected to total body irradiation15,16. Thymic transplants have also been relatively successful in large animal models. Several research groups have shown the acceptance, survival and function of
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS immunocompetent thymic grafts in miniature swine17–19, which were rendered tolerant to renal allografts syngeneic to the grafted thymus18. In addition, transplantation of composite organs, such as thymokidneys (a kidney with vascularized autologous thymic tissue under the capsule) and thymohearts, resulted in improved survival of secondary grafts. That is, transplantation of a thymus of donor origin together with a solid-organ transplant, such as a kidney or heart, increases the chance of graft acceptance17,19. These studies highlight that successful long-term acceptance of tissue grafts (in this case, a kidney or heart) requires a functioning thymus (donor derived or at least containing donor antigen-presenting cells, APCs) that can induce negative selection and/or the development of regulatory T cells. Although thymic grafts have been successful in promoting T-cell recovery and tolerance in preclinical and clinical studies, this strategy is limited by the availability of thymic tissue.
recognize a population of primordial TECs was first proposed by Blackburn et al.21 after they showed that this antibody stained the thymic remnant of nude mice. In separate studies, Gill et al.22 and Bennett et al.23 sorted embryonic TECs (from day 15 and day 12.5 embryos, respectively) into two groups: CD45–MTS24– MHC class II+ and CD45–MTS24+MHC class II+. Following in vitro reaggregation of a small number (as few as 500) of these cells, cells from each group were transplanted under the kidney capsule of recipient mice. After 3 weeks23 and 8 weeks22, capsulated, vascularized, ectopic thymi were present in those mice transplanted with MTS24+ TECs but not in those transplanted with MTS24– TECs. In these grafts, thymocyte development — as defined by the expression of CD4 and CD8, as well as CD25 and CD44 — seemed to be normal22,23. Immunohistological analysis showed that the thymi that were produced had a normal architecture and that all of the main stromal-cell components were present22,23. Importantly, thymi grafted into nude mice could produce peripheral T cells. Current strategies are aimed at more precisely defining the nature of TEC progenitors in mice as a basis for identifying the equivalent population of cells in the human thymus. The rapidly expanding horizons of stem-cell research might also enable the direct derivation of TEC progenitors from primitive adult or embryonic stem cells.
Thymic epithelial progenitor cells. The monoclonal antibody MTS24, which was raised against membrane preparations of mouse thymic stroma20, recognizes a glycoprotein that is differentially expressed throughout embryogenesis and adult life by mouse thymic epithelial cells (TECs). In the adult thymus, only isolated cells in the medulla are recognized by MTS24 (that is, MTS24+), but during early embryogenesis, the entire epithelium is MTS24+. Downregulation of expression of the glycoprotein in the thymus is coincident with the appearance of T cells. The idea that MTS24 might
Ex vivo culture systems
Several ex vivo culture systems have been developed to generate T cells from haematopoietic precursors: precursors are introduced into either a mouse fetal
Allograft
Host Extrathymic sites ?
Haematopoietic precursor
Bone marrow
T-cell precursor
?
Thymus
Naive T cells T-cell immunity and homeostasis
Non-alloreactive T cells
Homeostatic T-cell clonal expansion
T cells
Graft-versushost disease Alloreactive T cells
Alloreactive T-cell clonal expansion Graft rejection
Residual host T cells
Alloreactive T-cell clonal expansion
Figure 2 | T-cell reconstitution after allogeneic haematopoietic stem-cell transplantation. Five different T-cell populations might be present in recipients of an allogeneic haematopoietic stem-cell transplant (HSCT). Four of these are donor derived: thymic-dependent, newly generated T cells; non-alloreactive mature T cells that are infused with the allograft; alloreactive T cells that are also transferred with the allograft; and, possibly, a small population of donor T cells that are newly generated in extrathymic sites. The fifth population consists of residual host T cells that have survived the conditioning regimen (radiotherapy and/or chemotherapy) of the HSCT and can cause graft rejection.
NATURE REVIEWS | IMMUNOLOGY
VOLUME 4 | NOVEMBER 2004 | 8 5 9 ©2004 Nature Publishing Group
REVIEWS
Box 2 | Early T-cell lineage progenitors Sustained thymopoiesis requires continuous seeding of the thymus with bone-marrow progenitors120. In adult mice, lymphoid precursors enter the thymus periodically through the postcapillary venules in the corticomedullary junction121,122. The search for the elusive T-cell lineage progenitor cells has resulted in several candidates (reviewed in REF. 123).
Common lymphoid progenitors Common lymphoid progenitors (CLPs) are defined as being lineage (Lin)–CD44+cKITlowSCA1lowTHY1lowAA4+ interleukin-7 receptor α-chain (IL-7Rα)+ fms-related tyrosine kinase 3 (FLT3)+, and they constitute 0.02% of adult bonemarrow cells124. These cells were shown to have short-term (that is, early, rapid but not sustainable) reconstitution capability for T cells, B cells, natural killer (NK) cells, natural killer T (NKT) cells and dendritic cells (DCs), but they have a greater potential for lymphopoiesis of B cells than T cells. Also, they have not been isolated from the adult thymus. However, a cell population known as CLP-2 has been identified; these cells can be derived in short-term culture from CLPs (now known as CLP-1s)125. CLP-2s are bipotent precursors of T cells and B cells, but in contrast to CLP-1s, they are cKIT–B220+ and can be detected in the thymus.
Early T-cell lineage progenitors Early T-cell lineage progenitors (ETPs) are defined as being Lin–CD25–CD44+cKIThi SCA1hiIL-7Rαlow/–, and they constitute 0.01% of all thymocytes126. Within 3 weeks of being transferred to the thymus, they can undergo a 20,000–50,000-fold expansion. These ETPs can develop into T cells, B cells, NK cells, NKT cells and DCs, and to a lesser extent into myeloid cells. ETPs (but not CLPs) have weak myeloid differentiation potential, whereas CLPs have a stronger B-cell differentiation potential. After intrathymic injection, ETPs produce more double-positive thymocytes for a longer period of time than do CLPs. Ikaros-deficient mice (which do not produce B cells but have normal T-cell development) have ETPs but lack CLP-1 and CLP-2 populations, which indicates that ETPs can develop independently from CLPs. These data indicate that there are two progenitor populations that exist in parallel, both with B-cell and T-cell potential. This raises the question of whether a common progenitor of ETPs and CLPs exists in the bone marrow. CLPs arise from Lin–cKIThiSCA1hiFLT3+ bone-marrow progenitors, and this process requires FLT3 ligand127. However, the existence of a common progenitor for both ETPs and CLPs has not been shown, although two recently identified early lymphoid progenitors — recombination-activating gene 1 (RAG1)+Lin–cKIThiSCA1hiIL-7Rα–FLT3+ cells128 and Lin–cKIThiSCA1hiTHY1–CD62L+ cells129 — would be good candidates.
DOUBLE-NEGATIVE 1 (DN1) TO DN2 TRANSITION
Thymic precursors at the DN1 (CD3–CD4–CD8–CD25–CD44+) stage lose the ability to generate B cells, natural killer cells and dendritic cells after their transition to DN2 (CD3–CD4– CD8–CD25+CD44+) thymocytes.
860
thymic organ culture24 or a three-dimensional matrix (a thymic organoid)25, and then co-cultured with monolayers of thymic stroma26, peripheral-blood mononuclearcell feeder layers27 or combinations of growth factors and cytokines28. T cells that are generated in human culture systems could be administered to patients in addition to donor HSCs to expedite post-transplant T-cell recovery. However, most of these in vitro T-cell generation systems are difficult to establish, have a variable outcome and yield a small number of mature T cells. There are also uncertainties regarding whether positive and negative selection can occur appropriately in these culture systems. This makes the clinical application of these systems problematic, at least in the near future, until considerable improvements are made. One recently developed promising ex vivo culture system involves the co-culture of haematopoietic precursor cells with a Notch ligand (with or without a bone-marrow stromal cell line), resulting in large populations of thymic precursors and mature T cells.
| NOVEMBER 2004 | VOLUME 4
Notch ligand. Signalling through Notch is involved in various cell-fate decisions during the development of a multicellular organism, including survival, proliferation, lineage commitment and tissue architecture. In mammals, four members of the Notch family (the receptors Notch-1, -2, -3 and -4) and five ligands (Jagged-1 and -2, and Delta-like-1, -3 and -4) have been described. Notch-1 is essential for T-cell lineage commitment: the inhibition of Notch-1 results in a block in thymocyte differentiation at the DOUBLE-NEGATIVE 1 (DN1) TO DN2 TRANSITION and the accumulation of B cells in the thymus, whereas overexpression of constitutively active Notch-1 in haematopoietic progenitors inhibits B-cell development and promotes T-cell development up to the double-positive (DP) stage in the bone marrow. These and other studies indicate that Notch-1 is important for instructing bone-marrow-derived lymphoid precursors to select a T-cell versus B-cell fate in the thymus, as well as for promoting T-cell differentiation (reviewed in REF. 29). Consistent with this, Notch ligands are highly expressed by the thymus and less so by the bone marrow and fetal liver30. Several recent studies have shown that the Notch ligand Delta-like-1 can induce T-cell development from haematopoietic precursors in vitro 31,32. Incubation of mouse bone marrow with both the extracellular domain of Delta-like-1 fused to the Fc portion of human IgG and growth factors — stem-cell factor (SCF), interleukin-6 (IL-6), IL-11 and FLT3 (fms-related tyrosine kinase 3) ligand — inhibited myeloid differentiation and led to an increase in the number of precursors with short-term lymphoid and myeloid repopulation potential32. The addition of IL-7 further enhanced early T-cell development. Overexpression of Delta-like-1 by the OP9 bonemarrow stromal cell line (known as OP9-DL1 cells) — which can support haematolymphopoiesis from embryonic stem cells, and early haematopoiesis and B-cell lymphopoiesis from HSCs — allowed these cells to support the complete differentiation of fetal liver stem cells to mature CD8+ αβ and γδ T cells31. At present, the OP9-DL1 culture system has two drawbacks: it does not support the positive selection of functional CD4+ T cells or natural killer T (NKT) cells, because OP9-DL1 cells do not express MHC class II or CD1d; and defective negative selection of self-reactive T-cell clones could occur, because OP9-DL1 cells probably have a limited capacity to present self-antigens to developing T cells33. Additional genetic engineering of the OP9 stromal cell line to express MHC class II and CD1d molecules should further optimize this ex vivo culture system, which has potential for clinical use in recipients of an HSCT. Cultures could also be depleted of fully mature T cells before infusion so that only T-cell precursors are transferred, and these would then undergo positive and negative selection in the thymus of the recipient. Hormones and growth factors
Sex-steroids. The progressive loss of cell-mediated immunity during ageing can mostly be attributed to agerelated thymic atrophy34, which consists of a decrease in
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS
SCF
SCF
Lin– cKIThi SCA1hi THY1– CD62L+
CD62L+ thymic precursor
ETP
cKIThi SCA1hi CD44– IL-7Rα–
FLT3 ligand CLP-1 Ikaros?
Early lymphoid precursor?
cKITlow SCA1low THY1low CD44+ IL-7Rα+ FLT3+ AA4+
B-cell precursor
IL-7
T-cell development
IL-7
Notch-1?
Lin– cKIThi SCA1hi FLT3+
?
Notch ligand
?
CLP-2
cKIT– SCA1low THY1low CD44+ IL-7Rα+ FLT3+ AA4+ B220+
Thymus
B-cell development Spleen and lymph nodes
Bone marrow
Figure 3 | T-cell precursors in mice. The relationship between early lymphoid precursors, thymic precursors and common lymphoid progenitors (CLPs) is not yet clearly understood. Here, known and potential precursors of mouse T cells are shown, together with the cytokines and other factors that are important for early T-cell development. The phenotypic features of each T-cell precursor are also indicated. Interestingly, in contrast to mice, interleukin-7 (IL-7) does not seem to be involved in B-cell development in non-human primates and humans. ETP, Early T-cell lineage progenitor; FLT3, fms-related tyrosine kinase 3; IL-7Rα, IL-7 receptor α-chain; Lin, lineage; RAG1, recombination-activating gene 1; SCF, stem-cell factor.
thymic cellularity, as well as a marked alteration of the thymic microenvironment35. Thymic atrophy leads to a decrease in the number of recent thymic emigrants36. The ratio of naive T cells to memory T cells is decreased in the peripheral lymphoid tissues, and the TCR repertoire is restricted for both CD4+ and CD8+ T-cell subsets, resulting in diminished peripheral T-cell responses (reviewed in REF. 37). The main cause of thymic involution is thought to be the increased production of sex steroids (androgens, oestrogen and progesterone) after puberty, and it is well established in rodents that the ablation of sex steroids reverses age-related thymic atrophy35,38. Both surgical and chemical castration (using luteinizing-hormonereleasing hormone, LHRH) of old mice and rats lead to an increase in thymic cellularity and reformation of the thymic architecture35,38–40. Castration also leads to an increase in the number of peripheral T cells40. Castrated mice show increased reactivity to sheep red blood cells and enhanced rejection of skin grafts41. These increased responses were reversed by androgen treatment41. When castration and thymectomy were carried out together, rejection of skin grafts did not occur, confirming that the effects of sex steroids are linked to thymic output41. Olsen et al.42 have shown that functional androgen receptors at the cell surface of TECs, but not thymocytes, are essential for age-related thymic involution. From a clinical perspective, it is important that the decreased levels of sex steroids resulting from surgical castration can also be achieved chemically through the
NATURE REVIEWS | IMMUNOLOGY
administration of LHRH analogues, which are commonly used for the treatment of sex-steroid-exacerbated conditions, such as prostate cancer, breast cancer and endometriosis. Preliminary data from a phase II trial of patients with haematological malignancies who received an HSCT following myeloablative chemotherapy have revealed evidence of thymic rejuvenation in those treated with an LHRH agonist that is commonly used for the treatment of prostate cancer, as indicated by increased levels of naive CD4+ T cells (R.L.B., unpublished observations). Growth hormone and insulin-like growth factor 1. Apart from their anabolic effects as hormones involved in the regulation of metabolism, growth hormone and insulinlike growth factor 1 (IGF1) can enhance haematopoiesis, thymopoiesis, and T-cell and B-cell function43. The mechanisms responsible for the effects of growth hormone and IGF1 on T-cell development and function are poorly understood and could involve direct effects on T cells and their precursors, as well as stimulation of other cells (such as APCs or stromal cells) that can support T cells (reviewed in REF. 44). The main source of growth hormone is the anterior pituitary gland. But growth hormone is produced by many cells, and its receptor is also expressed by many cells (including haematopoietic cells) 45. Activation of the growthhormone receptor results in downstream activation of JAK2 (Janus activated kinase 2), STAT1 (signal transducer and activator of transcription 1), STAT3 and
VOLUME 4 | NOVEMBER 2004 | 8 6 1 ©2004 Nature Publishing Group
REVIEWS
ACTIVATION-INDUCED CELL DEATH
(AICD). The apoptotic cell death of activated lymphocytes. It ensures the rapid elimination of effector cells after their antigen-dependent clonal expansion. Defects in AICD result in lymphoproliferative diseases that are associated with autoimmune disorders.
862
STAT5 (REF. 45). Growth hormone mediates most of its effects on metabolism and haematopoiesis either directly or indirectly through the induction of IGF1. IGF1 can be secreted by haematopoietic cells, bone-marrow stromal cells and TECs, and IGF1 receptors are expressed by thymocytes, T cells, B cells, NK cells, monocytes and bonemarrow stromal cells46. IGF1 receptor expression is upregulated by T cells after activation by engagement of the TCR and CD28 (REF. 47). Under conditions of stress, Snell–Bagg dwarf mice, which have abnormal anterior-pituitary function, suffer from defects in T-cell immunity, including thymic hypoplasia and a decreased number of peripheral CD4+ T cells, which can be reversed by the administration of growth hormone48. Moreover, administration of IGF1 to 9-month-old mice was shown to promote engraftment of the thymus by lymphoid precursors and to increase thymic cellularity43. However, mice deficient in growth hormone or IGF1, or Snell–Bagg mice under non-stress conditions, have no defects in lymphoid development or function, which indicates that neither growth hormone nor IGF1 are required for normal lymphopoiesis and lymphoid homeostasis49. Both growth hormone and IGF1 can promote the survival and function of peripheral T cells, as well as increase the function of B cells, NK cells and macrophages44. For example, growth hormone can potentiate the antigen-specific proliferative and cytokine responses of human T-cell clones45. IGF1 and growth hormone seem to have an anti-apoptotic effect on peripheral T cells, because inhibition of the IGF1 receptor results in increased ACTIVATION-INDUCED CELL DEATH and FAS (CD95)mediated apoptosis, and both IGF1 and growth hormone can also partially inhibit dexamethasone-induced apoptosis of CD4+ T cells50. Studies using human cordblood T cells showed that administration of IGF1 could not only decrease FAS-dependent and FAS-independent apoptosis, but also enhance T-cell proliferation. This in vitro proliferative effect on the T-cell response to mitogens was confirmed in 9-month-old mice treated with IGF1 (REF. 43). Administration of growth hormone to normal adult mice increases the numbers of haematopoietic precursors in the spleen and bone marrow, and in recipients of bone-marrow transplants, it promotes multi-lineage reconstitution45,51. Serum levels of IGF1 are decreased in patients who have received an allogeneic HSCT, and T-cell recovery correlates with an increase in serum IGF1 levels52. In mouse models, post-transplant administration of IGF1 to recipients of a syngeneic or an allogeneic HSCT resulted in an increase in thymopoiesis, peripheral T-cell numbers and proliferation, and it did not aggravate GVHD53,54. Interestingly, post-transplant administration of another neuroendocrine hormone, prolactin, also resulted in increased thymic cellularity and improved T-cell (and B-cell) reconstitution55. Clinical studies in patients with AIDS showed that IGF1 and growth hormone were well tolerated, could increase lean body mass and could increase thymic volume in children, but they had only modest effects on T-cell function (as determined by in vitro production of
| NOVEMBER 2004 | VOLUME 4
IL-2 in response to peptides derived from HIV)45. So, growth hormone and IGF1 are probably not required for normal haematopoiesis, but under conditions of stress (such as during AIDS, high-dose chemotherapy or lethal conditioning for HSCT), they can promote engraftment, haematopoiesis and thymopoiesis. Further trials are required to examine the potential toxic effects, including glucose intolerance, oedema and arthralgia, and to determine the theoretical risks of autoimmunity, graft rejection, GVHD and tumour-growth enhancement. Keratinocyte growth factor. Keratinocyte growth factor (KGF; also known as fibroblast growth factor 7, FGF7) acts through its receptor (FGF receptor 2 IIIb isoform, FGFR2IIIb ) on a variety of epithelial tissues, including hepatocytes56, gut epithelial cells56 and skin keratinocytes57. In the thymus, KGF is produced by both thymic stromal cells58 and thymocytes58,59, but FGFR2IIIb is only expressed by TECs58,59. FGFR2IIIbdeficient mice have arrested thymic development, which leads to decreased thymic cellularity and abnormal T-cell development60. In models of bone-marrow transplantation, thymic reconstitution after a syngeneic or an allogeneic HSCT was considerably enhanced following treatment with KGF58. Thymic cellularity was increased, and the developmental block that is usually observed after an HSCT, between DN and DP thymocytes, was released58. KGFtreated HSCT-recipient mice had considerably more cells containing Il-7 mRNA in their thymi than untreated recipients of an HSCT, and treatment with KGF had no effect on thymic reconstitution in recipients of an HSCT that were deficient in IL-7. Together, these results indicate a role for IL-7 in KGF-enhanced immune reconstitution58, although the abnormal thymic phenotype of IL-7-deficient mice (that is, a marked decrease in thymocyte number and abnormal thymic microenvironment) might prevent KGF from having an effect. An increase in the number of donorderived T cells in the spleen and lymph nodes was observed after administration of KGF to recipients of a T-cell-depleted HSCT58, and these T cells — both CD4+ and CD8+ — expressed markers of naive T cells, indicating that the increased cell number results from an increase in thymic export and not peripheral clonal expansion58. T-cell-dependent antibody responses were also enhanced following a syngeneic or an allogeneic HSCT and pretreatment with KGF58. Because FGFRIIIb is expressed by many of the organs that are damaged during GVHD, several research groups have studied the effects of treatment with KGF in the setting of acute GVHD. Administration of KGF can facilitate allo-engraftment and ameliorate the development of GVHD through a variety of mechanisms that are not related to T-cell reconstitution, including protection and repair of epithelial-cell injury in the gut mucosa, reduction in inflammatorycytokine release and inhibition of the allogeneic T-cell response61–63. Treatment with KGF also largely protected the thymic microenvironment from the alterations that are usually seen during acute GVHD64.
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS
EFFECTOR MEMORY CELLS
Memory T cells that home to peripheral tissues and plasma cells that home to the bone marrow and secrete antibodies. They are responsible for immediate protection.
KGF is currently in phase II and III clinical trials to assess its effects on both mucositis (a side-effect of chemotherapy that results in mucosal ulceration of the digestive tract) after high-dose chemotherapy and GVHD following an allogeneic HSCT65. Hopefully, these studies will also provide data regarding the effects of KGF on T-cell recovery.
CENTRAL MEMORY CELLS
Memory T and B cells that home to secondary lymphoid organs. These cells are heterogeneous and do not have the full range of functions that are characteristic of effector T cells or plasma cells. They are responsible for secondary or chronic responses to antigen and might be involved in longterm maintenance of effector memory cells.
Cytokines and co-stimulation
Interleukin-2 and interleukin-12. IL-2 was the first T-cell growth factor to be used in clinical trials to enhance lymphocyte activity in patients with cancer or AIDS; however, the results were mixed, and marked toxicity was found (reviewed in REFS 66,67). Post-transplant administration of a low dose of IL-2 to recipients of an autologous or allogeneic HSCT had little effect on T cells but did increase the number of NK cells by 5–10-fold68. Moreover, it was not associated with marked toxicity and did not exacerbate GVHD. A phase III study is in progress to determine whether the administration of IL-2 can decrease the relapse rate of patients who receive an autologous HSCT to treat haematological malignancies. Interestingly, administration of IL-2 to patients with HIV/AIDS resulted in the emergence of CD4+ T cells with a CD4+CD25+ regulatory T-cell phenotype69. In addition, a recent study showed that high-level production of IL-2 increased the risk of acute GVHD in patients who received an unrelated bone-marrow transplant70. IL-12 is produced by thymic dendritic cells (DCs)71, and IL-12β-deficient mice have accelerated thymic involution, which is associated with an increased number of DN1 thymocytes, degeneration of the thymic extracellular matrix and blood vessels, a decreased thymic cortex to medulla ratio and an increased number of apoptotic cells in aged mice72. IL-12 has a synergistic effect on both IL-7-induced and IL-2-induced proliferation of thymocytes, which indicates that a combination therapy including IL-12 could have a thymopoietic effect. However, despite the potential of IL-2 and IL-12 to enhance post-transplant T-cell recovery, the toxicities observed when these cytokines were administered to patients have diminished the enthusiasm for further clinical development. Interleukin-7. IL-7 is produced by stromal cells (in the thymus and bone marrow), keratinocytes, intestinal epithelial cells and DCs (reviewed in REF. 73). The IL-7 receptor (IL-7R) consists of a specific α-chain and the common cytokine-receptor γ-chain, which is also a component of the receptors for IL-2, IL-4, IL-9, IL-15 and IL-21 (REFS 74,75). IL-7R is expressed at a high level by lymphocyte precursors (including CLPs) (BOX 2), thymocytes (except for CD4+CD8+ (DP) thymocytes, which express only low levels of IL-7R), naive and memory T cells, and immature B cells. IL-7-deficient and IL-7Rα-deficient mice have no γδ T cells and have a 100-fold reduction in thymic cellularity; however, a small number of αβ T cells can develop normally76,77. Patients with mutations in IL-7Rα or the common cytokine-receptor γ-chain develop a SCID syndrome
NATURE REVIEWS | IMMUNOLOGY
with a marked T-cell deficiency. IL-7 has a variety of effects on lymphocyte development and survival, and it is required at various stages in the development of T cells from lymphoid precursors to memory T cells. In the thymus, IL-7 promotes the survival (probably through the upregulation of expression of B-cell lymphoma 2, BCL-2), differentiation and proliferation of CD4–CD8– (DN) thymocytes, as well as the survival and proliferation of CD4+ and CD8+ (single-positive) thymocytes78,79. In the periphery, IL-7 has proliferative and anti-apoptotic effects on mature T cells, through upregulation of expression of the survival factors BCL-2 and lung Kruppel-like factor80,81. IL-7 is not required for the initiation of an antigen-specific T-cell response, but it controls the transition of CD8+ T cells from 82 EFFECTOR MEMORY CELLS to CENTRAL MEMORY CELLS . IL-7 has been identified as a key regulator of peripheral T-cell homeostasis (BOX 3) and is required for the homeostatic proliferation of CD4+ and CD8+ T cells during peripheral lymphopaenia. IL-7 secretion is relatively constant, and its regulation is still poorly understood, except for the inhibitory effect of TGF-β on IL-7 production by bone-marrow stromal cells83. Serum levels of IL-7 are increased during lymphopaenia, which is probably owing to a decrease in the available target cells that IL-7 can interact with and not necessarily to an increase in IL-7 production. IL-7Rα is expressed by naive T cells but is downregulated after the activation of these cells and their subsequent transition to effector cells80,84. However, IL-7Rα is re-expressed by a small proportion of effector cells and is important for the development and survival of memory T cells82. Administration of IL-7 has several stimulatory effects on T-cell development, including increased thymopoiesis in mice85,86 (both in vitro and in vivo) and humans87 (in a thymic organ culture), increased numbers of peripheral CD4+ and CD8+ T cells without activation88, enhanced antiviral or antitumour activity of cytotoxic T cells that are clonally expanded in vitro for adoptive T-cell therapy89,90, increased homeostatic proliferation of both CD4 + and CD8+ T cells, and increased survival and proliferation of CD8+ memory T cells84. These encouraging findings have resulted in considerable interest in the clinical development of IL-7 as a ‘lymphoid growth factor’ for clinical situations that require enhanced T-cell function, including bone-marrow transplantation, vaccination and treatment of AIDS. Preclinical studies in mouse models of HSC transplantation have shown that administration of IL-7 after transplantation can enhance the reconstitution of T cells in recipients of a syngeneic or an allogeneic HSCT through increased thymopoiesis, increased homeostatic proliferation of transferred and de novo-generated mature T cells, and decreased apoptosis of peripheral T cells84,86,91–93. Recipients of an HSCT that were treated with IL-7 had augmented antimicrobial and antitumour activity, but the potent effects on T cells in the posttransplant setting carry the risk of aggravating GVHD in recipients of an allogeneic HSCT. Several studies have addressed this concern and have shown that prolonged
VOLUME 4 | NOVEMBER 2004 | 8 6 3 ©2004 Nature Publishing Group
REVIEWS
Box 3 | Peripheral homeostasis Under normal circumstances, the number of peripheral T cells is tightly regulated, and this has led to the concept of T-cell homeostasis, which is supported by several observations. First, the number of peripheral T cells in mice remains constant and depends only on strain or age. Second, the number of T cells increases to a normal level following sub-lethal irradiation or viral infection, and naive T cells proliferate after transfer to T-cell-deficient mice130,131. Third, T-cell receptor (TCR)-transgenic mice maintain a normal number of T cells131. Importantly, thymic output seems to depend on overall thymic size and cellularity and is not affected by changes in the number of peripheral T cells. Therefore, alterations in the naive T-cell pool are sensed — through as-yet-undefined regulatory mechanisms that might involve interleukin-7 (IL-7) and IL-15 — and this results in the loss or proliferation of naive T cells in the periphery132. Expression of MHC class I and II molecules in the periphery is one of the requirements for the survival and proliferation of naive T cells133. Homeostatic proliferation of peripheral T cells is most probably driven by low-affinity interactions between TCR molecules and self-peptide–MHC complexes at the cell surface of dendritic cells, similar to the interactions during positive selection in the thymus134–136. Naive and memory T cells seem to have independent homeostatic set points and seem to occupy separate and independent homeostatic niches in the peripheral T-cell pool134. When memory T cells are transferred to T-cell-deficient mice, they proliferate until their numbers are equal to the memory T-cell numbers in normal mice, and the transfer of a large number of memory T cells is unable to increase the size of the memory T-cell pool, even when naive T cells are absent132. The persistence of memory T cells in the periphery seems to be independent of exposure to peptide–MHC complexes. For example, CD8+ memory T cells could proliferate in MHC-class-I-deficient lymphopenic hosts, and CD4+ memory T cells could survive indefinitely in MHC-class-II-deficient hosts134. The cytokines IL-7 and IL-15 have an important role in peripheral T-cell homeostasis. IL-7 is a non-redundant cytokine for the survival and homeostasis of CD4+ and CD8+ naive and memory T cells137–140, whereas IL-15 supports the homeostasis and survival of CD8+ naive and memory T cells131,139,141,142. Recent studies have indicated that naive T cells are not quiescent cells that can persist indefinitely but, instead, require signals to survive in the periphery. These signals could involve the transcription factor nuclear factor of activated T cells 4 (NFAT4)143, lung Kruppel-like factor144 and members of the B-cell lymphoma 2 (BCL-2) family145, which are all crucial for the maintenance of a naive T-cell pool.
administration of high doses of IL-7 can aggravate GVHD94, whereas administration of IL-7 at lower doses and at shorter intervals had no effect on morbidity and mortality from GVHD. Importantly, administration of IL-7 to recipients of T-cell-depleted allografts (a highly effective strategy to prevent GVHD) did not result in the development of GVHD92. Studies using non-human primates have shown that administration of IL-7 has more profound effects on peripheral T-cell proliferation than on thymopoiesis and is less effective at promoting B-cell development than in rodents95. Importantly, marked toxicity was only observed at doses more than tenfold higher than the therapeutic dose (M. Morre, unpublished observations). Despite the lack of evidence for increased thymopoiesis in higher species, the promising effects on peripheral T cells, together with the effects observed in preclinical studies, have stimulated interest in clinical trials in patients with AIDS, tumour-associated immune deficiency and post-transplant immune deficiency, which are scheduled to begin soon. Interleukin-15. IL-15 is a pleiotropic cytokine that is particularly important for the development, activation, trafficking and homeostasis of NK cells, NKT cells and
864
| NOVEMBER 2004 | VOLUME 4
CD8+ memory T cells (reviewed in REF. 96). IL-15 binds a receptor with three chains — a unique α-chain (IL-15Rα), the IL-2R β-chain and the common cytokine-receptor γ-chain. IL-15 is produced by APCs (such as macrophages and DCs) and by epithelial cells in the bone marrow, thymus, kidney, skin and intestines. Mice deficient in IL-15 or its receptor lack NK cells and have decreased numbers of CD8+ memory T cells97,98. In mice, administration or overexpression of IL-15 results in increased proliferation, survival, cytolytic activity and cytokine secretion (including tumour-necrosis factor, interferon-γ and granulocyte colony-stimulating factor) of NK cells99,100 and memory CD8+ T cells101, as well as increased antimicrobial and antitumour activity102–104. Administration of IL-15 to recipients of an HSCT can increase graft-versus-tumour activity after a syngeneic105 or an allogeneic106 HSCT and can increase reconstitution of CD8+ memory T cells, NK cells and NKT cells after an allogeneic HSCT — as long as the allograft was depleted of T cells to avoid GVHD106. In addition, IL-15 is less toxic than IL-2 (REF. 107). These data from mouse models indicate that the administration of IL-15 could be an effective immunotherapy to enhance the number and function of CD8+ memory T cells, NK cells and NKT cells in various settings, including post-transplant immune reconstitution and vaccination against pathogens or tumours. In vitro studies have shown that IL-15 could enhance the function of HIV-specific CD8+ T cells108, leading to the suggestion that immunotherapy with IL-15 could be effective for patients with AIDS, who have defective IL-15 production. Superagonistic CD28-specific antibody. Co-stimulation through CD28 promotes the proliferation of T cells that have been activated through TCR engagement, by providing proliferative as well as anti-apoptotic signals109. Although conventional agonistic CD28-specific antibodies have no mitogenic activity in the absence of TCR stimulation, recent studies have identified ‘superagonistic’ CD28-specific antibodies, which recognize a unique epitope on the CD28 molecule and can induce polyclonal T-cell proliferation in the absence of TCR engagement110. Administration of superagonistic CD28-specific antibodies results in polyclonal T-cell expansion, together with a marked (but transient) increase in the number of regulatory T cells and the expression of antiinflammatory cytokines, including IL-10 (REF. 111). When tested in a syngeneic model of HSC transplantation in rats, the administration of superagonistic CD28-specific antibodies accelerated the post-transplant proliferation of a small number of mature T cells that had been transferred with the syngeneic bone-marrow graft but did not affect thymic output110. Interestingly, the clonal expansion of CD4+ T cells was greater than that of CD8+ T cells, and TCR-repertoire diversity and T-cell function were sustained. In contrast to polyclonal T-cell activation through stimulation of the TCR–CD3 complex — which results in an initial clonal expansion followed by clonal contraction (involving apoptosis by activationinduced cell death) — T cells induced following
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS administration of superagonistic CD28-specific antibodies persisted, possibly through the well-described anti-apoptotic effects of CD28 stimulation, including the upregulation of BCL-XL112. Superagonistic antibodies specific for human CD28 have been developed, and clinical trials are being planned for patients with cancer who have received myeloablative therapy. Oncostatin M. Oncostatin M is a member of the IL-6 family and can stimulate haemangioblasts and fetal hepatic cells during fetal development. Athymic transgenic mice that express oncostatin M in the early T-cell lineage can transform their lymph nodes (but not their spleen, gut, liver or bone marrow) into a primary lymphoid organ similar to a normal thymus, possibly through neoangiogenesis of oncostatin-M-receptorexpressing postcapillary venules, thereby allowing the entry of T-cell precursors113. These lymph nodes can support extrathymic T-cell development and produce a diversified repertoire of functional T cells. However, although administration of oncostatin M could potentially enhance T-cell recovery, the expression of oncostatin M is increased during age-associated thymic atrophy, and administration of oncostatin M to adult mice results in thymic atrophy, possibly owing to enhanced production of corticosteroids114. Therefore, determining the clinical potential of oncostatin M requires further preclinical studies to analyse its thymic and extrathymic effects, as well as its toxicities.
1.
2.
3.
4.
5.
6.
7.
8.
9. 10.
11.
12.
13.
Berzins, S. P. et al. Thymic regeneration: teaching an old immune system new tricks. Trends Mol. Med. 8, 469–476 (2002). O’Reilly, R. J. et al. Biology and adoptive cell therapy of Epstein–Barr virus-associated lymphoproliferative disorders in recipients of marrow allografts. Immunol. Rev. 157, 195–216 (1997). Dudley, M. E. & Rosenberg, S. A. Adoptive-cell-transfer therapy for the treatment of patients with cancer. Nature Rev. Cancer 3, 666–675 (2003). Small, T. N. et al. Comparison of immune reconstitution after unrelated and related T-cell-depleted bone marrow transplantation: effect of patient age and donor leukocyte infusions. Blood 93, 467–480 (1999). Storek, J. et al. Immunity of patients surviving 20 to 30 years after allogeneic or syngeneic bone marrow transplantation. Blood 98, 3505–3512 (2001). Parkman, R. & Weinberg, K. I. Immunological reconstitution following bone marrow transplantation. Immunol. Rev. 157, 73–78 (1997). Storek, J., Gooley, T., Witherspoon, R. P., Sullivan, K. M. & Storb, R. Infectious morbidity in long-term survivors of allogeneic marrow transplantation is associated with low CD4 T cell counts. Am. J. Hematol. 54, 131–138 (1997). Maraninchi, D. et al. Impact of T-cell depletion on outcome of allogeneic bone-marrow transplantation for standard-risk leukaemias. Lancet 2, 175–178 (1987). Curtis, R. E. et al. Solid cancers after bone marrow transplantation. N. Engl. J. Med. 336, 897–904 (1997). BitMansour, A. et al. Myeloid progenitors protect against invasive aspergillosis and Pseudomonas aeruginosa infection following hematopoietic stem cell transplantation. Blood 100, 4660–4667 (2002). Arber, C. et al. Common lymphoid progenitors rapidly engraft and protect against lethal murine cytomegalovirus infection after hematopoietic stem cell transplantation. Blood 102, 421–428 (2003). Atkinson, K. et al. Thymus transplantation after allogeneic bone marrow graft to prevent chronic graft-versus-host disease in humans. Transplantation 33, 168–173 (1982). Markert, M. L. et al. Transplantation of thymus tissue in complete DiGeorge syndrome. N. Engl. J. Med. 341, 1180–1189 (1999).
Summary
With the remarkable progress in our understanding of lymphoid precursors, thymic development and peripheral T-cell homeostasis, as well as our improved understanding of the individual molecules involved in these processes, new targets for therapeutic intervention have become available. Several clinical trials aimed at improving thymic rejuvenation and T-cell immunity are currently in progress or are anticipated to begin soon; these include trials of thymic grafting, KGF administration and administration of superagonistic CD28-specific antibody. In particular, there is early evidence that LHRH analogues improve immune-system recovery in recipients of an HSCT after myeloablative chemotherapy for leukaemia or lymphoma. The improvements induced by sexsteroid ablation therapy, at the level of both the thymus and the bone marrow, could form a platform from which to administer other more specific therapies directed at peripheral T cells. Phase I trials of IL-7 administration to patients with cancer are currently underway and will soon be followed by studies treating recipients of an HSCT. Collectively, these novel approaches to restoring immune capacity through the translation of preclinical research could result in the development of one or more new strategies to improve the outcome for a variety of patients who incur considerable morbidity and mortality from T-cell deficiency.
14. Markert, M. L. et al. Thymus transplantation in complete DiGeorge syndrome: immunologic and safety evaluations in 12 patients. Blood 102, 1121–1130 (2003). This article reports on the first series of patients with no detectable thymus function who received allogeneic cultured thymic tissue, resulting in recovery of T-cell function in 7 of 12 patients. 15. Hong, R., Schulte-Wissermann, H., Jarrett-Toth, E., Horowitz, S. D. & Manning, D. D. Transplantation of cultured thymic fragments. II. Results in nude mice. J. Exp. Med. 149, 398–415 (1979). 16. Waer, M., Palathumpat, V., Sobis, H. & Vandeputte, M. Induction of transplantation tolerance in mice across major histocompatibility barrier by using allogeneic thymus transplantation and total lymphoid irradiation. J. Immunol. 145, 499–504 (1990). 17. Yamada, K. et al. Thymic transplantation in miniature swine. II. Induction of tolerance by transplantation of composite thymokidneys to thymectomized recipients. J. Immunol. 164, 3079–3086 (2000). 18. Kamano, C. et al. Vascularized thymic lobe transplantation in miniature swine: thymopoiesis and tolerance induction across fully MHC-mismatched barriers. Proc. Natl Acad. Sci. USA 101, 3827–3832 (2004). 19. Menard, M. T. et al. Composite ‘thymoheart’ transplantation improves cardiac allograft survival. Am. J. Transplant. 4, 79–86 (2004). 20. Godfrey, D. I., Izon, D. J., Wilson, T. J., Tucek, C. L. & Boyd, R. L. Thymic stromal elements defined by M.Abs: ontogeny, and modulation in vivo by immunosuppression. Adv. Exp. Med. Biol. 237, 269–275 (1988). 21. Blackburn, C. C. et al. The nu gene acts cell-autonomously and is required for differentiation of thymic epithelial progenitors. Proc. Natl Acad. Sci. USA 93, 5742–5746 (1996). 22. Gill, J., Malin, M., Hollander, G. A. & Boyd, R. Generation of a complete thymic microenvironment by MTS24+ thymic epithelial cells. Nature Immunol. 3, 635–642 (2002). 23. Bennett, A. R. et al. Identification and characterization of thymic epithelial progenitor cells. Immunity 16, 803–814 (2002). These two articles describe the identification of a TEC subpopulation based on expression of the glycoprotein
NATURE REVIEWS | IMMUNOLOGY
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
MTS24. These cells can fully reconstitute the thymic epithelial microenvironment and can support normal T-cell development. Ceredig, R., Jenkinson, E. J., MacDonald, H. R. & Owen, J. J. Development of cytolytic T lymphocyte precursors in organcultured mouse embryonic thymus rudiments. J. Exp. Med. 155, 617–622 (1982). Poznansky, M. C. et al. Efficient generation of human T cells from a tissue-engineered thymic organoid. Nature Biotechnol. 18, 729–734 (2000). Rosenzweig, M. et al. In vitro T lymphopoiesis of human and rhesus CD34+ progenitor cells. Blood 87, 4040–4048 (1996). Gardner, J. P., Zhu, H., Colosi, P. C., Kurtzman, G. J. & Scadden, D. T. Robust, but transient expression of adeno-associated virus-transduced genes during human T lymphopoiesis. Blood 90, 4854–4864 (1997). Pawelec, G., Muller, R., Rehbein, A., Hahnel, K. & Ziegler, B. L. Extrathymic T cell differentiation in vitro from human CD34+ stem cells. J. Leukoc. Biol. 64, 733–739 (1998). Radtke, F., Wilson, A., Mancini, S. J. & MacDonald, H. R. Notch regulation of lymphocyte development and function. Nature Immunol. 5, 247–253 (2004). This recent review provides an excellent overview of the role of the Notch family and their ligands in lymphocyte development. Harman, B. C., Jenkinson, E. J. & Anderson, G. Microenvironmental regulation of Notch signalling in T cell development. Semin. Immunol. 15, 91–97 (2003). Schmitt, T. M. & Zuniga-Pflucker, J. C. Induction of T cell development from hematopoietic progenitor cells by delta-like-1 in vitro. Immunity 17, 749–756 (2002). This article describes the capability of a bone-marrow stromal cell line ectopically expressing the Notch ligand Delta-like-1 to support the differentiation of haematopoietic progenitors into mature T cells in vitro. Varnum-Finney, B., Brashem-Stein, C. & Bernstein, I. D. Combined effects of Notch signaling and cytokines induce a multiple log increase in precursors with lymphoid and myeloid reconstituting ability. Blood 101, 1784–1789 (2003). Zuniga-Pflucker, J. C. T-cell development made simple. Nature Rev. Immunol. 4, 67–72 (2004).
VOLUME 4 | NOVEMBER 2004 | 8 6 5 ©2004 Nature Publishing Group
REVIEWS 34. Aspinall, R. & Andrew, D. Thymic atrophy in the mouse is a soluble problem of the thymic environment. Vaccine 18, 1629–1637 (2000). 35. Nabarra, B. & Andrianarison, I. Ultrastructural study of thymic microenvironment involution in aging mice. Exp. Gerontol. 31, 489–506 (1996). 36. Steffens, C. M., Al-Harthi, L., Shott, S., Yogev, R. & Landay, A. Evaluation of thymopoiesis using T cell receptor excision circles (TRECs): differential correlation between adult and pediatric TRECs and naive phenotypes. Clin. Immunol. 97, 95–101 (2000). 37. Linton, P. J. & Dorshkind, K. Age-related changes in lymphocyte development and function. Nature Immunol. 5, 133–139 (2004). 38. Fitzpatrick, F. T., Kendall, M. D., Wheeler, M. J., Adcock, I. M. & Greenstein, B. D. Reappearance of thymus of ageing rats after orchidectomy. J. Endocrinol. 106, R17–R19 (1985). 39. Greenstein, B. D., Fitzpatrick, F. T., Kendall, M. D. & Wheeler, M. J. Regeneration of the thymus in old male rats treated with a stable analogue of LHRH. J. Endocrinol. 112, 345–350 (1987). 40. Windmill, K. F. & Lee, V. W. Effects of castration on the lymphocytes of the thymus, spleen and lymph nodes. Tissue Cell 30, 104–111 (1998). 41. Castro, J. E. Orchidectomy and the immune response. II. Response of orchidectomized mice to antigens. Proc. R. Soc. Lond. B 185, 437–451 (1974). 42. Olsen, N. J., Olson, G., Viselli, S. M., Gu, X. & Kovacs, W. J. Androgen receptors in thymic epithelium modulate thymus size and thymocyte development. Endocrinology 142, 1278–1283 (2001). 43. Clark, R., Strasser, J., McCabe, S., Robbins, K. & Jardieu, P. Insulin-like growth factor-1 stimulation of lymphopoiesis. J. Clin. Invest. 92, 540–548 (1993). 44. Welniak, L. A., Sun, R. & Murphy, W. J. The role of growth hormone in T-cell development and reconstitution. J. Leukoc. Biol. 71, 381–387 (2002). 45. Murphy, W. J. & Longo, D. L. Growth hormone as an immunomodulating therapeutic agent. Immunol. Today 21, 211–213 (2000). 46. Stuart, C. A., Meehan, R. T., Neale, L. S., Cintron, N. M. & Furlanetto, R. W. Insulin-like growth factor-I binds selectively to human peripheral blood monocytes and B-lymphocytes. J. Clin. Endocrinol. Metab. 72, 1117–1122 (1991). 47. Walsh, P. T. & O’Connor, R. The insulin-like growth factor-I receptor is regulated by CD28 and protects activated T cells from apoptosis. Eur. J. Immunol. 30, 1010–1018 (2000). 48. Murphy, W. J., Durum, S. K. & Longo, D. L. Role of neuroendocrine hormones in murine T cell development. Growth hormone exerts thymopoietic effects in vivo. J. Immunol. 149, 3851–3857 (1992). 49. Foster, M., Montecino-Rodriguez, E., Clark, R. & Dorshkind, K. Regulation of B and T cell development by anterior pituitary hormones. Cell. Mol. Life Sci. 54, 1076–1082 (1998). 50. Dobashi, H., Sato, M., Tanaka, T., Tokuda, M. & Ishida, T. Growth hormone restores glucocorticoid-induced T cell suppression. FASEB J. 15, 1861–1863 (2001). 51. Tian, Z. G. et al. Recombinant human growth hormone promotes hematopoietic reconstitution after syngeneic bone marrow transplantation in mice. Stem Cells 16, 193–199 (1998). 52. Small, T. et al. Longitudinal analysis of serum levels of insulinlike growth factor-1 post bone marrow transplantation. Blood 90, 541a (1997). 53. Jardieu, P., Clark, R., Mortensen, D. & Dorshkind, K. In vivo administration of insulin-like growth factor-I stimulates primary B lymphopoiesis and enhances lymphocyte recovery after bone marrow transplantation. J. Immunol. 152, 4320–4327 (1994). 54. Alpdogan, O. et al. Insulin-like growth factor-I enhances lymphoid and myeloid reconstitution after allogeneic bone marrow transplantation. Transplantation 75, 1977–1983 (2003). 55. Sun, R. et al. Immunologic and hematopoietic effects of recombinant human prolactin after syngeneic bone marrow transplantation in mice. Biol. Blood Marrow Transplant 9, 426–434 (2003). 56. Housley, R. M. et al. Keratinocyte growth factor induces proliferation of hepatocytes and epithelial cells throughout the rat gastrointestinal tract. J. Clin. Invest. 94, 1764–1777 (1994). 57. Pierce, G. F. et al. Stimulation of all epithelial elements during skin regeneration by keratinocyte growth factor. J. Exp. Med. 179, 831–840 (1994). 58. Min, D. et al. Protection from thymic epithelial cell injury by keratinocyte growth factor: a new approach to improve thymic and peripheral T-cell reconstitution after bone marrow transplantation. Blood 99, 4592–4600 (2002).
866
59. 60.
61.
62.
63.
64.
65.
66. 67.
68.
69.
70.
71.
72.
73. 74.
75.
76.
77.
78.
79.
80.
81.
82.
This article describes how KGF can protect TECs against cytotoxic-therapy-induced damage, resulting in enhanced post-transplant T-cell recovery and function. Erickson, M. et al. Regulation of thymic epithelium by keratinocyte growth factor. Blood 100, 3269–3278 (2002). Revest, J. M., Suniara, R. K., Kerr, K., Owen, J. J. & Dickson, C. Development of the thymus requires signaling through the fibroblast growth factor receptor R2-IIIb. J. Immunol. 167, 1954–1961 (2001). Panoskaltsis-Mortari, A., Lacey, D. L., Vallera, D. A. & Blazar, B. R. Keratinocyte growth factor administered before conditioning ameliorates graft-versus-host disease after allogeneic bone marrow transplantation in mice. Blood 92, 3960–3967 (1998). Krijanovski, O. I. et al. Keratinocyte growth factor separates graft-versus-leukemia effects from graft-versus-host disease. Blood 94, 825–831 (1999). Panoskaltsis-Mortari, A. et al. Keratinocyte growth factor facilitates alloengraftment and ameliorates graft-versus-host disease in mice by a mechanism independent of repair of conditioning-induced tissue injury. Blood 96, 4350–4356 (2000). Rossi, S. et al. Keratinocyte growth factor preserves normal thymopoiesis and thymic microenvironment during experimental graft-versus-host disease. Blood 100, 682–691 (2002). Meropol, N. J. et al. Randomized phase I trial of recombinant human keratinocyte growth factor plus chemotherapy: potential role as mucosal protectant. J. Clin. Oncol. 21, 1452–1458 (2003). Nelson, B. H. IL-2, regulatory T Cells, and tolerance. J. Immunol. 172, 3983–3988 (2004). Dutcher, J. Current status of interleukin-2 therapy for metastatic renal cell carcinoma and metastatic melanoma. Oncology (Huntington, NY) 16, 4–10 (2002). Soiffer, R. J., Murray, C., Gonin, R. & Ritz, J. Effect of lowdose interleukin-2 on disease relapse after T-cell-depleted allogeneic bone marrow transplantation. Blood 84, 964–971 (1994). Sereti, I. et al. Long-term effects of intermittent interleukin 2 therapy in patients with HIV infection: characterization of a novel subset of CD4+/CD25+ T cells. Blood 100, 2159–2167 (2002). MacMillan, M. L. et al. High-producer interleukin-2 genotype increases risk for acute graft-versus-host disease after unrelated donor bone marrow transplantation. Transplantation 76, 1758–1762 (2003). Rizzitelli, A., Berthier, R., Collin, V., Candeias, S. M. & Marche, P. N. T lymphocytes potentiate murine dendritic cells to produce IL-12. J. Immunol. 169, 4237–4245 (2002). Li, L. et al. IL-12 inhibits thymic involution by enhancing IL-7- and IL-2-induced thymocyte proliferation. J. Immunol. 172, 2909–2916 (2004). Fry, T. J. & Mackall, C. L. Interleukin-7: from bench to clinic. Blood 99, 3892–3904 (2002). Noguchi, M. et al. Interleukin-2 receptor γ chain: a functional component of the interleukin-7 receptor. Science 262, 1877–1880 (1993). Goodwin, R. G. et al. Cloning of the human and murine interleukin-7 receptors: demonstration of a soluble form and homology to a new receptor superfamily. Cell 60, 941–951 (1990). von Freeden-Jeffry, U. et al. Lymphopenia in interleukin (IL)-7 gene-deleted mice identifies IL-7 as a nonredundant cytokine. J. Exp. Med. 181, 1519–1526 (1995). Peschon, J. J. et al. Early lymphocyte expansion is severely impaired in interleukin 7 receptor-deficient mice. J. Exp. Med. 180, 1955–1960 (1994). von Freeden-Jeffry, U., Solvason, N., Howard, M. & Murray, R. The earliest T lineage-committed cells depend on IL-7 for Bcl-2 expression and normal cell cycle progression. Immunity 7, 147–154 (1997). Akashi, K., Kondo, M., von Freeden-Jeffry, U., Murray, R. & Weissman, I. L. Bcl-2 rescues T lymphopoiesis in interleukin-7 receptor-deficient mice. Cell 89, 1033–1041 (1997). Schluns, K. S., Kieper, W. C., Jameson, S. C. & Lefrancois, L. Interleukin-7 mediates the homeostasis of naive and memory CD8 T cells in vivo. Nature Immunol. 1, 426–432 (2000). This is one of several articles to define the crucial role of IL-7 in the homeostasis of peripheral T cells. Schober, S. L. et al. Expression of the transcription factor lung Kruppel-like factor is regulated by cytokines and correlates with survival of memory T cells in vitro and in vivo. J. Immunol. 163, 3662–3667 (1999). Kaech, S. M. et al. Selective expression of the interleukin 7 receptor identifies effector CD8 T cells that give rise to longlived memory cells. Nature Immunol. 4, 1191–1198 (2003).
| NOVEMBER 2004 | VOLUME 4
83. Tang, J. et al. TGF-β down-regulates stromal IL-7 secretion and inhibits proliferation of human B cell precursors. J. Immunol. 159, 117–125 (1997). 84. Alpdogan, O. et al. IL-7 enhances peripheral T cell reconstitution after allogeneic hematopoietic stem cell transplantation. J. Clin. Invest. 112, 1095–1107 (2003). 85. Sempowski, G. D., Gooding, M. E., Liao, H. X., Le, P. T. & Haynes, B. F. T cell receptor excision circle assessment of thymopoiesis in aging mice. Mol. Immunol. 38, 841–848 (2002). 86. Bolotin, E., Smogorzewska, M., Smith, S., Widmer, M. & Weinberg, K. Enhancement of thymopoiesis after bone marrow transplant by in vivo interleukin-7. Blood 88, 1887–1894 (1996). 87. Okamoto, Y., Douek, D. C., McFarland, R. D. & Koup, R. A. Effects of exogenous interleukin-7 on human thymus function. Blood 99, 2851–2858 (2002). 88. Geiselhart, L. A. et al. IL-7 administration alters the CD4:CD8 ratio, increases T cell numbers, and increases T cell function in the absence of activation. J. Immunol. 166, 3019–3027 (2001). 89. Lynch, D. H. & Miller, R. E. Interleukin 7 promotes long-term in vitro growth of antitumor cytotoxic T lymphocytes with immunotherapeutic efficacy in vivo. J. Exp. Med. 179, 31–42 (1994). 90. Wiryana, P., Bui, T., Faltynek, C. R. & Ho, R. J. Augmentation of cell-mediated immunotherapy against herpes simplex virus by interleukins: comparison of in vivo effects of IL-2 and IL-7 on adoptively transferred T cells. Vaccine 15, 561–563 (1997). 91. Abdul-Hai, A. et al. Stimulation of immune reconstitution by interleukin-7 after syngeneic bone marrow transplantation in mice. Exp. Hematol. 24, 1416–1422 (1996). 92. Alpdogan, O. et al. Administration of interleukin-7 after allogeneic bone marrow transplantation improves immune reconstitution without aggravating graft-versus-host disease. Blood 98, 2256–2265 (2001). 93. Mackall, C. L. et al. IL-7 increases both thymic-dependent and thymic-independent T-cell regeneration after bone marrow transplantation. Blood 97, 1491–1497 (2001). 94. Sinha, M. L., Fry, T. J., Fowler, D. H., Miller, G. & Mackall, C. L. Interleukin 7 worsens graft-versus-host disease. Blood 100, 2642–2649 (2002). 95. Storek, J. et al. Interleukin-7 improves CD4 T-cell reconstitution after autologous CD34 cell transplantation in monkeys. Blood 101, 4209–4218 (2003). 96. Fehniger, T. A. & Caligiuri, M. A. Interleukin 15: biology and relevance to human disease. Blood 97, 14–32 (2001). 97. Kennedy, M. K. et al. Reversible defects in natural killer and memory CD8 T cell lineages in interleukin 15-deficient mice. J. Exp. Med. 191, 771–780 (2000). 98. Lodolce, J. P. et al. IL-15 receptor maintains lymphoid homeostasis by supporting lymphocyte homing and proliferation. Immunity 9, 669–676 (1998). 99. Carson, W. E. et al. Interleukin (IL) 15 is a novel cytokine that activates human natural killer cells via components of the IL-2 receptor. J. Exp. Med. 180, 1395–1403 (1994). 100. Carson, W. E. et al. A potential role for interleukin-15 in the regulation of human natural killer cell survival. J. Clin. Invest. 99, 937–943 (1997). 101. Zhang, X., Sun, S., Hwang, I., Tough, D. F. & Sprent, J. Potent and selective stimulation of memory-phenotype CD8+ T cells in vivo by IL-15. Immunity 8, 591–599 (1998). 102. Maeurer, M. J. et al. Interleukin-7 or interleukin-15 enhances survival of Mycobacterium tuberculosis-infected mice. Infect. Immun. 68, 2962–2970 (2000). 103. Yajima, T. et al. Memory phenotype CD8+ T cells in IL-15 transgenic mice are involved in early protection against a primary infection with Listeria monocytogenes. Eur. J. Immunol. 31, 757–766 (2001). 104. Klebanoff, C. A. et al. IL-15 enhances the in vivo antitumor activity of tumor-reactive CD8+ T cells. Proc. Natl Acad. Sci. USA 101, 1969–1974 (2004). 105. Katsanis, E. et al. IL-15 administration following syngeneic bone marrow transplantation prolongs survival of lymphoma bearing mice. Transplantation 62, 872–875 (1996). 106. Alpdogan, O. et al. Interleukin-15 enhances immune reconstitution after allogeneic bone marrow transplantation. Blood July 27 2004 (doi:10.1182/blood-2003-09-3344). 107. Munger, W. et al. Studies evaluating the antitumor activity and toxicity of interleukin-15, a new T cell growth factor: comparison with interleukin-2. Cell. Immunol. 165, 289–293 (1995). 108. Castelli, J., Thomas, E. K., Gilliet, M., Liu, Y. J. & Levy, J. A. Mature dendritic cells can enhance CD8+ cell noncytotoxic anti-HIV responses: the role of IL-15. Blood 103, 2699–2704 (2004). 109. Acuto, O. & Michel, F. CD28-mediated co-stimulation: a quantitative support for TCR signalling. Nature Rev. Immunol. 3, 939–951 (2003).
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS 110. Elflein, K., Rodriguez-Palmero, M., Kerkau, T. & Hunig, T. Rapid recovery from T lymphopenia by CD28 superagonist therapy. Blood 102, 1764–1770 (2003). This article shows that a novel class of superagonistic CD28-specific antibodies can induce polyclonal T-cell proliferation without TCR engagement. 111. Lin, C. H. & Hunig, T. Efficient expansion of regulatory T cells in vitro and in vivo with a CD28 superagonist. Eur. J. Immunol. 33, 626–638 (2003). 112. Kerstan, A. & Hunig, T. Distinct TCR- and CD28-derived signals regulate CD95L, Bcl-xL, and the survival of primary T cells. J. Immunol. 172, 1341–1345 (2004). 113. Clegg, C. H., Rulffes, J. T., Wallace, P. M. & Haugen, H. S. Regulation of an extrathymic T-cell development pathway by oncostatin M. Nature 384, 261–263 (1996). 114. Sempowski, G. D. et al. Leukemia inhibitory factor, oncostatin M, IL-6, and stem cell factor mRNA expression in human thymus increases with age and is associated with thymic atrophy. J. Immunol. 164, 2180–2187 (2000). 115. Mackall, C. L. & Gress, R. E. Pathways of T-cell regeneration in mice and humans: implications for bone marrow transplantation and immunotherapy. Immunol. Rev. 157, 61–72 (1997). 116. Dulude, G. et al. Thymic and extrathymic differentiation and expansion of T lymphocytes following bone marrow transplantation in irradiated recipients. Exp. Hematol. 25, 992–1004 (1997). 117. Roux, E. et al. Analysis of T-cell repopulation after allogeneic bone marrow transplantation: significant differences between recipients of T-cell depleted and unmanipulated grafts. Blood 87, 3984–3992 (1996). 118. Hebib, N. C. et al. Peripheral blood T cells generated after allogeneic bone marrow transplantation: lower levels of Bcl-2 protein and enhanced sensitivity to spontaneous and CD95-mediated apoptosis in vitro. Abrogation of the apoptotic phenotype coincides with the recovery of normal naive/primed T-cell profiles. Blood 94, 1803–1813 (1999). 119. Lin, M. T. et al. Increased apoptosis of peripheral blood T cells following allogeneic hematopoietic cell transplantation. Blood 95, 3832–3839 (2000). 120. Scollay, R., Smith, J. & Stauffer, V. Dynamics of early T cells: prothymocyte migration and proliferation in the adult mouse thymus. Immunol. Rev. 91, 129–157 (1986). 121. Foss, D. L., Donskoy, E. & Goldschneider, I. The importation of hematogenous precursors by the thymus is a gated phenomenon in normal adult mice. J. Exp. Med. 193, 365–374 (2001). 122. Petrie, H. T. Role of thymic organ structure and stromal composition in steady-state postnatal T-cell production. Immunol. Rev. 189, 8–20 (2002). 123. Bhandoola, A., Sambandam, A., Allman, D., Meraz, A. & Schwarz, B. Early T lineage progenitors: new insights, but old questions remain. J. Immunol. 171, 5653–5658 (2003).
124. Kondo, M., Weissman, I. L. & Akashi, K. Identification of clonogenic common lymphoid progenitors in mouse bone marrow. Cell 91, 661–672 (1997). This article describes the identification of a CLP in adult bone marrow that can give rise to T cells, B cells and NK cells. 125. Martin, C. H. et al. Efficient thymic immigration of B220+ lymphoid-restricted bone marrow cells with T precursor potential. Nature Immunol. 4, 866–873 (2003). 126. Allman, D. et al. Thymopoiesis independent of common lymphoid progenitors. Nature Immunol. 4, 168–174 (2003). This article shows that early T-cell lineage progenitors are present in the thymus, and therefore the production of T-cell lineage progeny could be sustained by a CLP-independent pathway. 127. Sitnicka, E. et al. Key role of flt3 ligand in regulation of the common lymphoid progenitor but not in maintenance of the hematopoietic stem cell pool. Immunity 17, 463–472 (2002). 128. Igarashi, H., Gregory, S. C., Yokota, T., Sakaguchi, N. & Kincade, P. W. Transcription from the RAG1 locus marks the earliest lymphocyte progenitors in bone marrow. Immunity 17, 117–130 (2002). 129. Perry, S. et al. L-selectin defines a bone marrow analog to the thymic early T-lineage progenitor. Blood 103, 2990–2996 (2004). 130. Rocha, B., Dautigny, N. & Pereira, P. Peripheral T lymphocytes: expansion potential and homeostatic regulation of pool sizes and CD4/CD8 ratios in vivo. Eur. J. Immunol. 19, 905–911 (1989). 131. Ku, C. C., Murakami, M., Sakamoto, A., Kappler, J. & Marrack, P. Control of homeostasis of CD8+ memory T cells by opposing cytokines. Science 288, 675–678 (2000). 132. Tanchot, C. & Rocha, B. The organization of mature T-cell pools. Immunol. Today 19, 575–579 (1998). 133. Freitas, A. A. & Rocha, B. Peripheral T cell survival. Curr. Opin. Immunol. 11, 152–156 (1999). 134. Goldrath, A. W. & Bevan, M. J. Selecting and maintaining a diverse T-cell repertoire. Nature 402, 255–262 (1999). 135. Viret, C., Wong, F. S. & Janeway, C. A. Jr. Designing and maintaining the mature TCR repertoire: the continuum of self-peptide:self-MHC complex recognition. Immunity 10, 559–568 (1999). 136. Dulude, G., Roy, D. C. & Perreault, C. The effect of graftversus-host disease on T cell production and homeostasis. J. Exp. Med. 189, 1329–1342 (1999). 137. Tan, J. T. et al. IL-7 is critical for homeostatic proliferation and survival of naive T cells. Proc. Natl Acad. Sci. USA 98, 8732–8737 (2001). 138. Schluns, K. S., Kieper, W. C., Jameson, S. C. & Lefrancois, L. Interleukin-7 mediates the homeostasis of naive and memory CD8 T cells in vivo. Nature Immunol. 1, 426–432 (2000).
NATURE REVIEWS | IMMUNOLOGY
139. Goldrath, A. W. et al. Cytokine requirements for acute and basal homeostatic proliferation of naive and memory CD8+ T cells. J. Exp. Med. 195, 1515–1522 (2002). 140. Seddon, B., Tomlinson, P. & Zamoyska, R. Interleukin 7 and T cell receptor signals regulate homeostasis of CD4 memory cells. Nature Immunol. 4, 680–686 (2003). 141. Alves, N. L., Hooibrink, B., Arosa, F. A. & van Lier, R. A. IL-15 induces antigen-independent expansion and differentiation of human naive CD8+ T cells in vitro. Blood 102, 2541–2546 (2003). 142. Tan, J. T. et al. Interleukin (IL)-15 and IL-7 jointly regulate homeostatic proliferation of memory phenotype CD8+ cells but are not required for memory phenotype CD4+ cells. J. Exp. Med. 195, 1523–1532 (2002). 143. Oukka, M. et al. The transcription factor NFAT4 is involved in the generation and survival of T cells. Immunity 9, 295–304 (1998). 144. Kuo, C. T., Veselits, M. L. & Leiden, J. M. LKLF: a transcriptional regulator of single-positive T cell quiescence and survival. Science 277, 1986–1990 (1997). 145. Veis, D. J., Sorenson, C. M., Shutter, J. R. & Korsmeyer, S. J. Bcl-2-deficient mice demonstrate fulminant lymphoid apoptosis, polycystic kidneys, and hypopigmented hair. Cell 75, 229–240 (1993).
Acknowledgements This work was supported by grants to M.R.M.B. from the National Institutes of Health, United States. M.R.M.B. is also the recipient of a Damon Runyon Scholar Award from the Damon Runyon Cancer Research Foundation, United Kingdom. This work was also supported by grants to R.L.B. from the National Health and Medical Research Council, Australia. The authors thank G. Goldberg for her many valuable contributions to the manuscript.
Competing interests statement The authors declare competing financial interests: see Web version for details.
Online links DATABASES The following terms in this article are linked online to: Entrez Gene: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=gene CD28 | growth hormone | IGF1 | IL-7 | IL-12 | IL-15 | KGF | oncostatin M | Notch-1 FURTHER INFORMATION Marcel van den Brink’s laboratory: http://www.mskcc.org/mskcc/html/10937.cfm Richard Boyd’s loboratory: http://www-personal.monash.edu.au/~malin/Boyd/ Access to this interactive links box is free online.
VOLUME 4 | NOVEMBER 2004 | 8 6 7 ©2004 Nature Publishing Group
REVIEWS
RETROVIRAL RESTRICTION BY APOBEC PROTEINS Reuben S. Harris and Mark T. Liddament Abstract | A powerful mechanism of vertebrate innate immunity has been discovered in the past year, in which APOBEC proteins inhibit retroviruses by deaminating cytosine residues in nascent retroviral cDNA. To thwart this cellular defence, HIV encodes Vif, a small protein that mediates APOBEC degradation. Therefore, the balance between APOBECs and Vif might be a crucial determinant of the outcome of retroviral infection. Vertebrates have up to 11 different APOBEC proteins, with primates having the most. APOBEC proteins include AID, a probable DNA mutator that is responsible for immunoglobulin-gene diversification, and APOBEC1, an RNA editor with antiretroviral activities. This APOBEC abundance might help to tip the balance in favour of cellular defences. TARGET CELL
Any cell lacking the virus under consideration that subsequently will be infected. APOBEC
(Apolipoprotein B (APOB) mRNA-editing, catalytic polypeptide). Although originally used to describe APOBEC1, which edits C6666 in APOB mRNA, this acronym has also become the accepted prefix for naming related vertebrate proteins that have the hallmark cytosinedeaminase motif His-XGlu–X23–28-Pro-Cys-X2–4-Cys.
University of Minnesota, Biochemistry, Molecular Biology and Biophysics Department, 321 Church Street South East, 6-155 Jackson Hall, Minneapolis, Minnesota 55455, USA. Correspondence to R.S.H. e-mail:
[email protected] doi:10.1038/nri1489
868
Nothing in biology makes sense except in the light of evolution. Theodosius Dobzhansky 1
Innate immune defences are integral components of eukaryotic systems, poised at all times to readily neutralize invaders. Eukaryotes have a wide variety of innate defences, including antimicrobial peptides, proteolytic cascades, signalling molecules such as interferons and specialized phagocytic cells; these all work together against pathogens (reviewed in REFS 2,3). In fact, even the outer membrane of a cell and the epithelial-cell surfaces of multicellular organisms can be considered innate immune defences, in that they are ever-present barriers to infection. In the past year, a novel mechanism of innate immunity has entered the spotlight — a potent cellular defence that actively blocks retroviral infection (FIG. 1). At least two proteins lie at the heart of this defence mechanism: APOBEC3F and APOBEC3G. These cellular proteins function by ‘hitchhiking’ with newly produced viral particles until a new TARGET CELL is found. Then, during synthesis of the first retroviral DNA strand (minus strand), which is an obligate step in the retroviral life cycle, APOBEC-dependent DEAMINATION of cytosine (C) residues results in the accumulation of excessive levels of uracil (U). This pre-mutagenic lesion leads to
| NOVEMBER 2004 | VOLUME 4
the demise of the invading RETROVIRUS on its replication, because uracil is recognized as thymine (T) by the viral reverse transcriptase and adenine (A) is incorporated into the newly synthesized second (plus) DNA strand rather than guanine (G). This process of lesion fixation can therefore produce a detrimental level of MUTATIONS in the viral genome. APOBEC3F and APOBEC3G are closely related to another cytosine deaminase, ACTIVATION-INDUCED DEAMINASE (AID), which also uses C to U deamination to initiate three distinct types of immunoglobulin-gene diversification: SOMATIC HYPERMUTATION, GENE CONVERSION and CLASS-SWITCH RECOMBINATION (reviewed in REFS 4,5). These processes are an integral part of the DNA-level modifications that drive maturation of the vertebrate antibody response to pathogens. The apparent deliberate use of DNA deamination by APOBEC3F, APOBEC3G and AID therefore constitutes a striking mechanistic parallel between innate and adaptive immune defences, both of which use this strategy to restrict infection. Here, we focus on mechanisms of (retro)viral restriction by APOBEC-family members. We aim to highlight some of the remarkable parallels between this innate cellular defence and the adaptive antibody defence that is underpinned by AID. We also consider the other putative and known cytosine deaminases that are encoded by the human genome. Finally, we note that
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS
Proteasome
Nucleus
Producer cell LTR Provirus
RNA
Vif
Incorporation
APOBEC
Ubiquitylation and degradation
Budding
Fusion CCT RT RNase H
Degradation? UCT
Degradation?
RT
Integration
C/G → T/A hypermutation Target cell
Figure 1 | APOBEC3G and Vif are key determinants of retroviral infectivity. This figure presents a model that depicts how the APOBEC proteins and Vif (virion infectivity factor) can influence the retroviral life cycle. APOBEC3G (red) that is expressed in the PRODUCER CELL is incorporated into the budding virion together with other components of the virus, including its genomic RNA. HIV-1 Vif (green) can reduce or eliminate APOBEC3G incorporation into budding virions by targeting it for proteasomal degradation. However, should APOBEC3G escape Vif, gain access to the virion and subsequently reach a target cell, it can deaminate cytosine residues in the first retroviral DNA strand (blue). The resulting uracil residues function as a template for the incorporation of adenine, which, in turn, can result in strand-specific C/G to T/A transition mutations that affect virus viability. Uracil residues have also been hypothesized to trigger degradation of the retroviral DNA before it can integrate into the host-cell genome16,18,19,104, although this hypothesis awaits rigorous experimentation. LTR, long terminal repeat; RT, reverse transcriptase.
a stunning increase in the number of APOBEC proteins has occurred since the diversification of vertebrates. This evolutionary perspective (see quote) is intended to encourage speculation and experimentation towards understanding the origin and function of the ancestral APOBEC-family member (or members), which probably served a dual role — both as an innate immune defender and as an antibody-response trigger. DEAMINATION
Removal of an amine group from a pyrimidine or purine nucleic-acid base. Deamination of cytosine and adenosine yields uracil and inosine, respectively. PRODUCER CELL
Any cell used to propagate viruses.
Retroviral restriction by APOBEC3G
Since the identification of HIV-1 two decades ago6–8, scientists have only been able to grow the virus in vitro using a subset of immortalized human T-cell lines. Such cell lines are considered ‘permissive’ for viral growth. Several factors are responsible for rendering cells ‘nonpermissive’, such as the absence of the viral receptor or coreceptor proteins — that is, CD4 and the chemokine
NATURE REVIEWS | IMMUNOLOGY
receptors CXC-chemokine receptor 4 (CXCR4) or CCchemokine receptor 5 (CCR5), respectively. Other less obvious explanations lack molecular detail. However, of the permissive T-cell lines that are able to replicate HIV-1, a subset is non-permissive for viruses that lack the small HIV ‘accessory’ protein Vif (virion infectivity factor). It is now appreciated that these cells are non-permissive owing to APOBEC protein expression9. Malim and colleagues identified APOBEC3G as a factor expressed by human T cells that failed to support the production of Vif-deficient HIV-1 (REF. 9). Its presence in these non-permissive T-cell lines was matched by its absence from T-cell lines that supported the growth of Vif-deficient viruses. Importantly, APOBEC3G expression alone could render permissive cell lines unable to replicate Vif-deficient HIV-1, whereas its expression had little discernable effect on the replication of Vif-proficient HIV-1. A close relationship was therefore established between APOBEC3G and Vif, which stimulated intense experimentation in search of a molecular explanation. APOBEC3G belongs to a larger family of bona fide and putative cytosine-deaminase proteins10 (FIG. 2a,b). Marked amino-acid similarities between APOBEC3G and the human mRNA-editing protein APOBEC1 (REF. 11) (BOX 1) indicated that the robust antiretroviral phenotype of APOBEC3G might also be mediated by the editing of a viral or cellular RNA9. However, APOBEC3G is also similar to AID, a protein that is thought to trigger immunoglobulin gene-diversification events by deaminating cytosine residues in immunoglobulin-gene DNA (discussed later and reviewed in REFS 4,5). Moreover, similar to AID, APOBEC3G can cause C/G to T/A transition mutations in DNA12,13. Such a transitionmutation bias could account for the retroviral mutation patterns that were previously observed in cell-based systems14 and in patients15, in which high frequencies of genomic strand G to A transition mutations had been reported (C to T on the minus strand). This phenomenon had therefore been appropriately termed RETROVIRAL 14 HYPERMUTATION . The main mechanism by which APOBEC3G restricts retroviral infection became clear when several groups sequenced retroviral cDNA from cells that had been infected with Vif-deficient retroviruses that were produced in the presence of APOBEC3G16–19. Retroviral DNA produced in the presence of APOBEC3G showed extremely high levels of plus-strand G to A transition mutations, in some cases involving up to 25% of all guanines sequenced16. By contrast, DNA from retroviruses that were produced in the absence of APOBEC3G showed little sign of HYPERMUTATION. These results unambiguously supported an antiretroviral mechanism in which APOBEC3G gains access to the virion and, on infection of a new cell, triggers the deamination of cytosine residues in the first cDNA strand of the replicating retroviral genome (FIG. 1). The resulting uracils then function as a template for the incorporation of second-strand adenines, thereby fixing the lesions as G to A hypermutations, which can affect the viability of the virus. In support of such a
VOLUME 4 | NOVEMBER 2004 | 8 6 9 ©2004 Nature Publishing Group
REVIEWS
RETROVIRUS
A class of enveloped RNA virus that is distinguished by the requirement for reverse transcription of the RNA genome by a viral reversetranscriptase enzyme to form a DNA intermediate that is then stably integrated into host chromosomal DNA. Lentiviruses such as HIV-1 are a subset of the retrovirus family that are further distinguished by numerous accessory proteins (for example, virion infectivity factor, Vif). MUTATIONS
Heritable changes in an organism’s nucleic acid. ACTIVATION-INDUCED DEAMINASE
(AID). A cytosine deaminase that catalyses a pivotal step in immunoglobulin genediversification reactions.
mechanism, purified APOBEC3G has been shown to preferentially deaminate cytosine residues in singlestranded DNA but not in DNA–RNA hybrids or doublestranded DNA substrates16,20,21. Interestingly, the robust cytosine-deamination activity of APOBEC3G does not depend on target-cell components, as purified viral particles that were incubated with dNTPs (deoxynucleotides) were capable of manifesting APOBEC3G-dependent retroviral hypermutations19. The mechanism of cytosine deamination by APOBEC3G is undoubtedly similar to that of APOBEC1 and AID (FIG. 2b,c). Cytosine and cytidine deaminases are characterized by a zinc-binding deaminase motif with the consensus amino-acid sequence His-X-Glu-X23–28Pro-Cys-X2– 4-Cys (where X denotes any amino acid)10. Based on the crystal structures of bacterial22,23 and yeast24,25 cytidine deaminases, the histidine and cysteine residues probably coordinate the zinc ion that is necessary for catalytic activity, and the glutamic-acid residue helps to produce the hydroxide ion that is required for amine-group removal (FIG. 2c). Unlike APOBEC1 and AID, which are characterized by a single deaminase motif, APOBEC3G is a double-deaminase-domain
protein. Altering conserved amino acids in either of the putative zinc-binding deaminase motifs reduced the ability of APOBEC3G to restrict retroviral infection18,19,26. However, it is unclear whether one or both of these motifs contributes directly to retroviral restriction. This is because G to A hypermutations were still apparent in the DNA of retroviruses that were produced in the presence of APOBEC3G with amino (N)- or carboxyterminal active-site glutamic acid to glutamine substitutions26. These data contrast with results that were obtained with various APOBEC1 and AID active-site mutants, which are catalytically inert13,27–30. So, future determination of the three-dimensional structure of APOBEC3G complexed with single-stranded DNA should help to resolve how this double-deaminasedomain protein catalyses C to U conversion. Incorporation of APOBEC3G into virions
The high levels of retroviral cDNA hypermutation clearly show that APOBEC3G gains access to retroviral particles16,18,19 (FIG. 1). Several groups have confirmed the presence of APOBEC3G in viral particles9,16,18, but it is not yet clear how it gets there. Several explanations are
SOMATIC HYPERMUTATION
High-frequency basesubstitution mutations found in B-cell immunoglobulin-gene variable regions.
a
GENE CONVERSION
Chromosome 6
A1
AID
Chromosome 12 A2
A non-reciprocal recombination event between homologous or partially homologous DNA sequences, leading to the templated replacement of one sequence with the other. CLASS-SWITCH
A3A
A3B
A3D
A3F
A3G
A3H
Chromosome 22 A3C
b
A3E
10 kb
His-X-Glu-X23–28-Pro-Cys-X2–4-Cys
Single-deaminasedomain APOBECs
AID, A1, A2, A3A, A3C, A3D and A3H
RECOMBINATION
(Class or isotype switching). A region-specific recombination process, which occurs in antigen-activated B cells. This occurs between ‘switch region’ DNA sequences and results in a change from the IgM class to IgG, IgA or IgE. This imparts flexibility to the humoral immune response and allows it to exploit the different capacities of the immunoglobulins to activate the appropriate downstream effector mechanisms. RETROVIRAL HYPERMUTATION
The high-frequency accumulation of mutations in a retroviral genome. They are distinguished from reversetranscriptase-dependent mutations in that they are predominantly plus-strand G to A substitutions (C to T in the minus strand).
870
Double-deaminasedomain APOBECs
A3B, A3D—3E, A3F and A3G
c Glu His X Glu
H
H O O
O– Cys Pro X
X
Zn2+
X Glu
O
Zn2+ H
H
O O–
Glu
NH2
Zn2+
O O–
H H
N O
Zn2+ X Cys
O–
H
H O
Cys Pro X Cys
O
His
N
O
O
NH2
N N
Glu O O–
Zn2+ H H
O
NH2 O– N N
NH3
Glu O O–
O H
O
N N
DNA Cytosine
Uracil
Active site of APOBEC
Figure 2 | The human APOBECs. a | A schematic of the human APOBEC genes that depicts their relative chromosomal locations. Activation-induced deaminase (AID) and APOBEC1 (A1) are separated on chromosome 12 by nearly 1 megabase. Single- and double-deaminase-domain proteins are indicated by green and red, respectively. b | A schematic of the domain organization of the human APOBEC proteins, which considers the single- and the double-deaminase-domain proteins separately. The consensus cytosine-deamination motif is indicated above the protein. c | A possible mechanism for APOBEC-dependent cytosine deamination. This model is based on the deamination mechanisms that were predicted from structural studies of the bacterial and yeast cytidine deaminases22–25. First, the cysteine (Cys) and histidine (His) residues of the active site coordinate a single zinc ion (Zn2+); an incoming water molecule (H2O) is shown. Second, a hydroxide ion (OH–) is produced when water reacts with glutamic acid (Glu) and zinc. Only the zinc ion and the side chain of the catalytic glutamic acid are illustrated beyond this step. Third, glutamic acid protonates N3 of the cytosine pyrimidine ring, which destabilizes the N3–C4 double bond and renders C4 susceptible to attack by the hydroxide ion. Fourth, this results in the formation of a tetrahedral (transition state) intermediate, as the proton (H+) of the water molecule is sequestered by glutamic acid. Fifth, the amino group (NH2) receives this proton in a reaction that causes cleavage of the C–N bond and formation of a C4–O double bond. Sixth, uracil and ammonia (NH3 ) are released from the active site.
| NOVEMBER 2004 | VOLUME 4
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS
Box 1 | The APOB mRNA-editing protein APOBEC1 Apolipoprotein B (APOB) mRNA-editing enzyme, catalytic polypeptide 1 (APOBEC1) normally functions as an mRNA-editing protein that deaminates C6666 in mammalian APOB mRNA, which yields a premature stop codon and thereby causes expression of a truncated protein (APOB48 instead of APOB100)70,96,97. At least one other protein known as APOBEC1-complementation factor (ACF), which is probably part of a larger complex, is important for ensuring that APOBEC1 reaches its mRNA target34,35. The unedited larger, and the edited smaller, proteins have distinct roles in lipid transport and metabolism. APOBEC1 can also catalyse C to U deamination in DNA13,20,98. Although this is not likely to be its main physiological function, dysregulated APOBEC1 expression might lead to DNA deamination and possible pathological consequences. Indeed, liverdirected APOBEC1 expression in transgenic mice triggered a high incidence of hepatocellular carcinoma99. Now that programmed DNA deamination has become a well-established mutational mechanism (see main text), it is reasonable to postulate that these tumours were precipitated, in part, by APOBEC1-induced mutations, although non-specific RNA-editing models have not been excluded100. Nevertheless, APOBEC1 has set the precedent for an APOBEC-family member to have several substrate specificities. Carcinogenic roles for some of the other APOBEC-family members have also been proposed13,101,102.
possible. First, based on its mainly cytoplasmic localization18,31–33, it is plausible that some APOBEC3G is simply engulfed by the budding retrovirus. Second, by analogy to APOBEC1, which requires at least one other cellular protein — APOBEC1-complementation factor (ACF) — to reach its physiological mRNA target34,35, the incorporation of APOBEC3G into viral particles might be facilitated by a cellular protein. Third, APOBEC3G might bind a component of the virus itself, providing the most assured route to the virion. However, from the cellular perspective, selfregulation of APOBEC3G, as for APOBEC1, would seem to be the most feasible strategy, as it would be important to protect cellular nucleic acids while targeting the MUTATOR activity of APOBEC3G to appropriate physiological substrates.
HYPERMUTATION
Levels of mutation significantly above spontaneous levels for a given system. Hypermutations are often characterized by specific local sequence preferences (biases). MUTATOR
A protein that actively promotes mutation. This is not to be confused with DNA-repair proteins, which, with the exception of the Y-family of error-prone DNA polymerases, work to discourage mutation.
APOBEC3G Vif APOBEC3G
Vif B
Ub E1
APOBEC3G
Ub E2
Vif B
C
Ub E2
C
CUL5
CUL5
RBX1
RBX1 Ubiquitylation
APOBEC3G
Ub Ub Ub Ub
Vif APOBEC3G
Ub Ub Ub Ub
Ub B
C
E2 CUL5
Proteasomal degradation
RBX1
Figure 3 | The mechanism of Vif-dependent APOBEC3G degradation. Vif (virion infectivity factor) binds APOBEC3G and forms a complex that recruits the cellular proteins elongin B and elongin C (B and C in the figure), which then mediate cullin-5 (CUL5)–ring-box-1 (RBX1)-dependent ligation of ubiquitin (Ub) to APOBEC3G. The E2 ubiquitin-conjugating enzyme receives its ubiquitin from an E1 enzyme. Ubiquitylated APOBEC3G is subsequently degraded by the proteasome. See recent reviews on Vif for further details55–57.
NATURE REVIEWS | IMMUNOLOGY
Several recent studies have begun to address this important area31,36,37. These groups used co-immunoprecipitation studies to show that the Gag (groupspecific antigen) protein of HIV-1 can interact with APOBEC3G. Two of the groups mapped this interaction to the nucleocapsid region of Gag36,37, whereas the third group showed that this region was dispensable for the incorporation of APOBEC3G into virus-like particles31. The studies further disagree with each other in that Patak and colleagues showed that the Gag–APOBEC3G interaction was sensitive to RNase A31, whereas Kleiman and colleagues showed that it was not 37. Both Gag and APOBEC3G had previously been shown to bind RNA10,38,39. So, it is plausible (in line with the results of Patak and colleagues31) that the Gag–APOBEC3G interaction is not direct but occurs, instead, through an RNA-containing complex. Whether this is the sole route used by APOBEC proteins to gain access to the virion has yet to be determined. This puzzle is further compounded by the fact that human APOBEC3G can restrict both human and mouse retroviruses16,18, whereas the mouse APOBEC3 protein can restrict HIV-1 but seems to be incapable of being incorporated into, or restricting, murine leukaemia virus (MLV)40–44. So, if either a Gag- or RNA-specific incorporation route is exploited by APOBEC3G, then MLV somehow avoids the mouse (but not the human) protein. Whatever the mechanism (or mechanisms) of APOBEC3G incorporation, it must accommodate the fact that, in many cases, APOBEC3G can cross inter-species barriers to restrict retroviruses40,44. APOBEC3G destruction is mediated by Vif
As described earlier, the inhibitory effect of APOBEC3G manifests most clearly in the absence of Vif (FIG. 1). Nearly a decade has passed since Vif was shown to be essential for HIV-1 infection45,46. The discovery of APOBEC3G9 has now opened the door to molecular studies showing that HIV-1 Vif functions by triggering the polyubiquitylation and proteasomal degradation of human APOBEC3G before virion incorporation32,33,47–49. This is accomplished by a UBIQUITIN-LIGASE complex that contains Vif and several cellular proteins such as elongin B and elongin C, cullin-5 (CUL5) and ring-box-1 (RBX1)49 (FIG. 3). Vif might also function to impede APOBEC3G protein synthesis, but this possible role is less well-defined32,50. Further mechanistic information has been obtained from inter-species comparisons. For example, although Vif from several different simian immunodeficiency viruses (SIVs) triggers the degradation of its host’s APOBEC3G, the same proteins fail to neutralize human APOBEC3G40. Similarly, HIV-1 Vif is unable to trigger the degradation of the corresponding simian APOBEC3G proteins. Remarkably, this apparent species specificity has been mapped to a single amino acid21,51–53. Substituting the negatively charged aspartic acid residue at position 128 of human APOBEC3G with the corresponding positively charged lysine residue from the simian protein simultaneously exchanged the species specificity, rendering the protein resistant to
VOLUME 4 | NOVEMBER 2004 | 8 7 1 ©2004 Nature Publishing Group
REVIEWS
Box 2 | Harnessing retroviral restriction as a therapeutic approach The functional interaction APOBEC Vif shield between Vif (virion Vif inhibito APOBEC infectivity factor) and the r APOBEC proteins is likely to be delicately balanced such that even minor disturbances Non-permissive Permissive for HIV for HIV could influence the outcome of an infection (see figure). It might even be possible to take advantage of this balance therapeutically. Possible strategies can be grouped into two categories.
Vif inhibitors Vif-binding compounds might be able to directly prevent Vif from functioning. This would presumably leave the cellular APOBEC proteins free to restrict infection. However, similar to anti-HIV-1 therapies that are directed towards viral reverse transcriptases or proteases, this approach might eventually succumb to viral ‘escape’ mutants. The intrinsically high level of genetic variation in a retroviral population (even without APOBECs) would probably undermine this approach by creating Vif variants that would no longer be bound by the inhibitors. However, in combination with other anti-retroviral drugs, such a compound would fortify the pharmaceutical anti-HIV-1 arsenal and further reduce the possibility of viral relapse.
APOBEC shields It has been argued that retroviruses such as HIV-1 are on the edge of a genetic abyss, with a mutation load so high that, if pushed higher, it might drive the virus to extinction103. Retroviral hypermutation by APOBEC3G has resulted in 10- to 1000-fold increases in the viral mutation load in model cell-culture systems, showing that it is possible to render the viral nucleic acid genetically inert through APOBEC-dependent deamination. In vivo, it might be possible to facilitate restriction by using a ‘molecular shield’, a compound that would protect the APOBECs from Vif but not interfere with their antiretroviral activities through cytosine deamination. Such an approach might have the advantage over Vif inhibitors, because the molecular shield would be targeted to a comparatively stable cellular target.
HIV-1 Vif and sensitive to SIV Vif 51–54. The simian APOBEC3G reacted in a corresponding manner. These studies have contributed to identifying the crucial regions of APOBEC3G that can influence Vif function. Several reviews that focus on Vif have been published recently and are recommended for further reading55–57. Retroviral restriction by APOBEC3B and -3F
UBIQUITIN LIGASE
An enzyme (E2 or E3) that catalyses the transfer of ubiquitin — an ~80-residue protein that is highly conserved among all living organisms — to a specific target protein, thereby modifying its function or marking it for degradation by the proteasome.
872
The human genome encodes at least ten other proteins that have strong similarities to APOBEC3G10 (FIG. 2). Seven of these (APOBECs 3A–3F and APOBEC3H) are encoded by genes that are located close to the APOBEC3G gene on chromosome 22q13.1. Strong similarities in the amino-acid sequences of these proteins indicated that these homologues might also be capable of restricting HIV-1 infection. Indeed, recent studies have shown that two other double-deaminasedomain APOBEC-family members — APOBEC3B and APOBEC3F — can also strongly inhibit retroviral infection41–43,58. The APOBEC3B protein is ~50% similar to APOBEC3G 10. Recently, Bishop and colleagues showed that APOBEC3B also has antiretroviral activity, which limits the infectivity of HIV-1- but not MLV-based retroviral substrates 41. By contrast, APOBEC3G readily restricted both types of retrovirus16,18. Moreover, APOBEC3B triggered retroviral
| NOVEMBER 2004 | VOLUME 4
cDNA deamination in a dinucleotide context (5′-TC, in which the underlined residue is deaminated)41 distinct from deamination by APOBEC3G (discussed next). The activity of APOBEC3B was not affected by Vif expression41 (discussed later). APOBEC3F is located adjacent to APOBEC3G on human chromosome 22, and these genes are more than 90% identical in their promoter regions10,58. These identities extend well into the coding sequences, as 59 out of 60 of their N-terminal amino acids are identical. Similar to APOBEC3G, APOBEC3F has shown strong DNA-editing activity in an Escherichia coli mutation assay58 and in model retroviral infectivity assays41–43,58. However, APOBEC3F differs from APOBEC3G in two important respects. First and most notably, similar to APOBEC3B, APOBEC3F triggers cytosine deamination in a distinct dinucleotide sequence, 5′-TC41,42,58, whereas APOBEC3G shows nearly exclusive preference for 5′-CC16,18,19,21. In both cases, the resulting uracils template the incorporation of adenines, triggering minus-strand transition mutations from 5′-TC to TT and 5′-CC to CT, respectively. On the plus strand of the retrovirus, these mutations manifest as 5′-GA to AA and 5′-GG to AG hypermutations. These dinucleotides dominate the hypermutation patterns that are observed in patientderived HIV-1 samples15,59. Second, similar to APOBEC3B but to a lesser extent, APOBEC3F is resistant to HIV-1 Vif 58. This property is consistent with APOBEC3B and/or APOBEC3F being more important than APOBEC3G in restricting HIV-1 in vivo, because GA to AA hypermutation is often more common than GG to AG hypermutation in patientderived retroviral DNA sequences15. Vif resistance might be partially explained by the amino-acid residues surrounding the APOBEC3B and APOBEC3F regions that correspond to the crucial Vif-interacting region (as defined by Asp128 ) of APOBEC3G58. In APOBEC3G, Asp128 lies in a Trp-Asp128-Pro-Asp motif, whereas the corresponding residue in APOBEC3B and APOBEC3F (Glu127) lies in a Trp-Glu127-Arg-Asp motif. It is probable that the positive charge of the arginine residue neutralizes the overall charge of this region and thereby renders the proteins resistant to HIV-1 Vif. Although local charge is a probable modifier, it is clearly not the only factor that controls the APOBEC–Vif interaction because APOBEC3B seems to be completely resistant to HIV-1 Vif 41, whereas APOBEC3F is only partially resistant58. So, although the precise molecular details of the Vif-resistance properties of APOBEC3B and APOBEC3F await experimental tests, it is clear that such variations in the human population could readily contribute to variations in HIV-1 resistance. Moreover, further studies along these mechanistic fronts will be crucial for identifying the dominant restrictors in vivo and for facilitating pharmaceutical intervention by delineating the key domains that are involved in this interaction (BOX 2). However, when these observations are taken together with the facts that APOBEC3F and APOBEC3G seem to be co-expressed42,58 and that they have significant sequence identity at the promoter level10,58, it is reasonable to hypothesize that these two proteins function
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS
Table 1 | The human APOBEC-family members Name*
Genomic position
Expression profile
Known editing activities
Probable physiological function
References
AID
12p13
Activated B cells, lower levels in other tissues
DNA deaminase‡
Immunoglobulin-gene diversification
63,79
APOBEC1
12p13.1
Gastrointestinal tissues
RNA or DNA deaminase
APOB mRNA editing
13,76
APOBEC2
6p21
Heart and skeletal muscle
Unknown
Unknown
65,66
APOBEC3A
22q13.1
Keratinocytes
Unknown
Unknown
APOBEC3B
22q13.1
Peripheral-blood cells, T cells and keratinocytes (not co-expressed with APOBEC3F or APOBEC3G)
DNA deaminase (with minor RNA-editing activity)
Retroviral cDNA editing
10,41
APOBEC3C
22q13.1
Many tissues and a variety of cancer cell lines
DNA deaminase
Unknown
10,70
APOBEC3D
22q13.1
Unknown
None
Unknown
10,13,70
APOBEC3D–3E
22q13.1
Unknown
None
Unknown
10,13,70
APOBEC3E
22q13.1
Unknown, probably a pseudogene
None
Unknown
APOBEC3F
22q13.1
Many tissues and probably co-expressed with APOBEC3G
DNA deaminase (with minor RNA-deaminase activity)
Retroviral cDNA editing
41,42,58
APOBEC3G
22q13.1
Many tissues and probably co-expressed with APOBEC3F
DNA deaminase
Retroviral cDNA editing
9,13,42,58
106
13,70
APOBEC3H
22q13.1
Unknown
Unknown
Unknown
70
APOBEC4
12q23
Unknown
Unknown
Unknown
70
*Many different names for these proteins have been used10,70. We have used the APOBEC (apolipoprotein B mRNA-editing enzyme, catalytic polypeptide) nomenclature simply as a name in and of itself and for the historical reason that the founding member of this class was APOBEC1 (REF. 11) (BOX 1). ‡An alternative model for immunoglobulin-gene diversification contends that activation-induced deaminase (AID) functions as an RNA editor4. This model is less probable than the DNA-deamination mechanism that is highlighted here because AID has not elicited an RNA-editing activity (even in direct biochemical comparisons with well-deaminated DNA substrates) and no mRNA substrate has been identified. See the main text for further discussion.
together to accomplish HIV-1 restriction in vivo58. Indeed, APOBEC3F and APOBEC3G have been shown to interact by co-immunoprecipitation42 and to synergize when co-expressed58. It is also paramount to consider the fact that many of the APOBEC-family members (including APOBEC3F and APOBEC3G42,58) have broad expression profiles that are indicative of these proteins having other functions (TABLE 1). The recent demonstrations that APOBEC3F and APOBEC3G hinder replication of hepatitis B virus60–62 could become the first of many other examples in which APOBEC-family members function to block the spread of foreign nucleic acids. This further highlights the need to fully understand the activities of the APOBEC proteins, including other possible mechanisms that the APOBECs might use to inhibit viral infection. Other APOBECs and (retro)viral restriction
The genes of five other human APOBEC3-family members are located near APOBEC3B, -3F and -3G on chromosome 22 (FIG. 2; TABLE 1). Three other APOBEC-family genes — which encode AID, APOBEC1 and APOBEC2 — have other genomic locations10,63–66 (FIG. 2; TABLE 1). Curiously, all of these other human APOBEC-family members (except APOBEC3D–3E) encode bona fide or predicted single-deaminase-domain proteins10–13,65–67, but none has shown antiretroviral activity so far 41–43. However, the rat homologue of APOBEC1 was recently shown to restrict HIV-1 regardless of Vif 41,68. In contrast to APOBEC3B, -3F and -3G — which are all clearly capable of potent retroviral cDNA cytosine deamination — rat APOBEC1 also triggered the deamination
NATURE REVIEWS | IMMUNOLOGY
of a large number of cytosine residues in HIV-1 genomic RNA68. Expression of the human APOBEC1 protein seemed not to mutate HIV-1 DNA or RNA41–43,68. The difference between the human and rat studies might therefore indicate that when APOBEC1 is expressed out of its normal context (for example, a rat protein in human cells), it might become capable of manifesting a broader functionality. Perhaps rat APOBEC1 is escaping regulation by rat ACF34,35 in this context. Alternatively, the cytosine-deaminase activity might be intrinsic to the rat (but not the human) APOBEC1 protein, as indicated by observations showing that the rat (but not the human) APOBEC1 protein can trigger C/G to T/A transitions in E. coli 13,69. Even if this surprising rat APOBEC1 result is not of direct physiological relevance to HIV-1, it will certainly help to broaden the scope of the search for further targets of APOBEC-family members. Some APOBEC-family members probably function preferentially on RNA substrates (such as RNA viruses or cellular mRNA), whereas others will prefer DNA substrates (such as DNA viruses or nuclear genes). Moreover, crossspecies comparisons such as this will continue to prove informative as they might allow the examination of APOBEC activities in the absence of potential negative regulators. APOBEC evolution
APOBEC-family members are found throughout the vertebrate lineage (FIG. 4). Humans and chimpanzees (and probably other non-human primates), for example, have at least eleven APOBEC genes: AID, APOBEC1, APOBEC2 and APOBECs 3A–3H 10,70. However, not far down the vertebrate tree, a precipitous decline in the
VOLUME 4 | NOVEMBER 2004 | 8 7 3 ©2004 Nature Publishing Group
REVIEWS number of APOBEC3-family members occurs, with rodents such as mice and rats encoding only one70. Mouse APOBEC3 is located on chomosome 15e1, which is syntenic to the APOBEC3 genes in humans (chromosome 22q13.1) and chimpanzees (chromosome 23). This indicates that a relatively recent and possibly unprecedented gene expansion has occurred. Although the strong homologies between the primate APOBEC3family members imply that this probably occurred by a series of gene duplications, the precise mechanism (or mechanisms) and selective pressures that drove the APOBEC3 gene expansion are unclear. These gross observations have been supported by studies on the genetic variability of primate APOBECfamily members71,72. Inter-species comparisons have revealed that APOBEC3G and most of the other APOBEC-family members show more amino-acidaltering (non-synonymous) base-substitution mutations than they show synonymous base-substitution mutations71,72. This phenomenon indicates that the APOBECfamily members have been under a positive-selection
Primates such as humans
Single-deaminase-domain APOBECs
Single- and double-deaminase-domain APOBECs
AID, A1, A2 and A3A–A3H
Artiodactyls such as cows AID, A1, A2 and A3
96 MY
Mammals
90 MY
Rodents such as mice AID, A1, A2 and A3
Birds such as chickens AID and A2 310 MY
Fish such as zebrafish AID and A2
425 MY
Figure 4 | APOBECs in vertebrates. A schematic depicting the phylogenetic relationships between several vertebrates and their APOBEC-family members. The timeline in millions of years (MY) was derived from studies by Nei and Glazko105 and is not illustrated to scale. The relative positioning of artiodactyls and rodents is controversial. However, in support of the arrangement shown, our sequence-database analyses have indicated that artiodactyls have more than one APOBEC3 (A3)-family member. AID, activation-induced deaminase.
874
| NOVEMBER 2004 | VOLUME 4
pressure, which has facilitated the rapid fixation of amino-acid changes. Sawyer and co-workers suggested that this selection pressure might be related to the possibility that the APOBECs function to prevent human endogenous retroviruses (HERVs) from causing wide-spread genomic instability72. Further examination of the APOBEC-family members that are present in the vertebrate lineage reveals another key transition point — from organisms containing singledeaminase-domain proteins to those containing both single- and double-deaminase-domain proteins (FIG. 4). The ancestral single-deaminase-domain duplication or fusion must have occurred after the divergence of birds, which contain only the single-deaminase-domain proteins AID73,74 and APOBEC2. This is supported by the fact that only AID- and APOBEC2-like sequences occur in the genomic sequences of more-primitive vertebrates such as zebrafish and pufferfish (FIG. 4). So, at present, the double-deaminase-domain proteins have been identified only in mammals. As retroviruses are not specific to mammals, this further indicates that APOBEC proteins such as APOBEC3G are multifunctional, capable not only of retroviral defence but possibly also of other important (protective) roles. Consistent with a possible role in reproduction, APOBEC3G expression has been observed in both human breast13 and placental42,58 tissues. It is curious that the most ancient APOBEC-family members are AID and APOBEC2, of which the former is responsible for triggering the vertebrate-specific immunoglobulin gene-diversification reactions and the function of the latter is still unknown (FIG. 4; TABLE 1). AID is slightly more conserved, showing 70–74% similarity between the human and fish proteins75, whereas APOBEC2 is 68–70% similar. Despite its interesting characteristics — the cytosine-deaminase motif 65,66, the cardiac-specific and smooth-muscle-specific expression profile65,66, its self-interaction capacity10 and its apparent ubiquitous presence in vertebrates — APOBEC2 has shown little enzymatic activity so far. However, partially purified APOBEC2 was able to bind RNA and displayed a weak cytidine-deaminase activity compared with the E. coli cytidine deaminase65,66. The latter observation indicated that a free base might not be its physiological substrate. This enigmatic protein has failed to show cytosine-deaminase activity with DNA13, RNA66 or retroviral substrates43. Perhaps APOBEC2 has an as-yet-unappreciated role in the provision of innate immunity, which might pre-date adaptive immunity. Whatever its function, by analogy to AID, it will probably be conserved throughout the vertebrate lineage. Therefore, as is the case for APOBEC1 and AID76,77, mice that lack APOBEC2 or APOBEC3 will be expected to greatly enrich our understanding of the physiological roles of the APOBEC-family members. In the light of evolution, we anticipate that the AID and APOBEC2 proteins of fish and other more primitive vertebrates will be multifunctional and will have the capacity to cause immunoglobulin-gene diversification, as well as the ability to use cytosine deamination as a direct defence against endogenous and exogenous invasive nucleic acids.
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS
Functional immunoglobulin gene
VDJ C
Sµ C
Cµ
Sγ C
AID
AID
AID
U
U C
U
Cγ
ΨV
Gene conversion
Somatic hypermutation
Class-switch recombination
Figure 5 | Programmed immunoglobulin-gene diversification by AID. Activationinduced deaminase (AID)-dependent cytosine deamination of functional immunoglobulin genes triggers gene conversion, somatic hypermutation and class-switch recombination. Uracil recognition and processing by the cellular uracil DNA glycosylase UNG2 is a pivotal step in these processes. V, D, J, S and C denote the immunoglobulin-gene variable, diversity, joining, switch and constant regions, respectively. See the main text for further details.
Immunoglobulin-gene diversification by AID
APOBEC-mediated cellular defences have several striking similarities to the vertebrate antibody affinitymaturation mechanism. Antibodies are an essential component of the vertebrate adaptive immune response to many pathogens, including viruses. At the molecular level, various DNA manipulations enable the formation of high-affinity antibodies of several classes (reviewed in REFS 4,5). These DNA-level manipulations, which are collectively known as immunoglobulin gene-diversification events, are summarized in FIG. 5. In response to antigen, the variable (antigen-binding) regions of the functional immunoglobulin genes of most vertebrates are diversified by base-substitution mutations. This process is known as somatic hypermutation. In some vertebrates such as birds, variable-region diversification is accomplished by gene conversion. Variable-region diversification coupled with repeated encounters between the antibody-expressing B cell and the antigen (selection) can produce higher-affinity antibodies (affinity maturation). Also, in response to antigen, the constant regions of functional immunoglobulin genes can be replaced by an alternative downstream constant region, which effectively changes the effector function of the antibody in a process known as class-switch recombination. Although somatic hypermutation, gene conversion and class-switch recombination seem, overtly, to be very different, they are linked by an absolute requirement for AID73,74,77,78 (reviewed in REFS 4,5). Honjo and colleagues isolated AID as a factor that was upregulated in cells that had been induced to undergo immunoglobulin-gene class-switch recombination79. AID was subsequently shown to be required for class switching and, surprisingly, also for somatic hypermutation in mice and humans77,78. Gene conversion was also shown to require AID73,74. This universal requirement indicated that these three processes might share a common molecular step. Neuberger and colleagues proposed a unifying model for immunoglobulin-gene diversification in which AID triggers all three of these distinct processes
NATURE REVIEWS | IMMUNOLOGY
through a common initiating lesion — DNA cytosine deamination12 (FIG. 5). In support of such a mechanism, AID expression triggered C/G to T/A transitions in E. coli DNA through a uracil intermediate12. Several model genetic and biochemical systems have since been used to show that the C to U DNA-deaminase activity of AID is specific for single-stranded DNA20,67,80–84. Moreover, the cellular uracil DNA glycosylase UNG2, which is a well-characterized base-excision-repair enzyme that functions to specifically remove uracil from single- or double-stranded DNA85, is also required for all three processes86–89. UNG2 is essential for class switching and aspects of somatic hypermutation in humans (shown by recessive mutations) and mice (shown by gene targeting), and it is required for somatic hypermutation and gene conversion in immortalized chicken B cells (shown by enzymatic inhibition)86–89. So, somatic hypermutation, gene conversion and class switching all require AID and UNG2. Precisely how the processing of uracil lesions results in these three distinct immunoglobulin gene-diversification events will require further experimentation. The fate of the uracils will probably be influenced by where they reside in the DNA (variable or switch region, plus or minus strand) and how they are processed (FIG. 5). Hypermutation might, in part, be due to trans-lesion DNA synthesis90,91, whereas gene conversion and class switching will clearly require some form of DNA breakage92,93. Recent work by Honjo and colleagues (who contend, instead, that AID functions as an RNA editor; see TABLE 1 footnote) has confirmed the requirement for UNG2 in class-switching but also revealed an interesting further twist 94. DNA double-stranded breaks in the immunoglobulin heavy-chain locus, but not class-switch recombination, seemed to occur in the absence of UNG2 uracil-excision activity, which seems to show that UNG2 has a role in class-switch recombination downstream of DNA cleavage. This observation can however be reconciled with the DNAdeamination model, as all of the UNG2 variants used in this study were still able to bind DNA without excising the uracil. Base-excision repair is thought to be a coordinated reaction in which UNG2 ‘passes’ the lesion to the next required repair protein, which is most often an endonuclease known as APEX (apurinic/ apyrimidinic endonuclease). Therefore, the UNG2 variants used by Honjo and colleagues might still be capable of recruiting an endonuclease such as APEX, which could then trigger DNA double-stranded breaks. Without doubt, further work on AID and immunoglobulin-gene diversification will fuel studies on the mechanism of APOBEC3-dependent retroviral restriction and vice versa. Conclusions
Programmed cytosine deamination is proving to be an integral part of the innate and adaptive immune responses, with several APOBECs able to restrict retroviruses and AID able to trigger immunoglobulin gene-diversification events. Further examples of deliberate nucleic-acid modification by APOBEC-family
VOLUME 4 | NOVEMBER 2004 | 8 7 5 ©2004 Nature Publishing Group
REVIEWS members will probably be described in the near future, as several of these proteins have yet to be assigned a physiological function (TABLE 1). Moreover, many lines of evidence have indicated that APOBEC-family members might have other functions in addition to their main physiological roles (for example, APOB mRNA editing70 and viral RNA editing68 by APOBEC1) (BOX 1). Both DNA and RNA substrates are possible cytosinedeamination targets, and such substrate selections are likely to be influenced by, and possibly even dictated by, many other cellular and non-cellular (viral) factors. Furthermore, common regulatory schemes such as ubiquitin-mediated degradation pathways might also be anticipated given the significant amino-acid similarities between APOBEC-family members. APOBEC3F and APOBEC3G underpin a new mechanism of cellular immunity to retroviruses such as HIV-1. The fact that the hypermutation signatures of these proteins are apparent in patient-derived HIV-1 DNA sequences indicates that, at least some of the time,
1.
2. 3. 4. 5. 6.
7. 8. 9.
10.
11.
12.
13.
14.
15.
876
Dobzhansky, T. Nothing in biology makes sense except in the light of evolution. Am. Biol. Teach. 35, 125–129 (1973). Beutler, B. & Hoffmann, J. Innate immunity. Curr. Opin. Immunol. 16, 1–3 (2004). Samuel, C. E. Antiviral actions of interferons. Clin. Microbiol. Rev. 14, 778–809 (2001). Honjo, T., Muramatsu, M. & Fagarasan, S. AID: how does it aid antibody diversity? Immunity 20, 659–668 (2004). Neuberger, M. S. et al. Immunity through DNA deamination. Trends Biochem. Sci. 28, 305–312 (2003). Barre-Sinoussi, F. et al. Isolation of a T-lymphotropic retrovirus from a patient at risk for acquired immune deficiency syndrome (AIDS). Science 220, 868–871 (1983). Gallo, R. C. Historical essay. The early years of HIV/AIDS. Science 298, 1728–1730 (2002). Montagnier, L. Historical essay. A history of HIV discovery. Science 298, 1727–1728 (2002). Sheehy, A. M. et al. Isolation of a human gene that inhibits HIV-1 infection and is suppressed by the viral Vif protein. Nature 418, 646–650 (2002). A landmark study identifying human APOBEC3G as a potent inhibitor of Vif-deficient HIV-1 replication and as a functional target of the HIV-1 Vif protein. Jarmuz, A. et al. An anthropoid-specific locus of orphan C to U RNA-editing enzymes on chromosome 22. Genomics 79, 285–296 (2002). Teng, B., Burant, C. F. & Davidson, N. O. Molecular cloning of an apolipoprotein B messenger RNA editing protein. Science 260, 1816–1819 (1993). Petersen-Mahrt, S. K., Harris, R. S. & Neuberger, M. S. AID mutates E. coli suggesting a DNA deamination mechanism for antibody diversification. Nature 418, 99–103 (2002). This paper proposed the DNA-deamination model for AID-mediated immunoglobin-gene diversification and showed that AID can trigger C/G to T/A transition mutations through a uracil intermediate. Harris, R. S., Petersen-Mahrt, S. K. & Neuberger, M. S. RNA editing enzyme APOBEC1 and some of its homologs can act as DNA mutators. Mol. Cell 10, 1247–1253 (2002). This paper was the first to show that APOBEC3G was capable of DNA cytosine deamination. Pathak, V. K. & Temin, H. M. Broad spectrum of in vivo forward mutations, hypermutations, and mutational hotspots in a retroviral shuttle vector after a single replication cycle: substitutions, frameshifts, and hypermutations. Proc. Natl Acad. Sci. USA 87, 6019–6023 (1990). Janini, M. et al. Human immunodeficiency virus type 1 DNA sequences genetically damaged by hypermutation are often abundant in patient peripheral blood mononuclear cells and may be generated during near-simultaneous infection and activation of CD4+ T cells. J. Virol. 75, 7973–7986 (2001). This paper provided a systematic analysis and discussion of HIV-1 G to A hypermutation in patientderived blood cells. The HIV-1 G to A hypermutations occurred exclusively in GA and GG dinucleotide contexts, with GA predominating.
the viral counterdefence by Vif can be thwarted58. Population-level variations in APOBEC3F, APOBEC3G and Vif are likely to influence both the outcome of an infection and the development of viral drug resistance. Addressing the APOBEC–Vif interaction pharmaceutically might one day fortify the arsenal of antiretroviral drugs (BOX 2). Retroviruses are not specific to humans or to organisms in which the double-deaminase-domain APOBECs occur. Given the potential potency of the APOBEC3F- and APOBEC3G-dependent retroviralrestriction mechanism and the overall high degree of conservation of the APOBEC-protein family in vertebrates, it is probable that the single-deaminase-domain proteins, AID and APOBEC2, have a similar function in vertebrates such as birds and fish. Many other robust mechanisms for preventing the spread of foreign and endogenous nucleic acids are likely to exist95; however, at present, it is difficult to imagine any with greater immunological impact.
16. Harris, R. S. et al. DNA deamination mediates innate immunity to retroviral infection. Cell 113, 803–809 (2003). 17. Lecossier, D. et al. Hypermutation of HIV-1 DNA in the absence of the Vif protein. Science 300, 1112 (2003). 18. Mangeat, B. et al. Broad antiretroviral defence by human APOBEC3G through lethal editing of nascent reverse transcripts. Nature 424, 99–103 (2003). 19. Zhang, H. et al. The cytidine deaminase CEM15 induces hypermutation in newly synthesized HIV-1 DNA. Nature 424, 94–98 (2003). References 16–19 showed that APOBEC3G is a potent retroviral cDNA deaminase, capable of triggering high levels of retroviral hypermutation. Reference 16 was also the first to report the singlestrand-specific DNA-cytosine-deaminase activity of APOBEC3G in vitro. 20. Beale, R. C. et al. Comparison of the differential contextdependence of DNA deamination by APOBEC enzymes: correlation with mutation spectra in vivo. J. Mol. Biol. 337, 585–596 (2004). 21. Yu, Q. et al. Single-strand specificity of APOBEC3G accounts for minus-strand deamination of the HIV genome. Nature Struct. Mol. Biol. 11, 435–442 (2004). 22. Betts, L. et al. Cytidine deaminase. The 2.3 Å crystal structure of an enzyme: transition-state analog complex. J. Mol. Biol. 235, 635–656 (1994). 23. Johansson, E. et al. Crystal structure of the tetrameric cytidine deaminase from Bacillus subtilis at 2.0 Å resolution. Biochemistry 41, 2563–2570 (2002). 24. Ko, T. P. et al. Crystal structure of yeast cytosine deaminase. Insights into enzyme mechanism and evolution. J. Biol. Chem. 278, 19111–19117 (2003). 25. Xie, K. et al. The structure of a yeast RNA-editing deaminase provides insight into the fold and function of activationinduced deaminase and APOBEC-1. Proc. Natl Acad. Sci. USA 101, 8114–8119 (2004). 26. Shindo, K. et al. The enzymatic activity of CEM15/Apobec3G is essential for the regulation of the infectivity of HIV-1 virion but not a sole determinant of its antiviral activity. J. Biol. Chem. 278, 44412–44416 (2003). 27. MacGinnitie, A. J., Anant, S. & Davidson, N. O. Mutagenesis of apobec-1, the catalytic subunit of the mammalian apolipoprotein B mRNA editing enzyme, reveals distinct domains that mediate cytosine nucleoside deaminase, RNA binding, and RNA editing activity. J. Biol. Chem. 270, 14768–14775 (1995). 28. Navaratnam, N. et al. Evolutionary origins of apoB mRNA editing: catalysis by a cytidine deaminase that has acquired a novel RNA-binding motif at its active site. Cell 81, 187–195 (1995). 29. Navaratnam, N. et al. Escherichia coli cytidine deaminase provides a molecular model for apoB RNA editing and a mechanism for RNA substrate recognition. J. Mol. Biol. 275, 695–714 (1998). 30. Ta, V. T. et al. AID mutant analyses indicate requirement for class-switch-specific cofactors. Nature Immunol. 4, 843–848 (2003).
| NOVEMBER 2004 | VOLUME 4
31. Svarovskaia, E. S. et al. Human apolipoprotein B mRNAediting enzyme-catalytic polypeptide-like 3G (APOBEC3G) is incorporated into HIV-1 virions through interactions with viral and nonviral RNAs. J. Biol. Chem. 279, 35822–35828 (2004). This paper and references 36 and 37 are beginning to shed light on the intriguing mechanism of how APOBEC3G is incorporated into the HIV-1 virion, through Gag, RNA and/or a Gag–RNA complex. 32. Stopak, K. et al. HIV-1 Vif blocks the antiviral activity of APOBEC3G by impairing both its translation and intracellular stability. Mol. Cell 12, 591–601 (2003). 33. Marin, M. et al. HIV-1 Vif protein binds the editing enzyme APOBEC3G and induces its degradation. Nature Med. 9, 1398–1403 (2003). 34. Lellek, H. et al. Purification and molecular cloning of a novel essential component of the apolipoprotein B mRNA editing enzyme-complex. J. Biol. Chem. 275, 19848–19856 (2000). 35. Mehta, A. et al. Molecular cloning of Apobec-1 complementation factor, a novel RNA-binding protein involved in the editing of apolipoprotein B mRNA. Mol. Cell. Biol. 20, 1846–1854 (2000). 36. Alce, T. M. & Popik, W. APOBEC3G is incorporated into virus-like particles by a direct interaction with HIV-1 Gag nucleocapsid protein. J. Biol. Chem. 279, 34083–34086 (2004). 37. Cen, S. et al. The interaction between HIV-1 Gag and APOBEC3G. J. Biol. Chem. 279, 33177–33184 (2004). 38. Li, J., Potash, M. J. & Volsky, D. J. Functional domains of APOBEC3G required for antiviral activity. J. Cell. Biochem. 92, 560–572 (2004). 39. Berkowitz, R. D. et al. Retroviral nucleocapsid domains mediate the specific recognition of genomic viral RNAs by chimeric Gag polyproteins during RNA packaging in vivo. J. Virol. 69, 6445–6456 (1995). 40. Mariani, R. et al. Species-specific exclusion of APOBEC3G from HIV-1 virions by Vif. Cell 114, 21–31 (2003). 41. Bishop, K. N. et al. Cytidine deamination of retroviral DNA by diverse APOBEC proteins. Curr. Biol. 14, 1392–1396 (2004). 42. Wiegand, H. L. et al. A second human antiretroviral factor, APOBEC3F, is suppressed by the HIV-1 and HIV-2 Vif proteins. EMBO J. 23, 2451–2458 (2004). 43. Zheng, Y. H. et al. Human APOBEC3F is another host factor that blocks human immunodeficiency virus type 1 replication. J. Virol. 78, 6073–6076 (2004). References 41–43 and 58 have shown that some other APOBEC-family members (in addition to APOBEC3G) can also function to restrict retroviral infection. An important implication is that APOBEC3F might be the dominant restrictor of HIV-1 infection in vivo, with APOBEC3G having a key supporting role. 44. Kobayashi, M. et al. APOBEC3G targets specific virus species. J. Virol. 78, 8238–8244 (2004). 45. Gabuzda, D. H. et al. Role of Vif in replication of human immunodeficiency virus type 1 in CD4+ T lymphocytes. J. Virol. 66, 6489–6495 (1992).
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS 46. von Schwedler, U. et al. Vif is crucial for human immunodeficiency virus type 1 proviral DNA synthesis in infected cells. J. Virol. 67, 4945–4955 (1993). 47. Conticello, S. G., Harris, R. S. & Neuberger, M. S. The Vif protein of HIV triggers degradation of the human antiretroviral DNA deaminase APOBEC3G. Curr. Biol. 13, 2009–2013 (2003). 48. Sheehy, A. M., Gaddis, N. C. & Malim, M. H. The antiretroviral enzyme APOBEC3G is degraded by the proteasome in response to HIV-1 Vif. Nature Med. 9, 1404–1407 (2003). 49. Yu, X. et al. Induction of APOBEC3G ubiquitination and degradation by an HIV-1 Vif–Cul5–SCF complex. Science 302, 1056–1060 (2003). This paper revealed that HIV-1 Vif functions to recruit a ubiquitin-ligase complex to accomplish the degradation of APOBEC3G. 50. Kao, S. et al. The human immunodeficiency virus type 1 Vif protein reduces intracellular expression and inhibits packaging of APOBEC3G (CEM15), a cellular inhibitor of virus infectivity. J. Virol. 77, 11398–11407 (2003). 51. Bogerd, H. P. et al. A single amino acid difference in the host APOBEC3G protein controls the primate species specificity of HIV type 1 virion infectivity factor. Proc. Natl Acad. Sci. USA 101, 3770–3774 (2004). 52. Mangeat, B. et al. A single amino acid determinant governs the species-specific sensitivity of APOBEC3G to Vif action. J. Biol. Chem. 279, 14481–14483 (2004). 53. Schrofelbauer, B., Chen, D. & Landau, N. R. A single amino acid of APOBEC3G controls its species-specific interaction with virion infectivity factor (Vif). Proc. Natl Acad. Sci. USA 101, 3927–3932 (2004). 54. Xu, H. et al. A single amino acid substitution in human APOBEC3G antiretroviral enzyme confers resistance to HIV-1 virion infectivity factor-induced depletion. Proc. Natl Acad. Sci. USA 101, 5652–5657 (2004). 55. Schrofelbauer, B., Yu, Q. & Landau, N. R. New insights into the role of Vif in HIV-1 replication. AIDS Rev. 6, 34–39 (2004). 56. Navarro, F. & Landau, N. R. Recent insights into HIV-1 Vif. Curr. Opin. Immunol. 16, 477–482 (2004). 57. Rose, K. M. et al. The viral infectivity factor (Vif) of HIV-1 unveiled. Trends Mol. Med. 10, 291–297 (2004). 58. Liddament, M. T. et al. APOBEC3F properties and hypermutation preferences indicate activity against HIV-1 in vivo. Curr. Biol. 14, 1385–1391 (2004). 59. Vartanian, J. P., Sommer, P. & Wain-Hobson, S. Death and the retrovirus. Trends Mol. Med. 9, 409–413 (2003). 60. Turelli, P. et al. Response to comment on ‘Inhibition of hepatitis B virus replication by APOBEC3G’. Science 305, 1403B (2004). 61. Rosler, C. et al. Comment on ‘Inhibition of hepatitis B virus replication by APOBEC3G’. Science 305, 1403; author reply 1403 (2004). 62. Turelli, P. et al. Inhibition of hepatitis B virus replication by APOBEC3G. Science 303, 1829 (2004). References 60–62 show that APOBEC3G can inhibit replication of HBV, possibly through both deamination-dependent and -independent mechanisms. 63. Muto, T. et al. Isolation, tissue distribution, and chromosomal localization of the human activation-induced cytidine deaminase (AID) gene. Genomics 68, 85–88 (2000). 64. Espinosa, R. et al. Assignment of the gene encoding the human apolipoprotein B mRNA editing enzyme (APOBEC1) to chromosome 12p13.1. Genomics 24, 414–415 (1994). 65. Anant, S. et al. ARCD-1, an apobec-1-related cytidine deaminase, exerts a dominant negative effect on C to U RNA editing. Am. J. Physiol. Cell Physiol. 281, C1904–C1916 (2001). 66. Liao, W. et al. APOBEC-2, a cardiac- and skeletal musclespecific member of the cytidine deaminase supergene family. Biochem. Biophys. Res. Commun. 260, 398–404 (1999). 67. Pham, P. et al. Processive AID-catalysed cytosine deamination on single-stranded DNA simulates somatic hypermutation. Nature 424, 103–107 (2003). 68. Bishop, K. N. et al. APOBEC-mediated editing of viral RNA. Science 305, 645 (2004). This paper has opened the door to the possibility that APOBEC proteins might also function to restrict
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85. 86.
87.
88.
89.
infection by RNA viruses. It showed that rat APOBEC1 can edit HIV-1 RNA, as well as deaminate its cDNA. Ramiro, A. R. et al. Transcription enhances AID-mediated cytidine deamination by exposing single-stranded DNA on the nontemplate strand. Nature Immunol. 4, 452–456 (2003). Wedekind, J. E. et al. Messenger RNA editing in mammals: new members of the APOBEC family seeking roles in the family business. Trends Genet. 19, 207–216 (2003). Zhang, J. & Webb, D. M. Rapid evolution of primate antiviral enzyme APOBEC3G. Hum. Mol. Genet. 13, 1785–1791 (2004). Sawyer, S. L., Emerman, M. & Malik, H. S. Ancient adaptive evolution of the primate antiviral DNA-editing enzyme APOBEC3G. PLoS Biol. 2, e275 (2004). Harris, R. S. et al. AID is essential for immunoglobulin V gene conversion in a cultured B cell line. Curr. Biol. 12, 435–438 (2002). Arakawa, H., Hauschild, J. & Buerstedde, J. M. Requirement of the activation-induced deaminase (AID) gene for immunoglobulin gene conversion. Science 295, 1301–1306 (2002). Saunders, H. L. & Magor, B. G. Cloning and expression of the AID gene in the channel catfish. Dev. Comp. Immunol. 28, 657–663 (2004). Nakamuta, M. et al. Complete phenotypic characterization of Apobec-1 knockout mice with a wild-type genetic background and a human apolipoprotein B transgenic background, and restoration of apolipoprotein B mRNA editing by somatic gene transfer of Apobec-1. J. Biol. Chem. 271, 25981–25988 (1996). Muramatsu, M. et al. Class switch recombination and hypermutation require activation-induced cytidine deaminase (AID), a potential RNA editing enzyme. Cell 102, 553–563 (2000). Revy, P. et al. Activation-induced cytidine deaminase (AID) deficiency causes the autosomal recessive form of the hyper-IgM syndrome (HIGM2). Cell 102, 565–575 (2000). References 77 and 78 show that AID is required for immunoglobulin-gene class-switch recombination and somatic hypermutation. Muramatsu, M. et al. Specific expression of activationinduced cytidine deaminase (AID), a novel member of the RNA-editing deaminase family in germinal center B cells. J. Biol. Chem. 274, 18470–18476 (1999). A landmark paper that identifies AID as a B-cellspecific factor, which could be upregulated in cells induced to undergo class-switch recombination. Yu, K., Huang, F. T. & Lieber, M. R. DNA substrate length and surrounding sequence affect the activation-induced deaminase activity at cytidine. J. Biol. Chem. 279, 6496–6500 (2004). Sohail, A. et al. Human activation-induced cytidine deaminase causes transcription-dependent, strand-biased C to U deaminations. Nucleic Acids Res. 31, 2990–2994 (2003). Dickerson, S. K. et al. AID mediates hypermutation by deaminating single stranded DNA. J. Exp. Med. 197, 1291–1296 (2003). Chaudhuri, J. et al. Transcription-targeted DNA deamination by the AID antibody diversification enzyme. Nature 422, 726–730 (2003). Bransteitter, R. et al. Activation-induced cytidine deaminase deaminates deoxycytidine on single-stranded DNA but requires the action of RNase. Proc. Natl Acad. Sci. USA 100, 4102–4107 (2003). Lindahl, T. & Wood, R. D. Quality control by DNA repair. Science 286, 1897–1905 (1999). Imai, K. et al. Human uracil-DNA glycosylase deficiency associated with profoundly impaired immunoglobulin classswitch recombination. Nature Immunol. 4, 1023–1028 (2003). Rada, C. et al. Immunoglobulin isotype switching is inhibited and somatic hypermutation perturbed in UNG-deficient mice. Curr. Biol. 12, 1748–1755 (2002). Di Noia, J. & Neuberger, M. S. Altering the pathway of immunoglobulin hypermutation by inhibiting uracil-DNA glycosylase. Nature 419, 43–48 (2002). Di Noia, J. M. & Neuberger, M. S. Immunoglobulin gene conversion in chicken DT40 cells largely proceeds through an abasic site intermediate generated by excision of the uracil produced by AID-mediated deoxycytidine deamination. Eur. J. Immunol. 34, 504–508 (2004).
NATURE REVIEWS | IMMUNOLOGY
90. Hochegger, H., Sonoda, E. & Takeda, S. Post-replication repair in DT40 cells: translesion polymerases versus recombinases. Bioessays 26, 151–158 (2004). 91. Bertocci, B. et al. DNA polymerases µ and λ are dispensable for Ig gene hypermutation. J. Immunol. 168, 3702–3706 (2002). 92. Petersen, S. et al. AID is required to initiate Nbs1/γ-H2AX focus formation and mutations at sites of class switching. Nature 414, 660–665 (2001). 93. Lahdesmaki, A. et al. Delineation of the role of the Mre11 complex in class switch recombination. J. Biol. Chem. 279, 16479–16487 (2004). 94. Begum, N. A. et al. Uracil DNA glycosylase activity is dispensable for immunoglobulin class switch. Science 305, 1160–1163 (2004). 95. Galagan, J. E. & Selker, E. U. RIP: the evolutionary cost of genome defense. Trends Genet. 20, 417–423 (2004). 96. Chester, A. et al. RNA editing: cytidine to uridine conversion in apolipoprotein B mRNA. Biochim. Biophys. Acta 1494, 1–13 (2000). 97. Anant, S. & Davidson, N. O. Molecular mechanisms of apolipoprotein B mRNA editing. Curr. Opin. Lipidol. 12, 159–165 (2001). 98. Petersen-Mahrt, S. K. & Neuberger, M. S. In vitro deamination of cytosine to uracil in single-stranded DNA by apolipoprotein B editing complex catalytic subunit 1 (APOBEC1). J. Biol. Chem. 278, 19583–19586 (2003). 99. Yamanaka, S. et al. Apolipoprotein B mRNA-editing protein induces hepatocellular carcinoma and dysplasia in transgenic animals. Proc. Natl Acad. Sci. USA 92, 8483–8487 (1995). 100. Yamanaka, S. et al. Hyperediting of multiple cytidines of apolipoprotein B mRNA by APOBEC-1 requires auxiliary protein(s) but not a mooring sequence motif. J. Biol. Chem. 271, 11506–11510 (1996). 101. Kunkel, T. A. & Diaz, M. Enzymatic cytosine deamination: friend and foe. Mol. Cell 10, 962–963 (2002). 102. Okazaki, I. M. et al. Constitutive expression of AID leads to tumorigenesis. J. Exp. Med. 197, 1173–1181 (2003). 103. Loeb, L. A. et al. Lethal mutagenesis of HIV with mutagenic nucleoside analogs. Proc. Natl Acad. Sci. USA 96, 1492–1497 (1999). 104. Harris, R. S. et al. DNA deamination: not just a trigger for antibody diversification but also a mechanism for defense against retroviruses. Nature Immunol. 4, 641–643 (2003). 105. Nei, M. & Glazko, G. V. The Wilhelmine E. Key 2001 Invitational Lecture. Estimation of divergence times for a few mammalian and several primate species. J. Hered. 93, 157–164 (2002). 106. Madsen, P. et al. Psoriasis upregulated phorbolin-1 shares structural but not functional similarity to the mRNA-editing protein Apobec-1. J. Invest. Dermatol. 113, 162–169 (1999).
Acknowledgements We are grateful to our laboratory and neighbouring colleagues for helpful comments, especially E. Hendrickson, D. Livingston, P. Magee and L. Mansky. We also thank T. Floss and the reviewers for helpful comments. R.S.H. is supported by a Burroughs–Wellcome Fund Hitchings Elion Fellowship (United States), the Searle Scholars Program (United States) and a new laboratory start-up award from the University of Minnesota (United States).
Competing interests statement The authors declare no competing financial interests.
Online links DATABASES The following terms in this article are linked online to: Entrez: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi ACF | AID | APEX | APOB | APOBEC1 | APOBEC2 | APOBEC3F | APOBEC3G | CCR5 | CD4 | CUL5 | CXCR4 | elongin B | elongin C | Gag | RBX1 | SIV Vif | UNG2 | Vif FURTHER INFORMATION Reuben Harris’s homepage: http://biosci.cbs.umn.edu/BMBB/faculty/Harris.R.html Access to this interactive links box is free online.
VOLUME 4 | NOVEMBER 2004 | 8 7 7 ©2004 Nature Publishing Group
REVIEWS
CLINICAL STRATEGIES FOR EXPANSION OF HAEMATOPOIETIC STEM CELLS Brian P. Sorrentino Abstract | Haematopoietic stem cells (HSCs) give rise to all blood and immune cells and are used in clinical transplantation protocols to treat a wide variety of diseases. The ability to increase the number of HSCs either in vivo or in vitro would provide new treatment options, but the amplification of HSCs has been difficult to achieve. Recent insights into the mechanisms of HSC self-renewal now make the amplification of HSCs a plausible clinical goal. This article reviews the molecular mechanisms that control HSC numbers and discusses how these can be modulated to increase the number of HSCs. Clinical applications of HSC expansion are then discussed for their potential to address the current limitations of HSC transplantation. AUTOLOGOUS HSCs
A transplant with autologous haematopoietic stem cells (HSCs) is a treatment in which transplanted HSCs are obtained directly from the patient. This is typically used to support intensive treatment with cytotoxic drugs, but it can also be used for gene therapy of genetic disorders.
St. Jude Children’s Research Hospital, Department of Hematology/Oncology, Division of Experimental Hematology, 332 North Lauderdale, Memphis, Tennessee 38120, USA. e-mail:
[email protected] doi:10.1038/nri1487
878
Haematopoietic stem cells (HSCs) are of interest both in clinical medicine and in basic developmental biology. All mature cells in the blood and immune system are derived from HSCs (FIG. 1a), and all of these cell types can be generated from a single HSC1. This tremendous potential for reconstituting the haematopoietic system has allowed the development of transplantation of HSCs as a clinical strategy. There are several different sources for obtaining HSCs for transplantation. The first transplants were carried out using bone-marrow cells2, but cells from the peripheral blood or umbilical cord blood can also be used. HSCs obtained directly from the patient (AUTOLOGOUS HSCs) are used for rescuing the patient from the effects of high doses of chemotherapy or used as a target for GENE-THERAPY vectors. HSCs obtained from another person (ALLOGENEIC HSCs) are used to treat haematological malignancies by replacing the malignant haematopoietic system with normal cells. Allogeneic HSCs can be obtained from siblings (MATCHED SIBLING TRANSPLANTS), parents or unrelated donors (MISMATCHED UNRELATEDDONOR TRANSPLANTS). Currently, ∼45,000 patients each year are treated by HSC transplantation, a number that has been increasing during the past decade. Although most of these cases have involved patients with haematological malignancies — such as lymphoma,
| NOVEMBER 2004 | VOLUME 4
myeloma and leukaemia — there is growing interest in using HSC transplantation to treat solid tumours and non-malignant diseases. For example, erythrocyte disorders such as β-THALASSAEMIA and SICKLE-CELL ANAEMIA have been successfully treated by transplantation of allogeneic HSCs3. In parallel with these clinical advances, there has been much interest in elucidating the molecular mechanisms by which HSC replication and differentiation are controlled (FIG. 1b). A better understanding of these mechanisms is of interest from the perspective of basic science and for developing new approaches for HSC transplantation. During the past decade, there have been considerable advances in our understanding of the molecular events that control HSC kinetics. When HSCs divide, one possible outcome is the generation of new HSCs — a process known as self-renewal. Alternatively, HSCs can differentiate into more-mature blood cells or can be lost by undergoing apoptosis. Through the use of forward-genetic approaches, particularly loss-of-function or gain-of-function mouse models, some of the genes that are important in these developmental outcomes have been identified (FIG. 1b). This review focuses on the molecular mechanisms that regulate HSC pool size and their potential relevance for clinical application. Although many genes can
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS
GENE THERAPY
a Haematopoietic stem cells (1 in 20,000)
In the context of haematopoietic disorders, this is a strategy for transducing autologous haematopoietic stem cells or lymphocyte progenitors with genetic vectors that express a therapeutic transgene. Genetically modified cells are then re-infused to reconstitute the haematopoietic system. The goal can be either to replace a defective gene, such as for treatment of sickle-cell anaemia, or to confer a new property to blood cells, such as resistance to cytotoxic drugs. ALLOGENEIC HSCs
A transplant with allogeneic haematopoietic stem cells (HSCs) is a treatment in which transplanted HSCs are obtained from a normal donor. This approach can be used to treat either malignant or non-malignant disorders. Mismatches between the histocompatibility antigens of the donor and patient can lead to adverse events, such as rejection of the transplanted graft or pathological immune responses to normal tissues in the patient. MATCHED SIBLING TRANSPLANTS
Transplants for which the donor is a sibling who has all of the same paternal and maternal major histocompatibility alleles as the patient. These cases are usually associated with the least transplant-related complications. MISMATCHED UNRELATEDDONOR TRANSPLANTS
Allogeneic transplants for which the donor is an unrelated individual. The donors are screened for phenotypic similarities in histocompatibility antigens; however, adverse immunological reactions are more problematic than for matched sibling transplants. β-THALASSAEMIA
An inherited disorder of erythrocytes caused by decreased or absent expression of β-globin and resulting in chronic anaemia. The most severe form of β-thalassaemia, sometimes called Cooley’s anaemia, is characterized by the requirement for regular blood transfusions to sustain life. β-thalassaemia is a relatively common cause of anaemia in Africa, Europe and Asia.
T cells Common lymphoid progenitor
NK cells Megakaryocytes, platelets, erythrocytes, dendritic cells, eosinophils, neutrophils and monocytes B cells
Common myeloid progenitor
CFU-S (1 in 5,000)
Clonogenic CFU-C (1 in 200)
b Differentiation • GATA1 • C/EBP-α • PAX5 • PU.1 • GFI1 • NFE2 Haematopoietic stem cell Self-renewal • HOXB4 • Notch-1 • GATA2 • MPL • AML1 • STAT5 • Sonic • WNT • p21 proteins hedgehog • p18
Apoptosis • BCL-2 • p53
Figure 1 | Basics of haematopoiesis. a | The haematopoietic cascade. The haematopoietic hierarchy is shown, with increasing maturity of cells from left to right. For several cell types, the approximate frequency in the total mouse bone-marrow cell population is shown; however, it is important to note that these frequencies are strain dependent. Primitive haematopoietic stem cells (HSCs) give rise to all formed blood elements, including both myeloid and lymphoid cells, and are defined by the ability to repopulate lethally irradiated recipients. A slightly more mature cell, CFU-S (colony-forming unit spleen), is defined by the ability to form myeloid colonies in the spleens of irradiated mice. There is functional heterogeneity within these primitive compartments. For example, some cells can repopulate irradiated animals but only for a relatively short time (short-term repopulating cells), whereas others lead to stable repopulation for the life of the animal. HSCs give rise to common lymphoid progenitors95, which in turn give rise to T cells, B cells and natural killer (NK) cells. HSCs also generate common myeloid progenitors96, which in turn generate all mature myeloid cells, such as granulocytes, monocytes, megakaryocytes (from which platelets are derived) and erythrocytes. Myeloid CFU-Cs (colony-formingunit cells) are a more mature intermediate in the myeloid differentiation pathway and are defined by the ability to form clonally derived colonies after culture in semi-solid media. b | HSC fate outcomes. There are three possible fate outcomes for HSCs: self-renewal, differentiation and apoptosis. Some of the genes involved in each of these pathways are shown. First, to maintain a pool of HSCs with time, HSCs undergo self-renewal divisions in which one or more of the daughter cells remain pluripotent HSCs, which themselves can undergo self-renewal. Second, daughter cells can differentiate or undergo a commitment to differentiate. When this occurs, daughter cells acquire a limited ability to self-renew or a restriction in developmental potential or both. Third, HSCs can undergo apoptosis and be lost. The relative rates at which HSCs undergo self-renewal, differentiation and apoptosis determine the number of HSCs that are present in an animal at any given time. AML1, acute myeloid leukaemia 1; BCL-2, B-cell lymphoma 2; C/EBP-α, CCAAT/enhancer binding protein-α; GATA, GATA-binding protein; GFI1, growth-factor independent 1; HOXB4, homeobox B4; MPL, product of the myeloproliferative leukaemia-virus oncogene; NFE2, nuclear factor (erythroid-derived 2); p18, a cell-cycle regulator, also known as INK4C; p21, a cell-cycle regulator, also known as CIP1; p53, p53 tumour-supressor protein; PAX5, paired box protein 5; PU.1, ETS transcription factor, also known as SPI1; STAT5, signal transducer and activator of transcription 5.
influence HSC kinetics, here, I focus on those factors that show the greatest promise for controlling the amplification of HSCs in a clinical setting. These genes can be divided into three main categories: genes that encode transcription factors, such as homeobox B4 (HOXB4); genes that encode signalling molecules, such as haematopoietic cytokines, the WNT-protein family and the Notch family of receptors; and genes that regulate the cell cycle, such as p18 and p21. There is evidence that modulating the expression of genes in each of these categories can be used to achieve expansion of mouse HSCs. There is less information regarding the expansion of human HSCs, in part because HSC assay systems are not
NATURE REVIEWS | IMMUNOLOGY
as well defined for human HSCs as for mouse HSCs. Investigators have used immunodeficient NOD–SCID mice (that is, mice on a non-obese diabetic and severe combined immunodeficient background) to study engraftment with human haematopoietic cells; however, haematopoietic progenitors, rather than HSCs, dominate the early phase of engraftment4. Another problem with the NOD–SCID model is the absence of human T-cell development; however, this limitation can be overcome using NOD–SCID mice that also lack the common cytokine-receptor γ-chain5. Eventually, it will be important to validate approaches for HSC expansion in non-human primate systems and, ultimately, in humans.
VOLUME 4 | NOVEMBER 2004 | 8 7 9 ©2004 Nature Publishing Group
REVIEWS
SICKLE-CELL ANAEMIA
An inherited disorder of erythrocytes, with a high prevalence in African and African American populations, that is caused by a mutation in the β-globin gene. A single nucleotide substitution (and the resultant amino-acid substitution) leads to the polymerization of haemoglobin when it is deoxygenated, ultimately resulting in the occlusion of small blood vessels. Disease manifestations include chronic anaemia, multiple painful crises, organ damage and increased susceptibility to bacterial infections. ACUTE MYELOID LEUKAEMIA
A clonal, malignant disease of blood cells that is characterized by proliferation of abnormal leukemic blast cells in the bone marrow. These blast cells express myeloid cell-surface markers and are usually present in high concentrations in the blood and bone marrow. Non-random cytogenetic abnormalities are usually present and define the causal genetic lesion underlying the malignancy.
Haematopoietic cytokines
The first attempts to amplify HSCs focused on the use of haematopoietic cytokines. These are signalling molecules that induce proliferation at various stages of haematopoietic development. Many of these cytokines are locally produced in the bone-marrow microenvironment by stromal cells, indicating that they might be useful for promoting HSC amplification in vitro. From the mid-1980s to the early 1990s, the rapid discovery of new haematopoietic growth factors and their receptors enabled researchers to test their ability to amplify HSCs in culture. Much of this work took place in the context of gene-therapy strategies, because to achieve efficient integration of oncoretroviral vectors requires not only the preservation of HSCs in culture but also the induction of self-renewal divisions. For example, mouse HSCs can be stimulated to undergo selfrenewal divisions by exposure to interleukin-3 (IL-3), IL-6 and stem-cell factor6,7, and human HSCs are induced to proliferate by exposure to FLT3 (fms-related tyrosine kinase 3) and thrombopoietin8. However, exhaustive studies using combinations of multiple cytokines have not yet identified conditions that result in more than a fourfold increase in HSC numbers9,10. It is possible that new factors will be identified that can achieve this goal. Vascular endothelial growth factor (VEGF) is required for HSC function in mice; however, it could be difficult to modulate this pathway because the effects of VEGF on HSC survival depend on intracellular receptor–ligand interactions11 that might not be possible to reproduce by simply adding VEGF to the culture medium. Fibroblast growth factor 1 (FGF1) also promotes long-term growth of HSCs in culture12. Mouse bone-marrow cells cultured for 4 weeks in the presence of FGF1, but no other exogenously added cytokines, accumulate HSCs, as defined phenotypically and by transplantation assays12. However, it is not certain to what extent HSCs were expanded compared with the input level or whether FGF1 functions mainly as a survival factor for HSCs, rather than by inducing proliferation. Bone-marrow microenvironmental factors
In the bone marrow, adult HSCs are supported by cell–cell interactions with non-haematopoietic cells, including endothelial cells, fibroblastic bone-marrow stromal cells and osteoblasts. This bone-marrow microenvironment (or HSC ‘niche’) is responsible for the localization of HSCs to specific anatomical regions in the bone marrow. The effects of these supporting cells are mediated through specific molecular interactions, some of which could be used to promote HSC expansion in vitro. Modulation of the Notch-signalling pathway provides one such approach. HSCs express receptors of the Notch family, which are transmembrane signalling molecules known to inhibit differentiation in other systems. The Notch-1 ligand Jagged-1 is expressed by bone-marrow stromal cells13, endothelial cells13 and osteoblasts14, indicating that cell-contact-dependent signalling through Notch might have a role in the regulation of the HSC population. Genetically engineered
880
| NOVEMBER 2004 | VOLUME 4
mice with increased numbers of osteoblasts in trabecular bone show a small increase in HSC numbers in the bone marrow15. Chemical inhibition of signalling through Notch-1 abolished this increase in HSCs, showing that this signalling pathway is required for osteoblast-mediated support14. It has recently been shown that angiopoietin-1-mediated signalling (through its receptor TIE2) also has a role in interactions between osteoblasts and HSCs16. TIE2 is expressed by HSCs, and osteoblasts express angiopoietin-1, which binds to TIE2 and results in the quiescence of HSCs and their protection against apoptosis. The expansion of HSCs has been achieved by experimental modulation of the Notch-1-signalling pathway. Constitutive expression of activated Notch-1 by mouse HSCs (using a retroviral vector) resulted in immortalized, cytokine-dependent cell lines that could reconstitute irradiated mice17. When fresh bone-marrow cells were transduced with activated Notch-1 constructs and transplanted directly into irradiated recipients, marked expansion of HSCs occurred in the bone marrow, which was mediated by a relative block in HSC differentiation18. Signalling through Notch-1 might inhibit the differentiation of HSCs by sustaining expression of the transcription factor GATA2 (GATA-binding protein 2)19. A soluble form of Jagged-1 can mediate the expansion of human HSCs when added to liquid cultures13, indicating the potential of soluble Jagged-1 for promoting the expansion of HSCs in vitro. Further studies are required to establish the utility of this approach and to determine whether the propensity of Notch-1 to favour lymphoid differentiation compared with myeloid differentiation18 will limit its clinical applications. Homeobox genes and HOXB4
HOX genes encode a large family of transcription factors that have a highly conserved DNA-binding motif known as the homeodomain. In mammals, there are four main families of HOX factors that are arranged in a co-linear manner at different loci (groups A, B, C and D). By convention, the genes in each group are numbered based on their chromosomal order. HOX transcription factors regulate early developmental processes, such as bodypart patterning along the spinal axis. Because certain HOX-family members are expressed by HSCs or dysregulated in patients with ACUTE MYELOID LEUKAEMIA20, HOX genes have been examined to determine whether they have a role in haematopoietic development. One particular factor, HOXB4, has been extensively studied as a way of increasing the self-renewal of HSCs. The effects of HOXB4 were first discovered using a retroviral-expression vector in a mouse HSCtransplantation assay21. When lethally irradiated mice were reconstituted with bone-marrow cells that were transduced with a control vector, HSC numbers in the bone marrow regenerated to only 5–10% of normal levels. By contrast, when bone-marrow cells were transduced with a retroviral vector expressing HOXB4 and then transplanted, normal numbers of HSCs were regenerated21 (FIG. 2). Expansion of HSCs did not continue after normal numbers of HSCs had been
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS
100
% Normal HSC number
Hoxb4 neor
0 Bonemarrow cell
Untransplanted donor
In vitro culture with a vector expressing Hoxb4 or neor
Lethal irradiation
Primary transplant recipient
Secondary transplant recipient
Figure 2 | Functional effects of enforced HOXB4 expression in transplanted mouse HSCs. Shown here are the effects of expression of a homeobox B4 (Hoxb4)-containing vector on haematopoietic stem-cell (HSC) numbers in a mouse assay of serial transplantation. Bonemarrow cells are removed from the hind limbs of a donor animal and cultured for several days in the presence of a retroviral vector expressing either Hoxb4 (yellow) or, as a control, the neomycinresistance gene neo r (green). When these cells are transplanted into a lethally irradiated primary recipient, full reconstitution of mature blood cells is obtained with cells cultured in the presence of either of the two vectors. However, the overall number of HSCs declines to 5–10% of normal levels after transplantation of cells transduced with the neo r-expressing vector (or any other control vector). By contrast, HSC numbers are restored to a normal level in mice transplanted with HSCs that were transduced with the Hoxb4-expressing vector. Importantly, HSC numbers do not expand beyond the normal range when they are transduced with HoxB4, indicating that normal physiological control mechanisms continue to operate. If the primary recipients are then used as donors for the transplantation of lethally irradiated secondary recipients, Hoxb4transduced HSCs again lead to considerably higher numbers of HSCs, although HSCs do not fully regenerate to normal levels. These findings21 led to the idea that HOXB4 expression could be used for the in vitro expansion of HSCs.
CHRONIC GRANULOMATOUS DISEASE
An inherited disorder caused by defective oxidase activity in the respiratory burst of phagocytes. It results from mutations in any of the four genes that are necessary to generate the superoxide radicals required for normal neutrophil function. Affected patients suffer from increased susceptibility to recurrent infections.
regenerated, showing that the effects of HOXB4 were still subject to normal homeostatic controls. The mechanism of HOXB4-mediated expansion of HSCs is not well understood. It is known that HOXB4 can cooperatively dimerize with PBX1 (pre-B-cell leukaemia transcription factor 1)22, which is encoded by a protooncogene that is required for maintenance of definitive (adult) haematopoiesis23. Downregulation of PBX1 expression using a vector that co-expressed both a PBX1 antisense sequence and HOXB4 resulted in a further increase in HSC expansion relative to that observed using HOXB4 alone24. However, a mutation in the HOXB4 gene that disrupts interaction with PBX1 did not lead to increased expansion of HSCs, indicating that the HOXB4 and PBX1 genes act on distinct pathways25. The fact that Hoxb4-knockout mice have only mild defects in haematopoiesis26 indicates that other HOX factors could have similar functions (that are usually redundant) or that the induction of self-renewal depends on enforced overexpression.
NATURE REVIEWS | IMMUNOLOGY
Not only can HOXB4 induce the expansion of mouse HSCs in vivo, but transduction of these cells with a Hoxb4-containing retroviral vector results in marked expansion during extended in vitro culture27. Usually, culturing bone-marrow cells for 14 days in the presence of serum and myeloid cytokines results in a progressive depletion of HSCs. However, when mouse HSCs were transduced with Hoxb4 and then cultured for 14 days, the number of HSCs increased 40-fold compared with initial numbers and ∼2,000fold relative to control cultures in which HSCs were lost27. These results indicate the possibility of using HOXB4 to expand HSCs in culture before transplantation. A practical approach for clinical application that avoids the use of retroviral vectors has been developed using a HOXB4 fusion protein that contains a plasma-membrane permeabilization sequence to allow entry of HOXB4 to the cell cytoplasm 28. When mouse bone-marrow cells were cultured in the presence of this protein, HSCs were expanded — albeit to a lesser extent than achieved by retroviral transduction of Hoxb4 (REF. 28). This strategy has also been used to expand human haematopoietic cells, which can then repopulate immunodeficient mice29. Another strategy that might be feasible in a clinical setting uses constructs that allow HOXB4 function to be controlled. A HOXB4 fusion protein containing a mutant hormone-responsive element derived from the oestrogen receptor has been used to achieve HSC expansion that can be induced with tamoxifen, which binds the variant oestrogen receptor and allows nuclear localization of the fusion protein (Y. Shou, L. Stepano, D. K. Srivastava and B.P.S., unpublished observations). This system might allow the expansion of HSCs with the subsequent elimination of HOXB4 activity after sufficient expansion has been achieved. The HOXB4 gene might also be useful for selection of genetically modified HSCs for gene-therapy strategies owing to the strong competitive-repopulation advantage of transduced HSCs30. One important limitation of current gene-therapy protocols is the relatively inefficient transduction of human HSCs, which results in a chimeric state in which many of the transplanted HSCs remain unmodified. Consequently, there is an inadequate proportion of transduced blood cells to achieve a therapeutic effect. For example, successful treatment of sickle-cell anaemia would require that ∼30–40% of circulating erythrocytes express a vector-encoded globin gene. This level has not been achievable in primate model systems. If the vector expressed both a globin protein and a selectable marker (such as HOXB4), then preferential expansion of the subpopulation of genetically modified HSCs would occur, leading to therapeutic levels of ‘corrected’ erythrocytes. This strategy could be applied to many other diseases in which corrected cells do not themselves have a survival advantage, such as CHRONIC GRANULOMATOUS DISEASE31 and LYSOSOMAL-STORAGE 32 DISORDERS . For example, mouse models of chronic granulomatous disease can be corrected by transferring genes that encode the superoxide-generating phagocyte NADPH oxidase, such as p47 Phox (also known as Ncf1)33
VOLUME 4 | NOVEMBER 2004 | 8 8 1 ©2004 Nature Publishing Group
REVIEWS
a Selection with drug-resistance gene Drug-resistance gene
Cytotoxic-drug treatment
• 50% of HSCs transduced • Only 50% of HSCs remaining
• 25% of HSCs transduced
b Selection with drug-resistance gene and Hoxb4 Self-renewal Drug-resistance gene HOXB4-mediated regeneration
Cytotoxic-drug treatment Hoxb4
• 25% of HSCs transduced
• 50% of HSCs transduced • Only 50% of HSCs remaining
• 75% of HSCs transduced • 100% of HSCs remaining
Figure 3 | Synergistic effects of Hoxb4 and a drug-resistance gene for in vivo selection of transduced HSCs. a | The effects of drug selection for haematopoietic stem cells (HSCs) transduced with a vector expressing a drug-resistance gene are shown. After transplantation, only 25% of HSCs are transduced with the drug-resistance vector. The animal is then treated with the appropriate cytotoxic drug, which eliminates most of the untransduced HSCs but not the transduced HSCs as these are resistant and survive. This selection process increases the proportion of modified HSCs to 50%; however, there is little to no regeneration of HSCs after drug selection, so the overall number of HSCs is reduced to 50% of normal. b | The effects of drug selection when HSCs are transduced with a vector expressing both a drug-resistance gene and the homeobox B4 (Hoxb4) gene are shown. Initially, 25% of HSCs are transduced, similar to the results of transduction with a drug-resistance gene alone. After drug treatment, enrichment for transduced HSCs is achieved by simple elimination of untransduced HSCs. The transduced HSCs then have a regenerative advantage conferred by expression of the Hoxb4 gene. As a result of enforced self-renewal, the transduced HSCs are amplified until HSC numbers are restored to normal. This results not only in a lack of HSC depletion but also in an improvement in the selection efficiency so that now 75% of HSCs are transduced following recovery from drug treatment40. LYSOSOMAL-STORAGE DISORDERS
A group of inherited disorders in which one or more tissues become progressively engorged with lipid. Mutations in lysosomal enzymes result in an accumulation of lipiddegradation products, which typically occurs in monocytes and macrophages derived from the bone marrow. Many of these disorders result in damage to the spleen, liver, brain and bone marrow. VIRAL 2A SEQUENCES
Peptide sequences found in picornaviruses and other viruses that mediate protein cleavage through a ribosomal skip mechanism. These relatively small sequences have been used in gene-therapy vectors to express multiple proteins from a single mRNA transcript.
882
or gp91Phox (also known as Cybb)34; however, clinical application has been limited by inadequate numbers of transduced cells31. Amplification of HSCs using HOXB4 can be used to augment selection systems based on genes that provide drug resistance. Vectors that express drug-resistance genes, such as genes that encode dihydrofolate reductase35 or methylguanine methyltransferase36–38, can be used to select transduced mouse HSCs by administering the appropriate cytotoxic drug to the transplant recipient39. Although drug-resistant HSCs preferentially survive, this selection process is accompanied by a decrease in the overall number of HSCs. This limitation can be overcome by co-expressing Hoxb4 with the drug-resistance gene to allow selective regeneration of transduced cells following treatment with cytotoxic drugs40. The inclusion of Hoxb4 in the vector results in a reduction in the overall depletion of HSCs following drug treatment, and the efficiency of selection of transduced HSCs is increased (FIG. 3). A technical challenge to this approach is the requirement for vectors that
| NOVEMBER 2004 | VOLUME 4
express three coding regions (that is, the therapeutic gene, the drug-resistance gene and Hoxb4) — although new technologies such as the use of VIRAL 2A SEQUENCES in the vector now make this possible41. Whether it is safe to use HOXB4 expression for the expansion of HSCs is an open question. Vectors that express other HOX genes, such as Hoxb8 or Hoxa10, can cause leukaemia in mouse models42,43 by inducing conditional immortalization and by altering normal differentiation. Although there have been no cases of leukemic transformation when using HOXB4 in similar assays, it is possible that transformation could occur with a low incidence or a long latency period. Another concern is that high levels of HOXB4 expression, achieved by using adenoviral vectors or retroviral vectors with strong promoters, have been associated with abnormalities in myeloid differentiation 44,45. Therefore, the use of a HOXB4 gene that can be regulated, or the direct use of a recombinant protein, might be necessary to minimize the risk of inadvertent effects.
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS
WNT-signalling pathways
Another strategy for amplifying HSCs is based on the WNT-signalling pathway. WNT-mediated signalling controls numerous biological processes, including cellfate determination, cell migration and cell adhesion. Genetic lesions in this pathway are associated with transformation in various malignancies including leukaemia46,47. Nineteen WNT genes are present in the genome of mammalian cells, and these encode lipidmodified signalling proteins that function as ligands for the frizzled family of receptors. In the absence of WNTmediated signalling, β-catenin is ubiquitylated and degraded by the proteasome. WNT signalling through frizzled inhibits the degradation of β-catenin, which results in the accumulated β-catenin forming a complex with T-cell factor (TCF)/lymphoid-enhancerbinding factor (LEF)-family members to regulate the transcription of downstream genes (FIG. 4). Various WNT proteins (WNT5A and WNT10B) and frizzled proteins are expressed by mouse fetal liver HSCs, and the addition of WNT proteins to HSC cultures results in increased proliferation48. WNT5A, WNT10B and WNT2B are also expressed by human stromal cells in fetal bone49. Co-culture of human CD34+ progenitors with stromal cells expressing these WNT genes caused a marked expansion of myeloid progenitor cells49. Injection of WNT5A into immunodeficient mice repopulated with human haematopoietic cells resulted in increased reconstitution and more primitive haematopoietic cells in vivo 50. Gene-expression profiles of purified mouse HSCs showed expression of TCF/ LEF-family members51, indicating that WNT-mediated
Cell-cycle genes
WNT frizzled
Cytoplasm AXIN APC GSK3β
signalling is ongoing in unmodified HSCs. A role for WNT proteins in controlling the expansion of HSCs is further indicated by the fact that WNT-mediated signalling induces the proliferation of primitive cells in the skin52 and gastrointestinal tract53. The use of induced WNT-mediated signalling for the expansion of HSCs was shown directly by retroviral transduction of HSCs with a vector expressing a constitutively active β-catenin gene51. These transduced HSCs expanded in myeloid cultures and were relatively refractory to differentiation, which predominated in control cultures. As few as 125 transduced HSCs were required to reconstitute irradiated mice, which involves a 5–50-fold expansion of HSCs in vivo. Interestingly, upregulation of mRNA encoding HOXB4 and Notch-1 was observed in HSCs that were transduced with the β-catenin-expressing vector, indicating that at least some of the effects of WNT might be mediated through these genes. The addition of WNT protein directly to cultures provides a practical approach to expand HSCs in vitro. Incubation of mouse bone-marrow cells with purified WNT3A resulted in a sixfold expansion of HSCs within 6 days (as defined by their phenotype)54. Transplantation of these expanded cells into irradiated mice resulted in a more than fivefold expansion. So, similar to HOXB4, soluble WNT protein could be used for the expansion of HSCs in liquid culture and, therefore, might provide a cell-processing approach to increase the numbers of HSCs for transplantation. Further research is required to determine which specific WNT molecules are best suited to inducing the expansion of HSCs.
β-catenin TCF/LEF
Destruction complex
Nucleus TCF/LEF β-catenin Target genes
Figure 4 | The WNT-signalling pathway. Binding of extracellular WNT proteins to the receptor frizzled prevents β-catenin from associating with a destruction complex, which consists of axis inhibitor (AXIN), adenomatosis polyposis coli (APC) and glycogen-synthase kinase 3β (GSK3β). Association with this destruction complex would otherwise lead to ubiquitylation of β-catenin and subsequent proteasomal degradation. When WNT-mediated signalling occurs, accumulated β-catenin associates with T-cell factor (TCF)/lymphoid-enhancer-binding factor (LEF)-family transcription factors, and after entering the nucleus, these proteins activate a series of downstream target genes leading to self-renewal of haematopoietic stem cells.
NATURE REVIEWS | IMMUNOLOGY
Because self-renewal of HSCs requires cell division, genes that control cell-cycle progression have been evaluated for their role in HSC kinetics. Advances in molecular oncology have identified many cell-cycle regulatory genes for which the gain of function or loss of function contributes to tumour formation. Although many of the relevant mutant mouse strains show mainly nonhaematopoietic abnormalities, transplantation of their bone-marrow cells into wild-type recipients has allowed the effects of some of these genes on haematopoiesis to be studied in otherwise normal hosts (FIG. 5). Cyclin-dependent kinase inhibitors (CKIs) function to inhibit cell division by antagonizing the activities of specific cyclin-dependent kinases (CDKs)55. One group of CKIs are the CIP (CDK interaction protein)-KIP (kinase-interacting protein) proteins, which include p21 (also known as CIP1 or WAF1) and p27 (also known as KIP1). These proteins are potent inhibitors of CDK2 (FIG. 5) that have been found to regulate haematopoietic proliferation. p21 has a dominant role in inhibiting the entry of HSCs into the cell cycle. Among bone-marrow cells from p21-deficient mice, there are an increased number of cycling HSCs, whereas committed myeloid progenitor cells are present in normal numbers and have normal to decreased rates of cycling56. So, p21 exerts relatively specific effects in the HSC compartment. Expansion of human HSCs in vitro has been
VOLUME 4 | NOVEMBER 2004 | 8 8 3 ©2004 Nature Publishing Group
REVIEWS
p16
Cyclin D–CDK4/CDK6
RB
HSC cycling
Cyclin E–CDK2 BMI1
ARF
p27
p21
MDM2
p53
HSC apoptosis
Figure 5 | Cell-cycle genes in the regulation of HSCs. BMI1 is a polycomb repressor that prevents the depletion of haematopoietic stem cells (HSCs) by inhibiting expression of both p16 and ARF (alternate reading frame) by HSCs. In the absence of BMI1, p16 is expressed and inhibits the action of the cyclin D–cyclin-dependent kinase 4 (CDK4)/CDK6 complex. When inactivated, this complex cannot inhibit the retinoblastoma protein (RB), which then directly inhibits cell-cycle progression. The cyclin D–CDK4/CDK6 complex is also inhibited by p18, which when absent, leads to increased HSC self-renewal. Loss of expression of BMI1 also leads to expression of ARF, which results in inhibition of MDM2 (mouse double minute 2 homologue) with subsequent activation of the p53 tumour-suppressor protein. The action of p53 can either directly lead to HSC apoptosis or can cause HSC senescence by activating p21, which in turn can activate RB through inhibition of the cyclin E–CDK2 complex. p27 can also inhibit the cyclin E–CDK2 complex. Decreased expression of p21 or p27 can result in increased selfrenewal of HSCs. On the basis of gene-knockout models in mice, it seems that mouse HSCs usually express BMI1, p18, p21 and p27, but not p16 or ARF.
COMPETITIVE-REPOPULATION ASSAYS
Functional assays for measuring the number of mouse haematopoietic stem cells (HSCs). The HSC content of an undefined population of cells is determined by mixing the cells with a defined number of fresh bone-marrow cells from another source. After transplantation into lethally irradiated mice, genetic markers are used to distinguish progeny from the two HSC sources that are present in the blood and haematopoietic organs. These measurements allow a quantitative assessment of HSC activity. SMALL INTERFERING RNA (siRNA) MOLECULES
Synthetic double-stranded RNA molecules of 19–23 nucleotides, which are used to ‘knockdown’ (silence the expression of) a specific gene. This is known as RNA interference (RNAi) and is mediated by the sequencespecific degradation of mRNA.
884
achieved by downregulating the endogenous expression of p21 using a lentiviral vector that expresses a p21 antisense construct57. Although loss of p21 expression results in an increase in HSC numbers in mice, serialtransplantation studies show an increased rate of HSC loss with time56, indicating that other mechanisms can counter the effects of p21 inactivation. Mice that are deficient in p27 have no alteration in the HSC compartment but do show increased cycling and expansion of myeloid progenitor cells in their bone marrow58. Despite the suggested progenitor-restricted effects of p27, culture of mouse HSCs in the presence of p27 antisense oligonucleotides and an antibody specific for transforming growth factor-β (TGF-β) resulted in increased transduction with a mouse oncoretroviral vector59, presumably reflecting increased self-renewal divisions of HSCs. The discrepancy between these results might be explained by hypomorphic expression of p27 in the antisense experiments or by the effects contributed by the inhibition of TGF-β60. Another family of CKIs includes the INK4 proteins — p16 (also known as INK4A), p15 (also known as INK4B), p18 (also known as INK4C) and ARF (alternate reading frame; p19ARF in mice and p14ARF in humans) — which function as specific inhibitors of the cyclin-D-dependent kinases CDK4 and CDK6 (FIG. 5) and, as such, can function as classical tumour suppressors. Several of these INK4 proteins have been implicated in regulating HSC number and self-renewal. The most clearly defined functional role is that of p18. Mice that are deficient in p18 have more HSCs in the bone marrow, as shown by phenotypic analysis and
| NOVEMBER 2004 | VOLUME 4
61
. In contrast to p21deficient HSCs, exhaustion of p18-deficient HSCs was not observed during serial bone-marrow transplants, indicating that p18 is required for the suppression of self-renewal divisions. On the basis of these results, downregulation of the expression of p18 and/or p21 by HSCs, using antisense oligonucleotides or SMALL INTERFERING RNA (siRNA) MOLECULES, could be useful for expanding HSCs in culture. A role for other INK4 proteins in the self-renewal of HSCs has been indicated by recent studies of mice deficient in the polycomb repressor BMI1(REF. 62). BMI1 is expressed by primitive bone-marrow cells and at relatively high levels by many myeloid leukaemias63. BMI1 functions by repressing downstream genes through chromatin-structure modulations, and the locus that includes the p16 and ARF genes is one such target64. This locus encodes two products: p16 a modulator of the retinoblastoma pathway; and ARF, a modulator of the p53 tumoursuppressor protein pathway 65. When BMI1 is absent, p16 expression is increased, and this prevents the binding of cyclin D to CDK4 and/or CDK6, which in turn leads to senescence and cell-cycle arrest (FIG. 5). Increased levels of ARF result in sequestration of the p53-binding protein MDM2 (mouse double minute 2 homologue), thereby inhibiting p53 degradation and causing cell-cycle arrest and apoptosis. BMI1 has an important role in regulating the p16/ ARF locus in HSCs. BMI1-deficient mice have overtly normal haematopoiesis at birth, but at ∼2 months of age, they develop a bone-marrow failure syndrome that is associated with the progressive loss of HSCs66. BMI1 is also required for the maintenance of neural stem cells67 and leukemic stem cells68, indicating that it has a broad role in stem-cell maintenance. Loss of HSCs in BMI1deficient mice was associated with increased expression of both p16 and ARF by haematopoietic cells66, indicating that, in the absence of BMI1, de-repression of these genes resulted in either senescence or death of HSCs. In turn, when HSCs were transduced with vectors expressing either ARF or p16, an increase in cell death or a decrease in proliferation was observed, respectively 66. The neural stem-cell failure observed in BMI1deficient mice was partially rescued when these animals were crossed onto a p16-deficient background67, indicating that pathological upregulation of p16 expression is partially responsible for the stem-cell phenotype but that additional pathways downstream of BMI1 (potentially including ARF) also have a role in the stem-cell loss observed after deletion of Bmi1. So, given these data, will enforced downregulation of the expression of p16, ARF or both allow HSC expansion in certain conditions? Assuming that a principal role of BMI1 is to suppress the expression of gene products of the p16/ARF locus by HSCs, then elimination of these genes when BMI1 function is preserved might have no effect on haematopoiesis. By contrast, if increased expression of the p16/ARF locus gene products is causally associated with HSC loss during times of proliferative stress — such as during serial COMPETITIVE-REPOPULATION ASSAYS
p18
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS
T-CELL-DEPLETED GRAFT
The removal of T cells from an allogeneic graft to prevent immunological complications, such as graft-versus-host disease. This is typically carried out for transplants from haploidentical and mismatched unrelated donors. GRAFT-VERSUS-HOST DISEASE
(GVHD). A potentially serious complication that arises when donor-derived T cells attack host tissues, typically resulting in hepatic, dermatological and gastrointestinal damage. Acute GVHD occurs within the first 100 days after transplantation, whereas chronic GVHD occurs later and has a different pathophysiology. HAPLOIDENTICAL TRANSPLANT
An allogeneic transplant in which the donor is matched for half of the major histocompatibility alleles of the recipient and is typically one of the parents of the patient. Because as many as half of the alleles are mismatched, specific treatment of the patient and processing of the haematopoietic stem-cell graft are required to avoid severe immunological consequences. CD34+ CELLS
Human haematopoietic cells that are immunopurified based on their expression of the CD34 antigen. These cells typically comprise 5% of the total bonemarrow cell population. Although this population is considerably enriched for haematopoietic stem cells (HSCs), most CD34+ cells are not HSCs. PERIPHERAL-BLOOD HSCs
Haematopoietic stem cells (HSCs) collected from the peripheral blood of the donor, usually after treatment with granulocyte colony-stimulating factor, which mobilizes HSCs to migrate from the bone marrow to the blood.
transplantation or prolonged culture, or in older animals — then loss of expression of p16, ARF or both could allow for HSC expansion. Initial experiments indicate that loss of expression of both p16 and ARF does not confer a survival advantage to HSCs in longterm culture (L. Stepanova and B.P.S., unpublished observations), indicating that enforced downregulation of expression of the p16/ARF locus gene products will not lead to considerable expansion of HSCs. Potential clinical applications
Clinical transplantation of HSCs has varying degrees of effectiveness in treating both malignant and nonmalignant disorders. Amplification of HSCs would be useful in several contexts, both to overcome existing limitations and to develop new transplantation approaches. In many transplantation settings, restoration of T-cell and B-cell numbers can take many months. During this lengthy time, patients are at risk of serious infectious complications. This problem is particularly evident when the donor graft is depleted of T cells (T-CELL-DEPLETED GRAFT) to reduce the incidence and severity of GRAFT-VERSUS-HOST DISEASE — a potentially serious complication that arises from mismatches in histocompatibility antigens between the donor and the host. When using HSC grafts that have not been depleted of T cells, the early stage of immune reconstitution arises from mature T cells that are present in the graft69. However, when T cells are depleted from donor HSC grafts, T-cell recovery is dependent on de novo production of T cells from the thymus, and considerably more time is required for immune reconstitution70–72. This delay is particularly evident when using a parental donor as an HSC source. The use of one of the parents of the patient as an HSC donor (HAPLOIDENTICAL TRANSPLANT) has the advantage of providing a donor for nearly all patients, but this approach is limited by a high degree of mismatch in histocompatibility antigens. For this reason, T-cell depletion of the graft is commonly used. To achieve this, one method involves a purification strategy based on immunoselection for expression of CD34 (REF. 73); this provides a relatively pure population of HSCs that has been sufficiently depleted of donor T cells. When such grafts were used in paediatric transplantation, the rate of T-cell recovery was dependent on the dose of CD34 CELLS that were administered. High doses of CD34+ HSCs (>20 × 106 cells per kg) led to a marked decrease in the time required for T-cell reconstitution compared with cases in which lower doses were administered74,75. However, in many cases, this high number of CD34+ HSCs could not be collected from the donor. Therefore, the ability to amplify donor HSCs in vitro could improve the rate of immune reconstitution and thereby broaden the use of parental donors in HSC transplantation. One important new source of HSCs is umbilicalcord blood that is collected during newborn deliveries. In addition to their widespread availability, these HSCs have several useful properties, including their decreased ability to induce immunological reactivity against the patient because of increased levels of immune tolerance +
NATURE REVIEWS | IMMUNOLOGY
in the fetus. Interest in this approach has increased since the first successful transplantation of cord-blood HSCs in 1988 (REF. 76), and there are now an estimated 70,000 units of cord blood that are stored and available for transplantation77. However, their use is limited by the number of HSCs that can be collected, and it is clear that engraftment is closely correlated with the number of cells that are infused78. Furthermore, cordblood transplantation is difficult to use for treating adult patients because of the limited number of cells that are available, so it has generally been limited to paediatric patients. To address these limitations, ex vivo expansion protocols have been developed that use cell culture in the presence of various haematopoietic cytokines before transplantation. In one study, a high rate of both acute and chronic graft-versus-host disease was noted79, and in another study, ex vivo culture of cord-blood cells was associated with a lack of engraftment80. Therefore, it is clear that more-effective strategies for the expansion of cord-blood HSCs are needed and would considerably improve the therapeutic potential of this approach. Even when using bonemarrow cells as a source of HSCs, donors must undergo several aspirations to collect several thousand millilitres of bone marrow, a procedure that is carried out in the operating room under general anaesthesia. An alternative source is HSCs from the peripheral blood (PERIPHERAL-BLOOD HSCs), collected after treating the donor with granulocyte colony-stimulating factor to increase the number of circulating HSCs. Both of these procedures entail some risk and are relatively costly. Successful HSC expansion could potentially allow the use of much smaller volumes of bone marrow, perhaps even collected with a single aspiration as an outpatient procedure, thereby reducing the cost and risk of HSC collection. Gene therapy involves collecting HSCs from the patient and genetically modifying them with a therapeutic transgene. This genetic modification is typically carried out using retroviral vectors derived from either the mouse leukaemia-virus family (oncoretroviruses) or HIV (a lentivirus). The use of oncoretroviral vectors is considerably limited by the requirement for active cell cycling to achieve integration of the vector into the genome of the host cell81. HSCs are mainly quiescent and the induction of cell cycling using cytokines often leads to their differentiation, explaining (at least in part) the low transduction efficiencies that have been noted in human clinical trials31,82. One advantage of lentiviral vectors is their ability to integrate into the genome of non-dividing cells83; however, studies of T cells indicate that this ability is limited and that cell cycling is required for transduction by lentiviral vectors in some cases84. Therefore, it will be interesting to test whether inclusion of self-renewal factors in the transduction medium improves the efficiency of HSC transduction using these vector systems. Gene therapy has been successfully used to treat patients with severe combined immunodeficiency (SCID)85,86. In SCID, there is a large selective growth advantage for transduced lymphocyte progenitors87
VOLUME 4 | NOVEMBER 2004 | 8 8 5 ©2004 Nature Publishing Group
REVIEWS
Box 1 | Safety issues regarding HSC expansion for clinical treatment As expansion techniques for human haematopoietic stem cells (HSCs) are developed, it will be crucial to establish the safety of these approaches. Uncontrolled self-renewal is a property of cancer cells, so it will be important that HSC expansion is self-limited and, ideally, pharmacologically controlled. The use of soluble factors in the culture medium during HSC expansion, as has been described for homeobox B4 (HOXB4) and WNT3A (see main text), is one way in which this can be achieved. When signalling pathways are involved, control can also be achieved using specific chemical dimerizers to activate the pathway91,92. The function of transcription factors can be controlled by fusion to hormone-binding receptors (such as the oestrogen receptor, which responds to tamoxifen). For gene-therapy strategies in which HSCs are stably transduced with a vector, the potential for insertional mutagenesis is a considerable risk. The degree to which vector-induced transformation occurs in gene therapy for conditions other than X-linked severe combined immunodeficiency (XSCID) remains to be elucidated, but it is a relevant issue93. The risk of inadvertent gene activation might be reduced by modifications of the retroviral vector, such as deletion of viral enhancer elements, introduction of strong polyadenylation sequences and incorporation of chromatin insulators that flank the transgene94.
that compensates for inefficient HSC transduction. However, in the case of X-linked SCID (XSCID), two patients have developed T-cell leukaemia as a result of insertional mutagenesis from the gene-therapy vector 88,89. Despite this, these occurrences might be specific to XSCID and therefore not directly relevant to other haematopoietic disorders, given that insertional mutagenesis has never been observed in other trials of human gene therapy or in studies of immunocompetent dogs and monkeys90. However, the ability to isolate and expand HSCs for targeting could reduce the risk of insertional mutagenesis in patients with SCID by eliminating transduced T-cell progenitors from the graft and thereby reducing the risk of T-cell transformation.
1.
2.
3.
4.
5.
6.
7.
8.
886
Osawa, M., Hanada, K., Hamada, H. & Nakauchi, H. Long-term lymphohematopoietic reconstitution by a single CD34-low/negative hematopoietic stem cell. Science 273, 242–245 (1996). Thomas, E. D., Lochte, H. L. Jr, Cannon, J. H., Sahler, O. D. & Ferrebee, J. W. Supralethal whole body irradiation and isologous marrow transplantation in man. J. Clin. Invest. 38, 1709–1716 (1959). Gaziev, J. & Lucarelli, G. Stem cell transplantation for hemoglobinopathies. Curr. Opin. Pediatr. 15, 24–31 (2003). Cashman, J. D. et al. Kinetic evidence of the regeneration of multilineage hematopoiesis from primitive cells in normal human bone marrow transplanted into immunodeficient mice. Blood 89, 4307–4316 (1997). Hiramatsu, H. et al. Complete reconstitution of human lymphocytes from cord blood CD34+ cells using the NOD/SCID/γcnull mice model. Blood 102, 873–880 (2003). Bodine, D. M., Karlsson, S. & Nienhuis, A. W. Combination of interleukins 3 and 6 preserves stem cell function in culture and enhances retrovirus-mediated gene transfer into hematopoietic stem cells. Proc. Natl Acad. Sci. USA 86, 8897–8901 (1989). Bodine, D. M., Orlic, D., Birkett, N. C., Seidel, N. E. & Zsebo, K. M. Stem cell factor increases colony-forming unit-spleen number in vitro in synergy with interleukin-6, and in vivo in Sl/Sld mice as a single factor. Blood 79, 913–919 (1992). Petzer, A. L., Zandstra, P. W., Piret, J. M. & Eaves, C. J. Differential cytokine effects on primitive (CD34+CD38–) human hematopoietic cells: novel responses to Flt3ligand and thrombopoietin. J. Exp. Med. 183, 2551–2558 (1996).
9.
10.
11.
12.
13.
14.
15.
16.
Concluding remarks
The ability to amplify HSCs remains one of the ‘holy grails’ of haematology and bone-marrow transplantation. This goal has remained elusive despite exhaustive attempts to define culture conditions that facilitate HSC expansion. However, recent discoveries regarding the molecular control of HSC self-renewal and cell division now place this goal within reach. Several molecules have been shown to regulate HSC pool size, as determined using mouse gene-knockout models and retroviralvector-mediated overexpression models. Gene-transfer strategies are being explored to achieve amplification of transduced HSCs for gene-therapy protocols. For other applications, gene transfer might not be a feasible approach, and other strategies for HSC amplification will be necessary. Possibilities include the direct addition of the WNT3A signalling protein to the culture medium, the use of a cell-membrane-permeable HOXB4 transcription factor, or the use of siRNA molecules or antisense oligonucleotides directed against cell-cycle inhibitors. The next important step will be to show the activity of these systems in large animal models — particularly in non-human primates — given the known differences in HSC cycling kinetics and genetransfer rates between mouse and primate models. The safety of this approach must also be established (BOX 1); however, current information indicates that normal physiological control can be maintained in existing models of induced HSC expansion. It seems reasonable to predict that these strategies will initially be tested in patients with an acceptable risk–benefit ratio, such as those who are undergoing transplantation for advanced malignancies. Should stem-cell expansion prove feasible and safe in this context, there is great potential for the treatment of a broad range of conditions, including less severe diseases and non-malignant genetic disorders.
Bhatia, M. et al. Quantitative analysis reveals expansion of human hematopoietic repopulating cells after short-term ex vivo culture. J. Exp. Med. 186, 619–624 (1997). This paper shows that human HSCs can be moderately expanded for short periods of time when cultured in an appropriate combination of cytokines, but longer periods of culture result in a progressive loss of repopulating HSCs, as defined using an immunodeficient mouse model of human HSC engraftment. Glimm, H. & Eaves, C. J. Direct evidence for multiple self-renewal divisions of human in vivo repopulating hematopoietic cells in short-term culture. Blood 94, 2161–2168 (1999). Gerber, H. P. et al. VEGF regulates haematopoietic stem cell survival by an internal autocrine loop mechanism. Nature 417, 954–958 (2002). de Haan, G. et al. In vitro generation of long-term repopulating hematopoietic stem cells by fibroblast growth factor-1. Dev. Cell 4, 241–251 (2003). Karanu, F. N. et al. The Notch ligand Jagged-1 represents a novel growth factor of human hematopoietic stem cells. J. Exp. Med. 192, 1365–1372 (2000). Calvi, L. M. et al. Osteoblastic cells regulate the haematopoietic stem cell niche. Nature 425, 841–846 (2003). Zhang, J. et al. Identification of the haematopoietic stem cell niche and control of the niche size. Nature 425, 836–841 (2003). Arai, F. et al. Tie2/angiopoietin-1 signaling regulates hematopoietic stem cell quiescence in the bone marrow niche. Cell 118, 149–161 (2004).
| NOVEMBER 2004 | VOLUME 4
17. Varnum-Finney, B. et al. Pluripotent, cytokine-dependent, hematopoietic stem cells are immortalized by constitutive Notch1 signaling. Nature Med. 6, 1278–1281 (2000). 18. Stier, S., Cheng, T., Dombkowski, D., Carlesso, N. & Scadden, D. T. Notch1 activation increases hematopoietic stem cell self-renewal in vivo and favors lymphoid over myeloid lineage outcome. Blood 99, 2369–2378 (2002). 19. Kumano, K. et al. Notch1 inhibits differentiation of hematopoietic cells by sustaining GATA-2 expression. Blood 98, 3283–3289 (2001). 20. Magli, M. C., Largman, C. & Lawrence, H. J. Effects of HOX homeobox genes in blood cell differentiation. J. Cell. Physiol. 173, 168–177 (1997). 21. Sauvageau, G. et al. Overexpression of HOXB4 in hematopoietic cells causes the selective expansion of more primitive populations in vitro and in vivo. Genes Dev. 9, 1753–1765 (1995). This is the first description of HSC expansion in transplanted mice using a Hoxb4-containing retroviral vector. HSC expansion is still responsive to normal control mechanisms and is limited after normal HSC numbers have been regenerated. 22. Krosl, J. et al. Cellular proliferation and transformation induced by HOXB4 and HOXB3 proteins involves cooperation with PBX1. Oncogene 16, 3403–3412 (1998). 23. DiMartino, J. F. et al. The Hox cofactor and proto-oncogene Pbx1 is required for maintenance of definitive hematopoiesis in the fetal liver. Blood 98, 618–626 (2001). 24. Krosl, J., Beslu, N., Mayotte, N., Humphries, R. K. & Sauvageau, G. The competitive nature of HOXB4transduced HSC is limited by PBX1: the generation of ultracompetitive stem cells retaining full differentiation potential. Immunity 18, 561–571 (2003).
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS 25. Beslu, N. et al. Molecular interactions involved in HOXB4induced activation of HSC self-renewal. Blood 29 June 2004 (doi:10. 1182/blood-2004-04-1653). 26. Brun, A. C. et al. Hoxb4-deficient mice undergo normal hematopoietic development but exhibit a mild proliferation defect in hematopoietic stem cells. Blood 103, 4126–4133 (2004). 27. Antonchuk, J., Sauvageau, G. & Humphries, R. K. HOXB4-induced expansion of adult hematopoietic stem cells ex vivo. Cell 109, 39–45 (2002). This paper describes a 2,000-fold expansion of mouse HSCs in culture after transduction with a Hoxb4-containing retroviral vector. This degree of HSC expansion far exceeds the levels that have been obtained using any other method. 28. Krosl, J. et al. In vitro expansion of hematopoietic stem cells by recombinant TAT–HOXB4 protein. Nature Med. 9, 1428–1432 (2003). 29. Amsellem, S. et al. Ex vivo expansion of human hematopoietic stem cells by direct delivery of the HOXB4 homeoprotein. Nature Med. 9, 1423–1427 (2003). This paper shows that a recombinant HOXB4 protein that can penetrate the cell membrane can be used to expand human HSCs in culture. This paper supports the rationale that transcription factors that increase self-renewal can be used as growth factors in culture media. 30. Antonchuk, J., Sauvageau, G. & Humphries, R. K. HOXB4 overexpression mediates very rapid stem cell regeneration and competitive hematopoietic repopulation. Exp. Hematol. 29, 1125–1134 (2001). 31. Malech, H. L. et al. Prolonged production of NADPH oxidase-corrected granulocytes after gene therapy of chronic granulomatous disease. Proc. Natl Acad. Sci. USA 94, 12133–12138 (1997). 32. Dunbar, C. E. et al. Retroviral transfer of the glucocerebrosidase gene into CD34+ cells from patients with Gaucher disease: in vivo detection of transduced cells without myeloablation. Hum. Gene Ther. 9, 2629–2640 (1998). 33. Mardiney, M. et al. Enhanced host defense after gene transfer in the murine p47phox- deficient model of chronic granulomatous disease. Blood 89, 2268–2275 (1997). 34. Bjorgvinsdottir, H. et al. Retroviral-mediated gene transfer of gp91phox into bone marrow cells rescues defect in host defense against Aspergillus fumigatus in murine X-linked chronic granulomatous disease. Blood 89, 41–48 (1997). 35. Allay, J. A. et al. In vivo selection of retrovirally transduced hematopoietic stem cells. Nature Med. 4, 1136–1143 (1998). This paper provides definitive proof, using a mouse model, that drug-resistance genes can be used to select HSCs in vivo. 36. Persons, D. A. et al. Successful treatment of murine β-thalassemia using in vivo selection of genetically modified, drug-resistant hematopoietic stem cells. Blood 102, 506–513 (2003). 37. Zielske, S. P. & Gerson, S. L. Lentiviral transduction of P140K MGMT into human CD34+ hematopoietic progenitors at low multiplicity of infection confers significant resistance to BG/BCNU and allows selection in vitro. Mol. Ther. 5, 381–387 (2002). 38. Davis, B. M., Koc, O. N. & Gerson, S. L. Limiting numbers of G156A O6-methylguanine-DNA methyltransferase-transduced marrow progenitors repopulate nonmyeloablated mice after drug selection. Blood 95, 3078–3084 (2000). 39. Sorrentino, B. P. Gene therapy to protect haematopoietic cells from cytotoxic cancer drugs. Nature Rev. Cancer 2, 431–441 (2002). 40. Sawai, N., Persons, D. A., Zhou, S., Lu, T. & Sorrentino, B. P. Reduction in hematopoietic stem cell numbers with in vivo drug selection can be partially abrogated by HOXB4 gene expression. Mol. Ther. 8, 376–384 (2003). This paper shows that in vivo selection of HSCs with cytotoxic drugs results in a decrease in the overall number of HSCs. This reduction can be overcome by co-expression of HOXB4, which also results in an increase in the efficiency of the selection process. 41. Szymczak, A. L. et al. Correction of multi-gene deficiency in vivo using a single ‘self-cleaving’ 2A peptide-based retroviral vector. Nature Biotechnol. 22, 589–594 (2004). 42. Perkins, A. C. & Cory, S. Conditional immortalization of mouse myelomonocytic, megakaryocytic and mast cell progenitors by the Hox-2.4 homeobox gene. EMBO J. 12, 3835–3846 (1993). 43. Thorsteinsdottir, U. et al. Overexpression of HOXA10 in murine hematopoietic cells perturbs both myeloid and lymphoid differentiation and leads to acute myeloid leukemia. Mol. Cell. Biol. 17, 495–505 (1997).
44. Brun, A. C., Fan, X., Bjornsson, J. M., Humphries, R. K. & Karlsson, S. Enforced adenoviral vector-mediated expression of HOXB4 in human umbilical cord blood CD34+ cells promotes myeloid differentiation but not proliferation. Mol. Ther. 8, 618–628 (2003). 45. Schiedlmeier, B. et al. High-level ectopic HOXB4 expression confers a profound in vivo competitive growth advantage on human cord blood CD34+ cells, but impairs lymphomyeloid differentiation. Blood 101, 1759–1768 (2003). 46. Giles, R. H., van Es, J. H. & Clevers, H. Caught up in a Wnt storm: Wnt signaling in cancer. Biochim. Biophys. Acta 1653, 1–24 (2003). 47. Muller-Tidow, C. et al. Translocation products in acute myeloid leukemia activate the Wnt signaling pathway in hematopoietic cells. Mol. Cell. Biol. 24, 2890–2904 (2004). 48. Austin, T. W., Solar, G. P., Ziegler, F. C., Liem, L. & Matthews, W. A role for the Wnt gene family in hematopoiesis: expansion of multilineage progenitor cells. Blood 89, 3624–3635 (1997). 49. van den Berg, D. J., Sharma, A. K., Bruno, E. & Hoffman, R. Role of members of the Wnt gene family in human hematopoiesis. Blood 92, 3189–3202 (1998). 50. Murdoch, B. et al. Wnt-5A augments repopulating capacity and primitive hematopoietic development of human blood stem cells in vivo. Proc. Natl Acad. Sci. USA 100, 3422–3427 (2003). 51. Reya, T. et al. A role for Wnt signalling in self-renewal of haematopoietic stem cells. Nature 423, 409–414 (2003). This study shows that mouse HSCs can be expanded in culture after transduction with an activated β-catenin gene. It also shows that the WNT-signalling pathway is used by HSCs. 52. Zhu, A. J. & Watt, F. M. β-catenin signalling modulates proliferative potential of human epidermal keratinocytes independently of intercellular adhesion. Development 126, 2285–2298 (1999). 53. Korinek, V. et al. Depletion of epithelial stem-cell compartments in the small intestine of mice lacking Tcf-4. Nature Genet. 19, 379–383 (1998). 54. Willert, K. et al. Wnt proteins are lipid-modified and can act as stem cell growth factors. Nature 423, 448–452 (2003). This study tested whether soluble WNT3A is an HSC growth factor in culture. By culturing single cells in the presence of stem-cell factor and WNT3A, considerable expansion of mouse HSCs was obtained. 55. Sherr, C. J. & Roberts, J. M. CDK inhibitors: positive and negative regulators of G1-phase progression. Genes Dev. 13, 1501–1512 (1999). 56. Cheng, T. et al. Hematopoietic stem cell quiescence maintained by p21cip1/waf1. Science 287, 1804–1808 (2000). This paper shows that the HSC pool size is increased in p21-deficient mice owing to an increase in HSC proliferation. These data provide clear evidence that cyclin-dependent kinase inhibitors regulate HSC kinetics. 57. Stier, S. et al. Ex vivo targeting of p21Cip1/Waf1 permits relative expansion of human hematopoietic stem cells. Blood 102, 1260–1266 (2003). 58. Cheng, T., Rodrigues, N., Dombkowski, D., Stier, S. & Scadden, D. T. Stem cell repopulation efficiency but not pool size is governed by p27kip1. Nature Med. 6, 1235–1240 (2000). 59. Dao, M. A., Taylor, N. & Nolta, J. A. Reduction in levels of the cyclin-dependent kinase inhibitor p27kip-1 coupled with transforming growth factor β neutralization induces cell-cycle entry and increases retroviral transduction of primitive human hematopoietic cells. Proc. Natl Acad. Sci. USA 95, 13006–13011 (1998). 60. Dao, M. A., Hwa, J. & Nolta, J. A. Molecular mechanism of transforming growth factor β-mediated cell-cycle modulation in primary human CD34+ progenitors. Blood 99, 499–506 (2002). 61. Yuan, Y., Shen, H., Franklin, D. S., Scadden, D. T. & Cheng,T. In vivo self-renewing divisions of haematopoietic stem cells are increased in the absence of the early G1phase inhibitor, p18INK4C. Nature Cell Biol. 6, 436–442 (2004). This paper shows that bone marrow from p18deficient mice contains more HSCs owing to increased self-renewal divisions. Unlike loss of p21 expression, loss of p18 expression does not lead to the exhaustion of HSCs with time. 62. Park, I. K., Morrison, S. J. & Clarke, M. F. Bmi1, stem cells, and senescence regulation. J. Clin. Invest. 113, 175–179 (2004). 63. Lessard, J. & Sauvageau, G. Polycomb group genes as epigenetic regulators of normal and leukemic hemopoiesis. Exp. Hematol. 31, 567–585 (2003).
NATURE REVIEWS | IMMUNOLOGY
64. Jacobs, J. J., Kieboom, K., Marino, S., DePinho, R. A. & van Lohuizen, M. The oncogene and Polycomb-group gene bmi-1 regulates cell proliferation and senescence through the ink4a locus. Nature 397, 164–168 (1999). 65. Lowe, S. W. & Sherr, C. J. Tumor suppression by Ink4a–Arf: progress and puzzles. Curr. Opin. Genet. Dev. 13, 77–83 (2003). 66. Park, I. K. et al. Bmi-1 is required for maintenance of adult self-renewing haematopoietic stem cells. Nature 423, 302–305 (2003). This paper shows that, in maturing adult mice, loss of the BMI1 polycomb repressor results in the progressive loss of HSCs with time. This phenotype occurs because of the activation of p16 and p19ARF expression by HSCs. 67. Molofsky, A. V. et al. Bmi-1 dependence distinguishes neural stem cell self-renewal from progenitor proliferation. Nature 425, 962–967 (2003). 68. Lessard, J. & Sauvageau, G. Bmi-1 determines the proliferative capacity of normal and leukaemic stem cells. Nature 423, 255–260 (2003). 69. Roux, E. et al. Analysis of T-cell repopulation after allogeneic bone marrow transplantation: significant differences between recipients of T-cell depleted and unmanipulated grafts. Blood 87, 3984–3992 (1996). 70. Dumont-Girard, F. et al. Reconstitution of the T-cell compartment after bone marrow transplantation: restoration of the repertoire by thymic emigrants. Blood 92, 4464–4471 (1998). 71. Douek, D. C. et al. Assessment of thymic output in adults after haematopoietic stem-cell transplantation and prediction of T-cell reconstitution. Lancet 355, 1875–1881 (2000). 72. Roux, E. et al. Recovery of immune reactivity after T-celldepleted bone marrow transplantation depends on thymic activity. Blood 96, 2299–2303 (2000). 73. Lang, P. et al. Clinical scale isolation of highly purified peripheral CD34+ progenitors for autologous and allogeneic transplantation in children. Bone Marrow Transplant. 24, 583–589 (1999). 74. Handgretinger, R. et al. Megadose transplantation of purified peripheral blood CD34+ progenitor cells from HLAmismatched parental donors in children. Bone Marrow Transplant. 27, 777–783 (2001). This paper describes the importance of using high doses of CD34+ stem cells when using parental donors for paediatric transplants. 75. Lang, P. et al. Transplantation of highly purified peripheralblood CD34+ progenitor cells from related and unrelated donors in children with nonmalignant diseases. Bone Marrow Transplant. 33, 25–32 (2004). 76. Gluckman, E. et al. Hematopoietic reconstitution in a patient with Fanconi’s anemia by means of umbilical-cord blood from an HLA-identical sibling. N. Engl. J. Med. 321, 1174–1178 (1989). 77. Gluckman, E. Hematopoietic stem-cell transplants using umbilical-cord blood. N. Engl. J. Med. 344, 1860–1861 (2001). 78. Wagner, J. E. et al. Transplantation of unrelated donor umbilical cord blood in 102 patients with malignant and nonmalignant diseases: influence of CD34 cell dose and HLA disparity on treatment-related mortality and survival. Blood 100, 1611–1618 (2002). 79. Shpall, E. J. et al. Transplantation of ex vivo expanded cord blood. Biol. Blood Marrow Transplant. 8, 368–376 (2002). 80. Fernandez, M. N. et al. Cord blood transplants: early recovery of neutrophils from co-transplanted sibling haploidentical progenitor cells and lack of engraftment of cultured cord blood cells, as ascertained by analysis of DNA polymorphisms. Bone Marrow Transplant. 28, 355–363 (2001). 81. Miller, D. G., Adam, M. A. & Miller, A. D. Gene transfer by retrovirus vectors occurs only in cells that are actively replicating at the time of infection. Mol. Cell. Biol. 10, 4239–4242 (1990). 82. Moscow, J. A. et al. Engraftment of MDR1 and NeoR gene-transduced hematopoietic cells after breast cancer chemotherapy. Blood 94, 52–61 (1999). 83. Naldini, L. et al. In vivo gene delivery and stable transduction of nondividing cells by a lentiviral vector. Science 272, 263–267 (1996). 84. Chinnasamy, D. et al. Lentiviral-mediated gene transfer into human lymphocytes: role of HIV-1 accessory proteins. Blood 96, 1309–1316 (2000). 85. Hacein-Bey-Abina, S. et al. Sustained correction of X-linked severe combined immunodeficiency by ex vivo gene therapy. N. Engl. J. Med. 346, 1185–1193 (2002). 86. Aiuti, A. et al. Correction of ADA-SCID by stem cell gene therapy combined with nonmyeloablative conditioning. Science 296, 2410–2413 (2002).
VOLUME 4 | NOVEMBER 2004 | 8 8 7 ©2004 Nature Publishing Group
REVIEWS 87. Bunting, K. D., Sangster, M. Y., Ihle, J. N. & Sorrentino, B. P. Restoration of lymphocyte function in Janus kinase 3deficient mice by retroviral-mediated gene transfer. Nature Med. 4, 58–64 (1998). 88. Hacein-Bey-Abina, S. et al. A serious adverse event after successful gene therapy for X-linked severe combined immunodeficiency. N. Engl. J. Med. 348, 255–256 (2003). 89. Hacein-Bey-Abina, S. et al. LMO2-associated clonal T cell proliferation in two patients after gene therapy for SCID-X1. Science 302, 415–419 (2003). This paper describes a molecular analysis of two cases of T-cell leukaemia that occurred in a human trial of gene therapy for XSCID. Both showed insertional activation of the LMO2 oncogene by the retroviral vector. 90. Kiem, H. P. et al. Long-term clinical and molecular follow-up of large animals receiving retrovirally transduced stem and progenitor cells: no progression to clonal hematopoiesis or leukemia. Mol. Ther. 9, 389–395 (2004).
888
91. Jin, L. et al. In vivo selection using a cell-growth switch. Nature Genet. 26, 64–66 (2000). 92. Jin, L. et al. Targeted expansion of genetically modified bone marrow cells. Proc. Natl Acad. Sci. USA 95, 8093–8097 (1998). 93. Kohn, D. B. et al. American Society of Gene Therapy (ASGT) ad hoc subcommittee on retroviral-mediated gene transfer to hematopoietic stem cells. Mol. Ther. 8, 180–187 (2003). 94. Baum, C. et al. Side effects of retroviral gene transfer into hematopoietic stem cells. Blood 101, 2099-2114 (2003). An excellent review of the safety issues associated with HSC-targeted gene therapy. There is a detailed consideration of how safety can be improved through modified gene-therapy vectors and protocols. 95. Kondo, M., Weissman, I. L. & Akashi, K. Identification of clonogenic common lymphoid progenitors in mouse bone marrow. Cell 91, 661–672 (1997). 96. Akashi, K., Traver, D., Miyamoto, T. & Weissman, I. L. A clonogenic common myeloid progenitor that gives rise to all myeloid lineages. Nature 404, 193–197 (2000).
| NOVEMBER 2004 | VOLUME 4
Acknowledgements I thank C. Sherr for his critical reading of this manuscript and for his many useful suggestions. This work was supported by grants from the National Institutes of Health (Bethesda, United States), a Program Project Grant (United States) and the American Lebanese Syrian Associated Charities (Memphis, United States).
Competing interests statement The author declares no competing financial interests.
Online links DATABASES The following terms in this article are linked online to: Entrez Gene: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=gene ARF | β-catenin | BMI1 | CD34 | GATA2 | HOXB4 | Notch-1 | p16 | p18 | p21 | p27 Access to this interactive links box is free online.
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS
FORKHEAD-BOX TRANSCRIPTION FACTORS AND THEIR ROLE IN THE IMMUNE SYSTEM Paul J. Coffer* and Boudewijn M. T. Burgering‡ Abstract | It is more than a decade since the discovery of the first forkhead-box (FOX) transcription factor in the fruit fly Drosophila melanogaster. In the intervening time, there has been an explosion in the identification and characterization of members of this family of proteins. Importantly, in the past few years, it has become clear that members of the FOX family have crucial roles in various aspects of immune regulation, from lymphocyte survival to thymic development. This review focuses on FOXP3, FOXN1, FOXJ1 and members of the FOXO subfamily and their function in the immune system.
*Department of Pulmonary Diseases and ‡Department of Physiological Chemistry, University Medical Centre, Heidelberglaan 100, 3584 CX Utrecht, The Netherlands. Correspondence to P.J.C. e-mail:
[email protected] doi:10.1038/nri1488
Forkhead proteins are a large family of functionally diverse transcription factors that have been implicated in a variety of cellular processes1. The name forkhead is dervived from the Drosophila melanogaster fork head (fkh) gene product, which is required for terminal pattern formation in the embryo2. Shortly after this discovery, a novel group of liver-specific transcription factors was identified (the hepatic nuclear factor 3 (HNF3) family), and these have DNA-binding domains with a high degree of similarity to that of the fkh gene product3. The discovery of this forkhead motif led to the definition of a new family of transcription factors that have now been identified in organisms ranging from yeast to humans. Forkhead transcription factors are commonly associated with the regulation of development, and it has been proposed that, with evolution, the increase in complexity of organisms has provided the driving force for the expansion of this protein family. In the past ten years, more than 100 members of the forkhead transcription-factor family have been identified. This led to a recent revision of the nomenclature used to identify these proteins4. FOX (forkhead box) is now used as the symbol for all chordate forkhead transcription factors. A phylogenetic analysis has resulted in the definition of 15 classes for all known FOX proteins, so these transcription factors are classified in terms of structure not function. In each subfamily (or class), a
NATURE REVIEWS | IMMUNOLOGY
gene is identified by a number; the genes that encode human transcription factors are denoted by uppercase letters (that is, FOX), whereas only the first letter is capitalized for the genes that encode mouse transcription factors (Fox). Both human and mouse proteins are denoted by uppercase letters (that is, FOX). The crystal structure of the forkhead domain bound to DNA has been solved and has been described as a ‘winged helix’ owing to its double-wing structure similar to the shape of a butterfly5 (FIG. 1a). It is a relatively invariant structure, with most of the amino acids being conserved between family members. This has made it difficult to understand the molecular mechanisms underlying the sequence specificity of the DNAbinding domains of different FOX proteins. Although all FOX proteins can bind DNA, the functional effect of this can be either the activation (transactivation) or the inhibition of gene transcription. In contrast to the DNA-binding domains, there is almost no sequence homology between the transactivation or repression domains of members of the FOX family (FIG. 1b), and little is known about their interactions with the transcriptional machinery. The growth of the FOX family during the past decade has been paralleled by an increased understanding of the importance of FOX proteins in the regulation of developmental processes. FOX gene mutations have VOLUME 4 | NOVEMBER 2004 | 8 8 9
©2004 Nature Publishing Group
REVIEWS
a
b
Amino terminus FOXJ1
W2
Carboxyl terminus
FKH 419
W1 FKH
FOXN1
648 H2
FOXO3A
FKH 673
H1 H3
FKH
FOXP3
431 100 amino acids
Figure 1 | The FOX protein winged-helix DNA-binding domain. a | Three-dimensional structure of the DNA-binding domain of forkhead box C2 (FOXC2), showing helical (H) sections and winged (W) sections. This image is redrawn with permission from REF. 79 (2002) Elsevier. b | A schematic representation of the location of the forkhead DNA-binding domain (FKH) of the various FOX factors described in this review. These proteins are classified in terms of structure, not function, and the location of the DNA-binding domain can vary. The FKH domain of FOXP3 is unique in that it is located very close to the carboxyl terminus of the protein. This has led to the proposal that FOXP3 might lack a bona fide transactivation domain and could function as a transcriptional inhibitor.
REGULATORY T (TReg) CELLS
A subset of CD4+ suppressor T cells that express high levels of CD25 (the interleukin-2 receptor α-chain), the role of which is to maintain selftolerance.
been found to be responsible for diverse phenotypes — from craniopharyngeal development (FOXE1), to speech and language development (FOXP2) and hearing (Foxi1)6. However, it has also become clear that these winged-helix proteins have crucial roles in various aspects of immune regulation. In the past few years, there have been many reports describing roles for distinct FOX transcription factors in the regulation of various aspects of the adaptive immune response. This article focuses on several members of the FOX family from different classes that have recently been shown to regulate diverse functions in the immune system: FOXP3, FOXN1, FOXJ1 and members of the FOXO subfamily. It reviews our current knowledge of the regulation and function of these important transcriptional regulators (TABLE 1). FOXP3 and regulatory T cells
Immune responses need to be constantly regulated — a balancing act that ensures appropriate reactivity to pathogens while preventing the development of an
unwanted autoimmune reaction. Autoreactive lymphocytes are normally deleted in the thymus, but selfreactive T cells are still found in the periphery in normal healthy individuals. However, these cells do not normally attack organs that express the corresponding autoantigen, and self-tolerance is generally maintained. Instead, T cells seem to be actively maintained in an unresponsive state. Recently, this has been shown to be mediated by REGULATORY T (T ) CELLS, which are characterized by expression of CD25 (the interleukin-2 (IL-2) receptor α-chain). These cells comprise ~5–10% of the total population of CD4+ T cells in mice and are thought to be crucial in the repression of autoimmune disorders, transplant rejection and inflammatory bowel disease7,8. Autoimmune diseases arise when self-reactive T cells somehow overcome the usual restraining mechanisms, leading to clonal expansion and activation. The activation of TReg cells is antigen specific; however, these lymphocytes can inhibit both CD4+ and CD8+ T cells in an antigen non-specific manner. This is thought to be through a combination of cell–cell contact and cytokine production, with a role for IL-10 and transforming growth factor-β (TGF-β). Reg
Scurfy mice and human IPEX. Scurfy (sf ) is an X-linked recessive mutation in mice that results in lethality of hemizygous males at 16–25 days after birth. There is an overproliferation of activated CD4+ T cells, resulting in multi-organ infiltration, and scurfy mice seem unable to properly regulate lymphocyte activity9,10. Adoptive transfer of CD4+CD8– T cells, but not CD4–CD8+ T cells, leads to disease, indicating that it is CD4+ T cells that are crucial for mediating the development of the scurfy phenotype in mice9. An intriguing early observation was that adoptive transfer of wild-type lymphocytes can control the T cells of scurfy mice, thereby preventing the development of disease9. So, a normal lymphoid compartment is somehow able to repress the scurfy phenotype. Recently, the gene that is defective in scurfy mice was identified and was found to encode FOXP3, a novel
Table 1 | Fox gene-knockout mice and their immune-related phenotypes Gene
Phenotype
Role in immune system
Disease
Foxp3 –/–
Scurfy-mouse phenotype; aggressive lymphoproliferative autoimmune syndrome; T cells have an activated phenotype; and lack of CD4+CD25+ TReg cells
Development and function of CD4+CD25+ TReg cells
IPEX; scurfy-mouse phenotype
References 11,13–17
Foxn1–/–
Congenital absence of hair and severe immunodeficiency; rudimentary thymus with lack of T-cell development; and defective thymic epithelial-cell differentiation and proliferation
Differentiation of thymic epithelial cells
Nude phenotype
30,33,37
Foxj1–/–
Embryonic lethal; fetal-liver chimerization results in systemic autoimmune inflammation with lymphocytic pathology, TH1-cell hyperactivation, NF-κB hyperactivation and cytokine overproduction
Suppression of T-cell activation and prevention of autoimmunity
Protein expression levels decreased in SLE-prone mouse strains
42–44
Foxo3a–/–
Spontaneous lymphoproliferation; mild, multi-system, non-lethal inflammation; TH1- and/or TH2-cell hyperactivation; NF-κB hyperactivation; and cytokine overproduction
Regulation of lymphocyte proliferation and apoptosis, suppression of T-cell activation and prevention of autoimmunity
Protein activity decreased in SLE-prone mouse strains
54,55
Fox, forkhead box; IPEX, immunodysregulation, polyendocrinopathy and enteropathy, X-linked syndrome; NF-κB, nuclear factor-κB; SLE, systemic lupus erythematosus; TH cell, T helper cell; TReg cell, regulatory T cell.
890
| NOVEMBER 2004 | VOLUME 4
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS member of the FOX transcription-factor family11. Analysis of Foxp3 expression revealed that it was most highly expressed in lymphoid organs, such as the thymus and spleen, and more specifically by CD4+ T cells. In scurfy mice, a frameshift mutation in the gene encoding FOXP3 results in the production of a functionally inactive, truncated protein product that lacks the DNAbinding domain. Interestingly, mice that lack a functional ∆K
a ∆E
Stop FOXP3
ZNF
A→T Stop
I→V
R→W F→C FKH
ZIP
C→Y
1
431
scurfy
ZNF
ZIP
Protein–protein interaction domain
b
DNA-binding domain
TGF-β TGF-β receptor
E2 Oestrogen
Cytoplasm E2
SMAD2/3
Termination of signalling through polyubiquitylation of receptor
ER
E2 ER
P SMAD2/3
E2 ER
TF
SMAD4 P SMAD2/3
FOXP3
TF
SMAD4
SMAD4 P SMAD2/3
SMAD7
ERE CD4+CD25+ T cell
CD4+CD25– T cell
Nucleus
Figure 2 | FOXP3 is a master regulator of regulatory T cells. a | Schematic representation of the forkhead box P3 (FOXP3) transcription factor, showing the location of various mutations (and the amino-acid changes) that have been identified in patients with IPEX (immunodysregulation, polyendocrinopathy and enteropathy, X-linked syndrome). Most mutations result in either a truncated protein or inhibition of DNA binding. Interestingly, two mutations (∆E and ∆K) occur within a putative protein–protein interaction domain, indicating that crucial regulation of FOXP3 is mediated by accessory proteins. The mouse scurfy protein is truncated as indicated, resulting in a lack of the DNA-binding domain. b | FOXP3 expression has been proposed to be regulated through the stimulation of CD4+ T cells. Oestrogen (E2) can induce FOXP3 expression, possibly through the direct interaction of the oestrogen receptor (ER) with the FOXP3 promoter. This could be particularly relevant during pregnancy when increased numbers of CD4+CD25+ regulatory T (TReg) cells are required. Transforming growth factor-β (TGF-β) can induce the expression of FOXP3, most probably through activation of SMAD (mothers against decapentaplegic homologue) transcription factors. Although TGF-β-mediated signalling is normally inhibited by an autoregulatory loop that involves SMAD7, this loop can be inhibited by FOXP3. Whether this results from direct effects on SMADs or an accessory transcription factor(s) (TF) is unknown. The result of induction of FOXP3 expression by TReg cells would be to prolong TGF-β-mediated signalling, perhaps allowing the stabilization or expansion of the TReg-cell pool. ERE, oestrogen-response element; FKH, forkhead DNA-binding domain; ZIP, leucine zipper; ZNF, zinc finger.
NATURE REVIEWS | IMMUNOLOGY
Foxp3 gene product have a much more severe phenotype than would be predicted from studies analysing the role of TReg cells. Scurfy mice die within a few weeks, whereas depletion of TReg-cell populations (by selecting for CD25+ cells) leads to autoimmunity but not rapid death. One possible explanation is that the expression of FOXP3 is essential in the thymic environment during development, allowing the differentiation of a crucial population of TReg cells. Loss of functional FOXP3 might result in the escape of self-reactive cells that would normally differentiate to become TReg cells, thereby greatly exacerbating the autoimmune phenotype12. Alternatively, FOXP3 might induce another, as-yet-undefined population of regulatory T cells, or perhaps depletion of CD25+ cell populations is insufficient to remove all Foxp3-expressing cells, so the resultant autoimmune disease is milder. Immunodysregulation, polyendocrinopathy and enteropathy, X-linked syndrome (IPEX; also known as X-linked autoimmunity and allergic dysregulation syndrome, XLAAD) is a fatal recessive disorder of humans that develops in early childhood. Symptoms manifest as protracted diarrhoea, dermatitis, insulin-dependent diabetes, thyroiditis and anaemia. Massive T-cell infiltration into the skin and gastrointestinal tract is also observed, as well as high serum levels of autoantibodies, which is indicative of autoimmune disease. Affected children also suffer from allergic manifestations including severe eczema, high IgE levels, eosinophilia and food allergies. Recent linkage studies have mapped the gene responsible for the development of IPEX to a conserved region on the X chromosome that contains the sf gene in mice13,14. Importantly, the scurfy syndrome in mice has several characteristics that are similar to IPEX, which leads to the possibility that FOXP3 mutations might be responsible for this fatal human disease (FIG. 2a). In various family studies, analysis of FOXP3 has revealed single nucleotide changes in the forkhead domain of FOXP3 in affected individuals, indicating disruption of DNA binding13,14. A further study identified a 3-base-pair deletion in the leucine-zipper dimerization domain, which might allow FOXP3 to homo- or heterodimerize; therefore, mutation could result in aberrant function15. Because the clinical phenotype of this patient is as severe as those with truncated FOXP3 proteins, this indicates that interaction with other proteins is crucial for transcription-factor function. Identification of such interacting proteins in these patients will certainly aid our understanding of the role of FOXP3 in autoimmune disease. Linking FOXP3 with TReg cells. The autoimmune pathologies observed in both humans and mice that lack functional FOXP3 indicate that this transcription factor has a crucial role in the regulation of T-cell function. Indeed, Fontenot et al. have generated Foxp3 –/– mice, which succumb to an aggressive lymphoproliferative autoimmune syndrome almost identical to that of scurfy mice16. It was found that Foxp3 –/– mice lack a discrete CD4+CD25+ T-cell population that is observed in wild-type mice, and Foxp3 –/– mice have many activated CD4+ T cells.
VOLUME 4 | NOVEMBER 2004 | 8 9 1 ©2004 Nature Publishing Group
REVIEWS
RECOMBINATION-ACTIVATINGGENE-KNOCKOUT MICE
Recombination-activating genes (Rag1 and Rag2) are expressed by developing lymphocytes. Mice that are deficient in either RAG protein fail to produce B and T cells owing to a developmental block in the gene rearrangement that is required for receptor expression.
892
Several recent reports have attempted to determine whether expression of FOXP3 is sufficient to confer TReg-cell activity on naive CD4+CD25– T cells16–18. Stimulation of Foxp3-transduced cells through the T-cell receptor (TCR) resulted in reduced proliferation relative to controls, as well as greatly reduced cytokine production. Furthermore, ectopic expression of FOXP3 resulted in increased expression of GITR (glucocorticoidinduced tumour-necrosis factor (TNF)-receptor-related protein) and CTLA4 (cytotoxic T-lymphocyte antigen 4), both of which are present on the cell surface of naturally occurring CD4+CD25+ T cells. These transduced cells could also suppress CD4+CD25– responder T cells when stimulated with CD3-specific monoclonal antibody. CD4+CD25– T cells that are retrovirally transduced with Foxp3 have also been injected intravenously into RECOMBINATION-ACTIVATING-GENE-KNOCKOUT MICE together with non-transduced CD4+CD25– T-cell populations16. Control animals that received CD4+CD25– T-cell populations transduced with an ‘empty’ virus developed severe wasting disease, whereas those that received Foxp3-transduced and non-transduced T cells were healthy. In contrast to control animals, mice that received Foxp3-transduced cells generated a population of CD4+CD25+ TReg cells, which were found to express increased levels of mRNA encoding IL-10. Because protection against inflammatory bowel disease can be mediated by wild-type CD4+CD25+ T cells and depends on the expression of IL-10 (REF. 19), IL-10 production might be an important mechanism of action of FOXP3. However, the relevance of IL-10 remains unclear because, in a separate study, the levels of IL-10 were found to be markedly decreased in Foxp3-transduced cells17. Although there have been numerous studies investigating the in vivo role of FOXP3, there are still few clues to the molecular mechanisms that underly its function as a transcription factor. Schubert et al. have found that FOXP3 binds DNA and localizes mainly in the nucleus and that this requires the presence of its forkhead domain20. The forkhead domain of FOXP3 is unique in that it is located in close proximity to the carboxyl terminus of the protein. This has led to the proposal that FOXP3 might lack a bona fide transactivation domain and could function as a transcriptional inhibitor. Indeed, experiments have shown that FOXP3 can, to some degree, reduce the transcriptional activity of a synthetic forkhead reporter construct20. Furthermore, the expression of FOXP3 by Jurkat cells resulted in a marked reduction in IL-2 production after the crosslinking of cell-surface CD3 (REF. 20). Normally, the expression of IL-2 depends on NFAT (nuclear factor of activated T cells), and a significant reduction in the level of transcription of an NFAT reporter construct was also observed in FOXP3-expressing cells. Importantly, the binding site of FOXP3 in the IL-2 promoter overlaps with that of NFAT, indicating a potential ‘competitive mechanism’ of inhibition. Potential FOXP3-binding sites have also been identified in the promoters of the genes encoding IL-4, TNF and GM-CSF (granulocyte/ macrophage colony-stimulating factor), all cytokines
| NOVEMBER 2004 | VOLUME 4
that are dysregulated in scurfy mice. However, the functional consequences of these observations have not yet been fully determined. Most of the studies described so far were carried out using mice and examined FOXP3 expression by mouse TReg cells. But is there evidence for a similar role for this transcription factor in humans? CD4+CD25+ TReg cells in humans comprise ~1–3% of the total number of CD4+ T cells. In humans, FOXP3 has been found to be expressed by both CD4+ and CD8+ T cells, although the level of expression by CD4+ T cells is much higher21. Similar to the observations made for mice, FOXP3 was found to be expressed only by CD4+CD25+ T cells and not by CD4+CD25– T cells, and the addition of CD4+CD25+ T cells to CD4+CD25– T-cell cultures suppressed proliferation triggered by TCR ligation. However, in humans, it is specifically a population of CD4+CD25hi T cells that has suppressor activity22. So, it seems that at least some of the lessons learned from the study of scurfy mice are applicable to the human immune system. Stimulation of human CD4+CD25– T cells through ligation of the TCR and co-stimulation of CD28 leads to proliferation, production of cytokines and expression of CD25. However, in contrast to mice, this can also lead to FOXP3 expression. Furthermore, CD4+CD25+ T cells generated from activated CD4+CD25– T cells could suppress the proliferation of freshly isolated CD4+CD25– T cells21. The observed differences between mice and humans could result from the tendency to use peripheralblood-derived T cells in studies of humans compared with T cells derived from the spleen or lymph nodes in studies of mice. The data indicate two possible pathways for the generation of TReg cells in humans: thymic selection or peripheral development through immune stimulation. Indeed, Khattri et al. showed that limiting Foxp3 expression to the thymus failed to rescue scurfy mice from disease18. This lends support to the idea that constant expression of Foxp3 is required for suppressive function. Regulation of FOXP3 expression. What is the precise signal that triggers T cells to express FOXP3? In the mouse thymus, CD4+CD25+ thymocytes preferentially transcribe Foxp3, and similarly in the periphery, CD4+, but not CD8+, T cells were found to express Foxp3 (REFS 16,17). Expression does not seem to be a consequence of T-cell activation in mice, because in vitro stimulation of CD4+CD25– T cells did not result in the expression of Foxp3 (in contrast to humans). Importantly, this separates FOXP3 from other TReg-cell markers (such as CD25, CTLA4 and GITR), the expression of which can also be upregulated by activated T cells. It should be noted that CD4+CD25– T-cell populations can also express low levels of FOXP3; however, these levels are about 100-fold lower than those of CD4+CD25+ T cells17. Interestingly, after TCR ligation, the level of FOXP3 can also increase in CD4+CD25+ T cells, apparently through a post-transcriptional mechanism16. This indicates a potentially important link between TCR signalling and the level of FOXP3 expression.
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS
ARREST
Any process by which progression through the cell cycle is halted during one of the normal phases — G1 (gap 1), S (synthesis), G2 (gap 2) or M (mitosis). WNT PROTEINS
WNTs are glycoproteins related to the Drosophila melanogaster protein Wingless, a ligand that regulates the temporal and spatial development of the embryo. WNT-mediated signalling has been shown to regulate cell-fate determination, proliferation, adhesion, migration and polarity during development. In addition to their crucial role in embryogenesis, WNTs and their downstream signalling molecules have been implicated in tumorigenesis and have causative roles in human colon cancers. BONE MORPHOGENETIC PROTEINS
(BMPs). The genes encoding BMPs constitute a subfamily of the transforming growth factor-β gene superfamily. BMPs have a crucial role in the modulation of mesenchymal differentiation and the induction of cartilage and bone formation.
Although its precise role remains controversial, TGF-β has been implicated in the regulation of TReg-cell function23. Recent studies have shown that the conversion of human and mouse peripheral naive CD4+CD25– T cells to CD4+CD25+ TReg cells correlates with TGF-β-mediated induction of Foxp3 expression when cells are stimulated through the TCR and CD28 (REFS 24,25). By contrast, Peng et al. found that TGF-β promotes the clonal expansion of a FOXP3-expressing CD4+CD25+ TReg-cell pool but not a pool of CD25 – T cells26. The mechanisms underlying this unusual pro-proliferative role of TGF-β are unknown, but a recent study has indicated that there might be a link between FOXP3 and SMAD7 (mothers against decapentaplegic homologue 7), which is an inhibitor of TGF-β-mediated signalling25. In primary CD4+ T cells, mRNA encoding SMAD7 was induced by TGF-β in control cells, but this was abrogated in cells ectopically expressing FOXP3 (REF. 25). This indicates that FOXP3 might induce a positive autoregulatory loop of TGF-βmediated signalling in the absence of SMAD7 (FIG. 2b). This might function to stabilize the regulatory phenotype of these TReg cells induced in the periphery or to expand the FOXP3-expressing population of TReg cells. Mamura and co-workers have recently shown that CD4+CD25+ T cells that express FOXP3 develop normally in TGF-β1deficient mice27. These TGF-β1-deficient CD4+CD25+ T cells were found to be functional as suppressors, but they did have a reduced capacity to suppress inflammation in vivo. So, the in vivo relevance of TGF-β in the regulation of TReg-cell clonal expansion and function is still unclear. Most recently, a report showed that oestrogen can also increase FOXP3 expression both in vitro and in vivo, and this correlates with the clonal expansion of the TReg-cell compartment28. During pregnancy, there is an increased challenge to peripheral tolerance from fetal antigens. Concomitantly, levels of oestrogen rise, and it is possible that oestrogen-mediated induction of FOXP3 expression might be responsible for expanding the TReg-cell compartment, thereby resulting in increased immunosuppression and preventing fetal rejection (FIG. 2b). In summary, although FOXP3 seems to be crucial for the generation of TReg cells, the molecular mechanisms that regulate the expression and function of this transcription factor remain elusive. TGF-β probably has an important role in modulating the expression of FOXP3; however, it is probable that T cells can respond to various extracellular signals that result in the upregulation of FOXP3 expression and the clonal expansion of TReg-cell populations when appropriate. Identifying these signals, and importantly the target genes of FOXP3 activation, is crucial for understanding the function of TReg cells in the maintenance of immune tolerance. FOXN1 and thymic development
The development and selection of T cells in the thymus requires both cell-intrinsic signals and extrinsic signals delivered by the stroma of the thymus. In the stroma, thymic epithelial cells (TECs) create the correct
NATURE REVIEWS | IMMUNOLOGY
microenvironment for the unique function of the thymus29. Histologically, the stroma can be divided into the cortex and the medulla, which contain distinct types of TEC. These distinct types of TEC support, in a spatially defined manner, the different stages of T-cell development. The ‘nude’ phenotype is due to a recessive mutation in the Foxn1 gene that causes hairlessness and, importantly, also results in congenital athymia in mice, rats and humans30. A nucleotide deletion in the Foxn1 gene of nude mice causes the reading frame to shift, resulting in loss of the DNA-binding domain of this FOX transcription factor. Athymia resulting from a loss of Foxn1 function is characterized by the absence of all of the main TEC types, indicating that Foxn1 regulates the growth and differentiation of all TECs in a cell-autonomous manner. In nude mice, cells that are thought to be the progenitors of TECs undergo ARREST at a stage corresponding to a putative multipotent TEC progenitor29. So, what could be the functional targets of FOXN1 that regulate thymic development? Microdissection-based geneexpression profiling has been carried out comparing wild-type and nude mice to identify FOXN1-regulated genes31. In addition to several as-yet-uncharacterized cDNAs, a member of the B7 family, PD1 (programmed death 1) ligand, was identified, which binds to the immunoinhibitory receptor PD1. Expression of mRNA that encodes PD1 is restricted to the thymus in adult mice, and cell-surface expression of PD1 has been shown on activated T and B cells, myeloid-lineage cells and also endothelial cells treated with inflammatory agents such as interferon-γ (IFN-γ) or TNF. In the thymus, PD1 is thought to have a role in positive selection and prevention of autoimmunity32. However, the relevance of this observation to the development of the athymic nude phenotype remains unclear. At present, there is limited understanding of the molecular mechanisms by which FOXN1 activity is regulated. Recently, WNT PROTEINS and BONE MORPHOGENETIC PROTEINS (BMPs) have been implicated in the transcriptional control of Foxn1 expression33. BMPs are members of the TGF-β superfamily and have been implicated in cell-fate determination and patterning of the embryo. Treatment of intact thymic lobes with BMP4 results in upregulation of Foxn1 expression. It is suggested that FOXN1 might subsequently upregulate the expression of fibroblast growth factor (FGF) receptors, which then modulate the thymic stroma and regulate thymopoiesis34 (FIG. 3). WNTs are a large group of secreted glycoproteins that have important roles in cell-fate specification35. They are expressed by TECs and developing thymocytes, and co-culture experiments and transfection studies show that WNT-mediated signalling can induce Foxn1 transcription33. Overexpression of β-catenin (a downstream target of WNT) was in itself sufficient to increase the expression of Foxn1. In addition, this study showed that WNT-induced signalling through phosphatidylinositol 3-kinase (PI3K) contributes to Foxn1 expression33. PI3K can inhibit glycogen-synthase kinase 3β, the kinase that phosphorylates and destabilizes
VOLUME 4 | NOVEMBER 2004 | 8 9 3 ©2004 Nature Publishing Group
REVIEWS
Stromal cell
FGF7 and FGF10
BMP4
WNT frizzled
BMP receptor
FGFR2
Thymic epithelial cell P Degradation
β-catenin
PI3K
TCF/LEF β-catenin
?
FOXN1
FOXN1
FGFR2
Nucleus
Figure 3 | Thymic development and FOXN1. WNT proteins and bone morphogenetic proteins (BMPs) have been implicated in the transcriptional control of forkhead box N1 (FOXN1). BMP4 has been shown to upregulate the expression of FOXN1, and it has been suggested that FOXN1 might subsequently upregulate the expression of fibroblast growth factor (FGF) receptors (FGFRs), which are then responsible for modulating the thymic stroma and regulating thymopoiesis. WNT proteins expressed by thymic epithelial cells and developing thymocytes can induce the expression of FOXN1. This seems to result from stabilization of β-catenin, and it can be further regulated by phosphatidylinositol 3-kinase (PI3K) through an undefined mechanism. TCF/LEF, member of the T-cell factor (TCF)/lymphoid-enhancer-binding factor (LEF) family.
β-catenin. Inhibition of PI3K, with the specific pharmacological inhibitor LY294002, resulted in decreased transcription of Foxn1. However, the precise mechanism by which PI3K is involved in regulating Foxn1 expression is unclear (FIG. 3). A study investigating the role of FOXN1 in keratinocyte terminal differentiation found that FOXN1 protein expression is repressed by extracellular signal-regulated kinase 1 (ERK1) and ERK2 (also known as p44 MAPK and p42 MAPK, respectively)36. In this system, it is proposed that, before the initiation of keratinocyte differentiation, the level of FOXN1 expression is suppressed by the ERK signalling cascade and that, at the start of keratinocyte differentiation, inactivation of this cascade allows the induction of FOXN1 expression, thereby initiating the early terminal stages of differentiation. Because ERK activity is important in the regulation of proliferation, this supports a link between inhibition of proliferation and induction of terminal differentiation of epithelial cells. Whether ERK has a similar regulatory role in TEC differentiation remains to be seen. Recently, a Foxn1-knock-in mutant (denoted Foxn1∆/∆) — the gene product of which lacks the amino (N)-terminal regulatory domain preceding the
894
| NOVEMBER 2004 | VOLUME 4
DNA-binding domain — was described37. Foxn1∆/∆ mice seem to be normal with respect to hair and skin but show a thymus-specific phenotype that is milder than that observed in the case of complete loss of expression of Foxn1. In Foxn1∆/∆ mice, TEC differentiation is initiated but does not progress, apparently being halted at the stage at which TEC–thymocyte crosstalk is required for TEC differentiation. As a result, some lymphoid progenitors do enter the thymus, but development of these progenitors is severely impaired. The Foxn1∆/∆ mice do not show signs of autoimmune disease, indicating either that negative selection is occurring normally or that there is a decreased repertoire of CD4+ T cells being produced37. These results indicate that the N-terminal region of FOXN1 is involved in a thymus-specific function and that FOXN1 expression is not only required for the initial stages of TEC development but also for the later stages. Interestingly, evolutionary evidence supports a specific function for the N-terminal domain of FOXN1 in the thymus. The cephalochordate Branchiostoma lanceolatum does not have a thymus or hair, but it has a Foxn1 orthologue that is conserved with vertebrate homologues with respect to the DNA-binding domain but not the N-terminal regulatory domain38. Although FOXN1 expression seems to be regulated by factors that are important in cell-fate specification (such as WNTs), the functional targets of FOXN1 activity still remain unclear. The observation that, in Foxn1∆/∆ mice, TEC development is halted at the stage of TEC–thymocyte crosstalk indicates that thymocyte-derived signals might affect the N-terminal domain of FOXN1. Alternatively, it is also possible that the N-terminal domain of FOXN1 interacts with cofactors that are crucial for the transactivation of specific target genes. Understanding the precise role of this functional domain will not only shed light on the process of TEC differentiation but might also highlight yet another possible mode of FOX-mediated regulation. FOXJ1 and T-helper-1-cell activation
FOXJ1 is a FOX-family member the expression of which was, until recently, thought to be restricted to the lungs, spermatids, oviducts and choroid plexus — all structures that contain ciliated cells39,40. FOXJ1 seems to be a potent transcriptional activator that can bind a DNA sequence distinct from that recognized by other FOX transcription factors41. Foxj1–/– mice die during embryonic development or within a few days of birth, and any surviving animals are readily distinguished from their litter-mates by their smaller size, wasted appearance and poor postnatal weight gain42,43. These animals have both severe internal abnormalities, with reversal of abdominal organs, and a complete absence of ciliated epithelia. Although FOXJ1 has previously been linked with epithelial-cell development in the lungs, its role in the differentiation of other cell types has only recently been characterized. Peng and co-workers made the surprising observation that Foxj1 expression was markedly downregulated by lymphocytes isolated from mice prone to
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS
T-cell stimulation
Cytoplasm P
IKK-γ
P
IKK-α IKK-β
P
P FOXJ1
P
P
P
P
IκB
FOXO3A T-cell quiescence
p50
P
IκB
p65
Proteasomal degradation
NF-κB ?
FOXJ1 Forkhead box
?
IκB
p50
p65
IL-2 IFN-γ Cyclin D
T-cell activation
NF-κB-response element
Nucleus
Figure 4 | Regulation of TH1-cell activation by FOXJ1 and FOXO3A. Mice with a targeted mutation in the gene encoding forkhead box J1 (FOXJ1) have systemic autoimmune inflammation, including activated CD4+ T cells. In Foxj1–/– mice, nuclear factor-κB (NF-κB) activity was found to be greatly increased, resulting in the overproduction of T helper 1 (TH1)cell cytokines. These mice also had reduced levels of the β-subunit of inhibitor of NF-κB (IκBβ). It is proposed that FOXJ1 can directly induce the expression of IκB, resulting in repression of NF-κB activation. This can be further modulated by FOXO3A. IκB kinase (IKK) can phosphorylate IκB, resulting in its proteasomal degradation and the activation of NF-κB. IKK can also phosphorylate and thereby inhibit FOXO3A. Importantly, FOXO3A can also inhibit NF-κB activity through indirectly modulating IκB expression, possibly by interacting with FOXJ1. So, T-cell stimulation results in a decrease in expression of FOXJ1 and in inactivation of FOXO3A, leading to NF-κB activation, cytokine-gene transcription and T-cell activation. IFN-γ, interferon-γ; IL-2, interleukin-2.
systemic lupus erythematosus (SLE), indicating a possible role for FOXJ1 in the prevention of autoimmune reactions44. Because Foxj1–/– mice die in utero or in the early neonatal period, a fetal-liver chimerization approach was used to allow the generation of a Foxj1–/– lymphoid system. Mice with this targeted mutation of Foxj1 showed evidence of systemic autoimmune inflammation. Although these Foxj1–/– chimeric mice had relatively normal lymphocyte development, their CD4+ T cells were found to have an activated phenotype. Indeed, when stimulated with CD3-specific antibody, CD4+ T helper (TH) cells from these Foxj1–/– chimeric mice produced more IL-2 and IFN-γ, but not IL-4, than wild-type mice, indicating skewing to a TH1-type response. Furthermore, during primary stimulation, Foxj1–/– CD4+ T cells upregulated expression of the TH1-cell transcription factor T-bet, which is also consistent with a TH1-cell bias44. How could FOXJ1 regulate the proliferation and differentiation of CD4+ TH cells? One possibility is that it could directly regulate the genes that are involved in these processes. Another possibilty is that it could modulate the function of other crucial transcription factors (FIG. 4). NFAT and nuclear factor-κB (NF-κB) have both been shown to have an important role in the differentiation of CD4+ TH cells that is induced following TCR NATURE REVIEWS | IMMUNOLOGY
ligation45,46. Because other FOX transcription factors have been found to be transcriptional repressors, it is possible that FOXJ1 could inhibit NF-κB- or NFATmediated transcription. NF-κB has been shown to be crucial for the production of IFN-γ, and it is thought to be required for TH1-cell responses47. Constitutive activation of this transcription factor is found in various inflammatory conditions: for example, in rheumatoid arthritis. FOXJ1 was found to inhibit both constitutive and TNF-induced NF-κB activity, but not NFAT activity, in a lymphoma cell line44. In Foxj1–/– CD4+ T cells, there were increased levels of expression of NF-κB target genes, including those encoding cyclin D1 (a cellcycle regulator), GADD45β (growth arrest and DNAdamage-inducible 45β), IL-2 and IFN-γ. The activity of NF-κB is regulated through binding the cytosolic inhibitor protein IκB (inhibitor of NF-κB), which blocks nuclear translocation. Overexpression of FOXJ1 was shown to result in a marked upregulation of expression of the β-subunit of IκB at both the mRNA and protein levels44. The expression pattern of IκBβ also correlates with that of Foxj1, being reduced in response to T-cell stimulation with CD3-specific antibody or IL-2. These data support a model in which FOXJ1 regulates NF-κB activity and CD4+ TH-cell activation through modulating the expression levels of IκBβ (FIG. 4). It is still not known what could trigger T cells to downregulate their expression of FOXJ1. Brody and coworkers have reported analysis of transcriptional regulatory regions in the 5′ flanking region of Foxj1 (REF. 48). Potential regulatory elements include an NF-κB-binding site, an AP2 (activator protein 2)-binding site, an AP1binding site and E-box motifs. The NF-κB-binding site could regulate a negative-feedback pathway by which NF-κB could downregulate its own activity by upregulating the expression of FOXJ1. This would result in a subsequent increase in transcription of IκB and in inhibition of NF-κB-mediated transcription. The FOS and JUN transcription factors form tight complexes of homo- or heterodimers (known as AP1) that bind specifically to the AP1-recognition sequence. Recent studies have determined that the strength of TCR signalling can lead to differential expression and/or activation of AP1 components49,50, so it is possible that TCR-mediated changes in AP1 activity are crucial for the downregulation of Foxj1 expression that is observed after activation of naive CD4+ T cells. The initial observation that Foxj1 expression was downregulated in mice prone to SLE led to the suggestion that FOXJ1 might have a role in the prevention of autoimmune reactions. Indeed, Foxj1–/– chimeric mice have systemic autoimmune inflammation, including activated T cells. As discussed earlier, this seems to result from the ability of FOXJ1 to inhibit the transcription factor NF-κB, which is itself known to have a role in the production of inflammatory cytokines. The current data indicate that modulation of the function of FOXJ1 might have important therapeutic implications for various inflammatory disorders.
VOLUME 4 | NOVEMBER 2004 | 8 9 5 ©2004 Nature Publishing Group
REVIEWS
a
b
PKB
SGK
IKK
P
P
Cytoplasm PI3K
TNF
PKB
IKK
P SGK
FOXO1 T24 FOXO3A T32 FOXO4 T28? FOXO
BIM FASL
Nucleus RBL2
Apoptosis BCL-6 TRAIL
SOD2
Cyclin G2
Catalase p27
Cyclin D
c
P
FKH S256 S319 S253 S315 S193 S258
• NF-κB • ER • Glucocorticoid receptor • Progesterone receptor
S644
• STAT3 • ER • Retinoic acid • Thyroid hormone • SMAD3/SMAD4
GADD45 FOXO Cell protection
Cell-cycle arrest
• PPAR-γ-C1α • SMAD3/SMAD4
• FOXG1 • ER
Figure 5 | Regulation and function of FOXO transcription factors. a | Forkhead box O (FOXO) transcription factors are regulated through direct phosphorylation by several intracellular kinases, including serum/glucocorticoid-regulated kinase (SGK), protein kinase B (PKB) and IKK (inhibitor of nuclear factor-κB (NF-κB) kinase). This results in inhibition of DNA binding and nuclear export (PKB and SGK) or targeting for proteasomal degradation (IKK), thereby preventing FOXO-mediated transcription. In resting, unstimulated cells, FOXO transcription factors are active and can modulate the expression of genes regulating diverse processes, including proliferation, apoptosis and response to cellular stress. b | Schematic representation of the phosphorylation sites in FOXO and the respective kinases that mediate this phosphorylation. The identity of the phosphorylated amino acids and their locations are indicated for several members of the FOXO subfamily. c | FOXO proteins have been shown to interact with a large variety of transcription factors. These interactions can result in activation or repression of either binding partner as indicated. BCL-6, B-cell lymphoma 6; BIM, BCL-2-interacting mediator of cell death; ER, oestrogen receptor; FASL, FAS ligand; FKH, forkhead DNA-binding domain; GADD45, growth arrest and DNAdamage-inducible 45; PPAR-γ-C1α, peroxisome-proliferator-activated receptor-γ co-activator 1α; PI3K, phosphatidylinositol 3-kinase; RBL2, retinoblastoma-like 2; SMAD, mothers against decapentaplegic homologue; SOD2, superoxide dismutase 2; STAT3, signal transducer and activator of transcription 3; TNF, tumour-necrosis factor; TRAIL, TNF-related apoptosis-inducing ligand.
FOXOs and immune homeostasis
FOX transcription factors of the O class (FOXOs) have been the subject of a recent burst of research interest: first, because the Caenorhabditis elegans FOXO homologue DAF-16 seems to be an important regulator of longevity in this organism; and second, because FOXO1, FOXO3A and FOXO4 are regulated by the PI3K and PKB (also known as AKT) signalling pathway in mammalian cells (FIG. 5a). Several recent reviews have highlighted the role of DAF-16 or FOXOs in cellular signalling and disease51–53. Here, we restrict our discussion to recent advances in our understanding of the biological role of FOXO proteins in the immune system and, in particular, the regulation of these transcription factors in T and B cells. FOXO proteins are ubiquitously expressed, but there is some tissue-specific expression of different isoforms. In many cell types, FOXO proteins might function in a redundant manner, and for in vitro studies, no clear functional differences between the isoforms have been reported with respect to gene regulation and cellular responses. Whereas FOXO1 is relatively ubiquitously expressed, recent reports indicate that FOXO3A is probably the main isoform that is expressed by cells of the immune system54. However, there have been no published studies that have carefully investigated the expression patterns of FOXOs in myeloid- and lymphoid-cell lineages. Here, we focus
896
| NOVEMBER 2004 | VOLUME 4
mainly on FOXO3A, but it is probable that other FOXO isoforms also have a role in the regulation of immune-cell function. Despite the overlapping tissue expression, and the apparent redundancy of function in vitro, genetic loss of the different FOXOs does result in specific phenotypes. Foxo3a –/– mice provide a starting point for understanding the general roles of FOXOs that might be relevant to immune homeostasis. Foxo3a –/– mice were initially described to have a defect in ovarian function owing to the accelerated differentiation of ovarian follicles, which rapidly depletes the pool of these cells55. Primordial follicles can be regarded as being in a state of developmental arrest, and to maintain this arrest, FOXO3A seems to be required. In mammalian cells in culture, activation of FOXO3A also induces a reversible arrest in the G1 (gap 1) phase of the cell cycle56–58. To maintain this quiescent state, it is important that arrested cells can protect themselves against possible damaging insults, such as oxidative stress. In C. elegans and mammalian cells, the gene that encodes superoxide dismutase 2 (SOD2; also known as MnSOD) — which is involved in protection against oxidative stress — is regulated by DAF-16 and FOXO3A, respectively, and it is required for maintaining quiescence59. Although Foxo3a –/– mice are initially healthy, after 8 months they develop lymphoproliferative disease with multi-organ infiltrates, resulting in
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS
ANERGY
A state of T cells that have been stimulated through their T-cell receptors in the absence of ligation of CD28. On restimulation, these T cells are unable to produce interleukin-2 or to proliferate, even in the presence of co-stimulatory signals. PROGRAMMED CELL DEATH
A common form of cell death, which is also known as apoptosis. Many physiological and developmental stimuli cause apoptosis, and this mechanism is frequently used to delete unwanted, superfluous or potentially harmful cells, such as those undergoing transformation. Apoptosis involves cell shrinkage, chromatin condensation in the periphery of the nucleus, plasma-membrane blebbing and DNA fragmentation into segments of about 180 base pairs. Eventually, the cell breaks up into many membrane-bound ‘apoptotic bodies’, which are phagocytosed by neighbouring cells. HYPERMORPHIC
A type of mutation in which the altered gene product has an increased level of activity or in which the wild-type gene product is expressed at an increased level.
multisystem inflammation54 : increased lymphocyte proliferation was evident in these animals, resulting in an enlarged spleen and lymph nodes. Furthermore, Lin et al. reported that FOXO3A activity was greatly reduced in T cells from SLE-prone mouse strains compared with non-autoimmune animals54. These data indicate that, in vivo, FOXO3A can regulate lymphocyte homeostasis, and the molecular mechanism underlying these observations is discussed next (FIG. 4). Within the immune system, various cells can become reversibly arrested, being able to re-enter proliferation and/or differentiation programmes. A well-known example of this is anergic T cells60. Induction of T-cell ANERGY has been associated with increased expression of the cell-cycle inhibitor p27 (also known as KIP1)61. p27 is a direct transcriptional target of all FOXOs, indicating a possible role for FOXOs in the maintenance of this immune-unresponsive state58,62. Co-stimulation of T cells through cell-surface CD28 can block the induction of anergy, and ligation of CD28 has also been shown to both activate the PI3K–PKB signalling pathway (which is known to regulate FOXOs) and reduce p27 expression levels63. However, FOXOs have not been definitively determined to have a role in T-cell anergy, and it is unclear whether the observed events are crucial for costimulation through CD28 and reversal of the anergic state64. FOXO3A has also been shown to regulate cell division through a cyclin-G2-dependent mechanism65. Cyclin G2 inhibits cell-cycle progression, and it has been shown to have a role in maintaining the quiescent state of differentiated cells. In lymphocytes, cyclin G2 has been proposed to be an important negative regulator of proliferation: for example, it is upregulated during B-cell receptor (BCR)-mediated growth arrest66. Similar to p27, regulation of cyclin G2 expression by FOXO3A might be important in the regulation of lymphocyte anergy. It is also probable that FOXO-mediated induction of expression of stress-response genes (such as those encoding SOD2, catalase or GADD45) could help to maintain anergic T cells in a quiescent state without the acquisition of cellular damage. The regulation of PROGRAMMED CELL DEATH has an important role in the homeostasis of the immune system. For example, the number of T cells that eventually enters the periphery is only a fraction of the total that is initially generated, and to enable fast and robust responses, the haematopoietic system is continuously counterbalancing stem-cell renewal and differentiation with apoptosis. During all stages of T- and B-cell development, lineage-specific survival factors and/or deathinducing factors are operating. Important survival factors such as IL-2 and IL-3 have been shown to inactivate FOXO3A through PI3K–PKB-mediated phosphorylation62,67, and activated FOXO3A has been shown to control the expression of several pro-apoptotic genes, including FAS ligand (also known as CD95 ligand), BCL-6 (B-cell lymphoma 6) and BIM (BCL-2-interacting mediator of cell death). For example, specific ligandindependent activation of FOXO3A is sufficient to induce BIM expression and to recapitulate all known elements of the apoptotic programme67. In TH2 cells,
NATURE REVIEWS | IMMUNOLOGY
co-stimulatory signals induced by ligation of CTLA4 can activate PI3K and result in the inhibition of FOXO3A, leading to increased cell survival68. Consistent with the role of the PI3K–PKB signalling pathway and FOXOmediated signalling in balancing anti- and pro-apoptotic signalling in the immune system, respectively, expression of a HYPERMORPHIC PKB allele (known as gagPKB) by T cells results in increased lymphocyte survival, most probably owing to the upregulation of expression of BCL-XL69. Interestingly, the FOXO target BCL-6 has been shown to function as a repressor of BCL-XL expression, so inactivation of FOXO in gagPKB-expressing T cells (through PKB-mediated phosphorylation) might relieve the inhibition of BCL-XL expression by BCL-6, thereby resulting in increased survival70. However, although deletion of Foxo3a results in increased lymphocyte proliferation, Foxo3a –/– lymphocytes have a surprising lack of apoptotic defects54. It is possible that FOXO3A is not a physiological regulator of lymphocyte apoptosis in vivo, but we think it is more probable that the lack of FOXO3A is compensated for by FOXO1, which can also regulate programmed cell death. The nuclear exclusion and inhibition of FOXOs seems not to be mediated exclusively through PKBmediated phosphorylation (FIG. 5b). A recent study found that IκB kinase-β (IKK-β) could also induce phosphorylation of FOXO3A, resulting in nuclear exclusion accompanied by proteasome-mediated degradation71. IKK-β has itself been implicated in various immune functions, including the maintenance of mature B cells72. Most of these functions have been attributed to the role of IKK-β in regulating NF-κB transcriptional activity, but it now seems that some of these functions might be mediated by FOXOs. There might also be a feedback loop between FOXOs and NF-κB. In Foxo3a –/– cells, NF-κB transcriptional activity is increased, indicating that FOXO3A expression might modulate NF-κB activity54. Although the molecular mechanism by which this could occur is unclear, in Foxo3a –/– cells, expression of some Iκβ isozymes (IκBβ and IκBε) is reduced. Alternatively, FOXO3A might regulate other proposed regulators of IκBβ and IκBε gene expression such as FOXJ1, although FOXJ1 has been shown to regulate IκBβ but not IκBε44 (FIG. 4). Until recently, most studies of FOXOs have been carried out using established T- or B-cell lines, and these studies might not accurately represent the situation in vivo. Recent studies are extending the established paradigms by using primary immune cells. Yusuf et al. have shown that the inactivation of FOXO1 and FOXO3A by PI3K–PKB signalling is functionally important in BCR-triggered proliferation of primary B cells 73. Consistent with this, they also observed regulation of several known FOXO target genes — those encoding retinoblastoma-like 2 (RBL2; also known as p130) and cyclin G2 — through BCR signalling. Finally, although most studies of FOXO function in the immune system have focused on the T- and B-cell lineages of the haematopoietic compartment, FOXO function is not restricted to these lineages. During erythroid differentiation of the p53-deficient I/11 erythroid cell line,
VOLUME 4 | NOVEMBER 2004 | 8 9 7 ©2004 Nature Publishing Group
REVIEWS
BCR-ABL
A tyrosine-kinase oncogene. The Abelson leukaemia-virus protein (ABL) is fused with the breakpoint-cluster region (BCR) in the Philadelphia-chromosome translocation found in chronic myeloid leukaemia.
increased expression and nuclear localization of FOXO3A is observed, indicating that FOXO3A activation is required for erythropoiesis. In these I/11 cells, FOXO3A regulates expression of B-cell translocation gene 1 (BTG1), which encodes an arginine methyltransferase74. The precise molecular targets of this methyltransferase activity and their potential role in erythropoiesis remain unclear at present. Taken together, the studies discussed here start to provide a basis for our understanding of the PI3K–PKB signalling module and the role of FOXO inactivation in the immune system. Central to this is the reciprocal effect of FOXO and PI3K–PKB on cell-cycle progression and survival, which allows regulated switching between an arrested state and a proliferative state by extracellular signals. Importantly, in immune cells, it seems that FOXO3A activity tends to be inversely correlated with cellular activation. FOX hunting: the future of forkhead
The importance of FOX transcription factors in the immune system is highlighted not only by the various null-mutant mice described here but also by the linkage of different FOX proteins to specific disease phenotypes. The autoimmune disorder of scurfy mice and patients with IPEX results from mutations in the gene encoding FOXP3, and mice with a Foxj1–/– lymphoid system have systemic autoimmune inflammation. In addition, nude mice lack Foxn1, and a human equivalent of this nude phenotype has also been described75. But what about FOXOs? Spontaneous autoimmune disease and T-cell hyperactivity were observed in Foxo3a –/– animals, and diminished FOXO3A activity was found in SLE-prone strains of mice54. So far, no diseases have been directly linked to mutations in specific FOXO transcription factors, perhaps owing to their more generalized function in the regulation of cellular proliferation and survival. However, a recent study showed that, in an in vitro model of acute myeloid leukaemia, constitutive inhibition of FOXO3A promotes cellular proliferation, inhibits apoptosis and blocks the differentiation of transformed cells76. Similarly, BCR-ABL-induced loss of FOXO3A function has been proposed to be involved in the oncogenic transformation that occurs in chronic myeloid leukaemia, partially through downregulation of
1. 2.
3.
4.
5.
6.
898
Kaufmann, E. & Knochel, W. Five years on the wings of fork head. Mech. Dev. 57, 3–20 (1996). Weigel, D., Jurgens, G., Kuttner, F., Seifert, E. & Jackle, H. The homeotic gene fork head encodes a nuclear protein and is expressed in the terminal regions of the Drosophila embryo. Cell 57, 645–658 (1989). Lai, E., Prezioso, V. R., Smith, E., Litvin, O., Costa, R. H. & Darnell, J. E. HNF-3A, a hepatocyte-enriched transcription factor of novel structure is regulated transcriptionally. Genes Dev. 4, 1427–1436 (1990). Kaestner, K. H., Knochel, W. & Martinez, D. E. Unified nomenclature for the winged helix/forkhead transcription factors. Genes Dev. 14, 142–146 (2000). Clark, K. L., Halay, E. D., Lai, E. & Burley, S. K. Co-crystal structure of the HNF-3/fork head DNArecognition motif resembles histone H5. Nature 364, 412–420 (1993). Lehmann, O. J., Sowden, J. C., Carlsson, P., Jordan, T. & Bhattacharya, S. S. Fox’s in development and disease. Trends Genet. 19, 339–344 (2003).
expression of the cell-cycle inhibitor p27 (REF. 77). FOXOs themselves could be considered potential tumour suppressors, with an inhibitory mutation possibly contributing to neoplasia. So far, no such role has been described, but future studies might identify disease-causing mutations in members of the FOXO subfamily that are similar to those found for other FOX transcription factors. Although studies have focused on the regulation of FOX function and the generation of null-mutant mice, little is known about the specific target genes of these proteins. Owing to the general function of FOXO proteins in the regulation of cell survival and proliferation of various cell types, several FOXO-regulated genes have now been identified. However, for FOXP3, FOXN1 and FOXJ1, the identification of target genes is still in its infancy. This is a crucial step for fully appreciating the important function of these transcription factors in regulating immune homeostasis. FOX proteins can function as monomeric transcription factors and hetero- or homodimerization is not required for DNA-binding specificity. However, studies using FOXO mutants that can no longer bind DNA have indicated that FOXOs can also regulate transcription indirectly78. In support of this, several transcription factors have recently been proposed to bind FOXOs with specific functional consequences (FIG. 5c). FOXP3 contains a leucine-zipper dimerization domain, indicating that it can also associate with accessory molecules. Furthermore, a point mutation in this region was found in a family with IPEX, indicating that it is indeed crucial for normal FOXP3 function15. The hunt for FOX-associating proteins is likely to be crucial for clarifying the molecular mechanisms underlying the diverse functions of the FOX proteins. In the past few years, we have started to fully appreciate the importance of this large family of FOX transcription factors for immune homeostasis. The evolutionary conservation and diverse biological functions of the family members highlight the importance of these proteins in developmental processes. With a greater understanding of the molecular mechanisms that regulate FOX activity and the identification of specific target genes, the design of targeted therapies for immune disorders that directly affect FOX function might become a reality.
7.
Maloy, K. J. & Powrie, F. Regulatory T cells in the control of immune pathology. Nature Immunol. 2, 816–822 (2001). Shevach, E. M. CD4+CD25+ suppressor T cells: more questions than answers Nature Rev. Immunol. 2, 389–400 (2002). 9. Blair, P. J. et al. CD4+CD8– T cells are the effector cells in disease pathogenesis in the scurfy (sf) mouse. J. Immunol. 153, 3764–3774 (1994). 10. Godfrey, V. L., Wilkinson, J. E., Rinchik, E. M. & Russell, L. B. Fatal lymphoreticular disease in the scurfy (sf) mouse requires T cells that mature in a sf thymic environment: potential model for thymic education. Proc. Natl Acad. Sci. USA 88, 5528–5532 (1991). 11. Brunkow, M. E. et al. Disruption of a new forkhead/wingedhelix protein, scurfin, results in the fatal lymphoproliferative disorder of the scurfy mouse. Nature Genet. 27, 68–73 (2001). This study identified the gene that is defective in scurfy mice by combining high-resolution genetic and physical mapping with large-scale sequence analysis.
8.
| NOVEMBER 2004 | VOLUME 4
12. 13.
14.
15.
16.
Genetic complementation showed that the protein product of Foxp3 is essential for normal immune homeostasis. Ramsdell, F. Foxp3 and natural regulatory T cells: key to a cell lineage? Immunity 19, 165–168 (2003). Wildin, R. S. et al. X-linked neonatal diabetes mellitus, enteropathy and endocrinopathy syndrome is the human equivalent of mouse scurfy. Nature Genet. 27, 18–20 (2001). Bennett, C. L. et al. The immune dysregulation, polyendocrinopathy, enteropathy, X-linked syndrome (IPEX) is caused by mutations of FOXP3. Nature Genet. 27, 20–21 (2001). Chatila, T. A. et al. JM2, encoding a fork head-related protein, is mutated in X-linked autoimmunity-allergic disregulation syndrome. J. Clin. Invest. 106, R75–R81 (2000). Fontenot, J. D., Gavin, M. A. & Rudensky, A. Y. Foxp3 programs the development and function of CD4+CD25+ regulatory T cells. Nature Immunol. 4, 330–336 (2003).
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS 17. Hori, S., Nomura, T. & Sakaguchi, S. Control of regulatory T cell development by the transcription factor Foxp3. Science 299, 1057–1061 (2003). References 16 and 17 showed that FOXP3 is specifically expressed by CD4+CD25+ TReg cells and is required for their development. Furthermore, expression of FOXP3 was found to confer suppressor function on peripheral CD4+CD25– T cells. In addition, reference 17 showed that FOXP3, which encodes a transcription factor that is genetically defective in an autoimmune and inflammatory syndrome in humans and mice, is specifically expressed by naturally arising CD4+ regulatory T cells. 18. Khattri, R. et al. The amount of scurfin protein determines peripheral T cell number and responsiveness. J. Immunol. 167, 6312–6320 (2001). 19. Asseman, C., Mauze, S., Leach, M. W., Coffman, R. L. & Powrie, F. An essential role for interleukin 10 in the function of regulatory T cells that inhibit intestinal inflammation. J. Exp. Med. 190, 995–1004 (1999). 20. Schubert, L. A., Jeffery, E., Zhang, Y., Ramsdell, F. & Ziegler, S. F. Scurfin (FOXP3) acts as a repressor of transcription and regulates T cell activation. J. Biol. Chem. 276, 37672–37679 (2001). 21. Walker, M. R. et al. Induction of FoxP3 and acquisition of T regulatory activity by stimulated human CD4+CD25– T cells. J. Clin. Invest. 112, 1437–1443 (2003). 22. Baecher-Allan, C., Brown, J. A., Freeman, G. J. & Hafler, D. A. CD4+CD25hi regulatory cells in human peripheral blood. J. Immunol. 167, 1245–1253 (2001). 23. Nakamura, K., Kitani, A. & Strober, W. Cell contactdependent immunosuppression by CD4+CD25+ regulatory T cells is mediated by cell surface-bound transforming growth factor β. J. Exp. Med. 194, 629–644 (2001). 24. Chen, W. et al. Conversion of peripheral CD4+CD25– naive T cells to CD4+CD25+ regulatory T cells by TGF-β induction of transcription factor Foxp3. J. Exp. Med. 198, 1875–1886 (2003). 25. Fantini, M. C. et al. TGF-β induces a regulatory phenotype in CD4+CD25– T cells through Foxp3 induction and downregulation of Smad7. J. Immunol. 172, 5149–5153 (2004). This study showed that TGF-β induces a regulatory phenotype in CD4+CD25– T cells through the induction of FOXP3 expression, which can subsequently inhibit SMAD7 and thereby activate a positive autoregulatory loop of TGF-β signalling. 26. Peng, Y., Laouar, Y., Li, M. O., Green, E. A. & Flavell, R. A. TGF-β regulates in vivo expansion of Foxp3-expressing CD4+CD25+ regulatory T cells responsible for protection against diabetes. Proc. Natl Acad. Sci. USA 101, 4572–4577 (2004). 27. Mamura, M. et al. CD28 disruption exacerbates inflammation in Tgf-β1–/– mice: in vivo suppression by CD4+CD25+ regulatory T cells independent of autocrine TGF-β1. Blood 103, 4594–4601 (2004). 28. Polanczyk, M. J. et al. Estrogen drives expansion of the CD4+CD25+ regulatory T cell compartment. J. Immunol. 173, 2227–2230 (2004). 29. Blackburn, C. C. & Manley, N. R. Developing a new paradigm for thymus organogenesis. Nature Rev. Immunol. 4, 278–289 (2004). 30. Nehls, M., Pfeifer, D., Schorpp, M., Hedrich, H. & Boehm, T. New member of the winged-helix protein family disrupted in mouse and rat nude mutations. Nature 372, 103–107 (1994). 31. Bleul, C. C. & Boehm, T. Laser capture microdissectionbased expression profiling identifies PD1-ligand as a target of the nude locus gene product. Eur. J. Immunol. 31, 2497–2503 (2001). 32. Nishimura, H., Honjo, T. & Minato, N. Facilitation of β selection and modification of positive selection in the thymus of PD-1deficient mice. J. Exp. Med. 191, 891–898 (2000). 33. Balciunaite, G. et al. Wnt glycoproteins regulate the expression of FoxN1, the gene defective in nude mice. Nature Immunol. 3, 1102–1108 (2002). This paper reported that secreted WNT glycoproteins expressed by TECs and thymocytes can regulate FOXN1 expression in the epithelium through both autocrine and paracrine mechanisms. 34. Tsai, P. T., Lee, R. A. & Wu, H. BMP4 acts upstream of FGF in modulating thymic stroma and regulating thymopoiesis. Blood 102, 3947–3953 (2003). 35. Staal, F. J. & Clevers, H. C. Wnt signaling in the thymus. Curr. Opin. Immunol. 15, 204–208 (2003). 36. Baxter, R. M. & Brissette, J. L. Role of the nude gene in epithelial terminal differentiation. J. Invest. Dermatol. 118, 303–309 (2002). 37. Su, D. M., Navarre, S., Oh, W. J., Condie, B. G. & Manley, N. R. A domain of Foxn1 required for crosstalkdependent thymic epithelial cell differentiation. Nature Immunol. 4, 1128–1135 (2003).
38. Schlake, T., Schorpp, M. and Boehm, T. Formation of regulator/target gene relationships during evolution. Gene 256, 29–34 (2000). 39. Hackett, B. P. et al. Primary structure of hepatocyte nuclear factor/forkhead homologue 4 and characterization of gene expression in the developing respiratory and reproductive epithelium. Proc. Natl Acad. Sci. USA 92, 4249–4253 (1995). 40. Clevidence, D. E. et al. Members of the HNF-3/forkhead family of transcription factors exhibit distinct cellular expression patterns in lung and regulate the surfactant protein B promoter. Dev. Biol. 166, 195–209 (1994). 41. Lim, L., Zhou, H. & Costa, R. H. The winged helix transcription factor HFH-4 is expressed during choroid plexus epithelial development in the mouse embryo. Proc. Natl Acad. Sci. USA 94, 3094–3099 (1997). 42. Chen, J., Knowles, H. J., Hebert, J. L. & Hackett, B. P. Mutation of the mouse hepatocyte nuclear factor/forkhead homologue 4 gene results in an absence of cilia and random left–right asymmetry. J. Clin. Invest. 102, 1077–1082 (1998). 43. Brody, S. L., Yan, X. H., Wuerffel, M. K., Song, S. K. & Shapiro, S. D. Ciliogenesis and left–right axis defects in forkhead factor HFH-4-null mice. Am. J. Respir. Cell. Mol. Biol. 23, 45–51 (2000). 44. Lin, L., Spoor, M. S., Gerth, A. J., Brody, S. L. & Peng, S. L. Modulation of TH1 activation and inflammation by the NF-κB repressor Foxj1. Science 303, 1017–1020 (2004). This study showed that FOXJ1 can inhibit NF-κB signalling through induction of IκB proteins. These results indicate that FOXJ1 might modulate inflammatory reactions and prevent autoimmunity by antagonizing the transcription of genes that encode pro-inflammatory cytokines. 45. Rao, A., Luo, C. & Hogan, P. G. Transcription factors of the NFAT family: regulation and function. Annu. Rev. Immunol. 15, 707–747 (1997). 46. Li, Q. & Verma, I. M. NF-κB regulation in the immune system. Nature Rev. Immunol. 2, 725–734 (2002). 47. Kojima, H. et al. An essential role for NF-κB in IL-18-induced IFN-γ expression in KG-1 cells. J. Immunol. 162, 5063–5069 (1999). 48. Brody, S. L., Hackett, B. P. & White, R. A. Structural characterization of the mouse Hfh4 gene, a developmentally regulated forkhead family member. Genomics 45, 509–518 (1997). 49. Schade, A. E. & Levine, A. D. Extracellular signal-regulated kinases 1/2 function as integrators of TCR signal strength. J. Immunol. 172, 5828–5832 (2004). 50. Jorritsma, P. J., Brogdon, J. L. & Bottomly, K. Role of TCRinduced extracellular signal-regulated kinase activation in the regulation of early IL-4 expression in naive CD4+ T cells. J. Immunol. 170, 2427–2434 (2003). 51. Burgering, B. M. & Medema, R. H. Decisions on life and death: FOXO forkhead transcription factors are in command when PKB/Akt is off duty. J. Leukoc. Biol. 73, 689–701 (2003). 52. Birkenkamp, K. U. & Coffer, P. J. FOXO transcription factors as regulators of immune homeostasis: molecules to die for? J. Immunol. 171, 1623–1629 (2003). 53. Accili, D. & Arden, K. C. FoxOs at the crossroads of cellular metabolism, differentiation, and transformation. Cell 117, 421–426 (2004). 54. Lin, L., Hron, J. D. & Peng, S. L. Regulation of NF-κB, TH activation, and autoinflammation by the forkhead transcription factor Foxo3a. Immunity 21, 203–213 (2004). This was the first study that identified a role for FOXO transcription factors in immune homeostasis in vivo. FOXO3A can inhibit NF-κB activation and maintain T-cell tolerance through an undefined mechanism. 55. Castrillon, D. H., Miao, L., Kollipara, R., Horner, J. W. & DePinho, R. A. Suppression of ovarian follicle activation in mice by the transcription factor Foxo3a. Science 301, 215–218 (2003). 56. Medema, R. H., Kops, G. J., Bos, J. L. & Burgering, B. M. AFX-like forkhead transcription factors mediate cell-cycle regulation by Ras and PKB through p27kip1. Nature 404, 782–787 (2000). This paper showed that FOXO transcription factors can directly modulate proliferation. 57. Kops, G. J. et al. Control of cell cycle exit and entry by protein kinase B-regulated forkhead transcription factors. Mol. Cell. Biol. 22, 2025–2036 (2002). 58. Dijkers, P. F. et al. Forkhead transcription factor FKHR-L1 modulates cytokine-dependent transcriptional regulation of p27KIP1. Mol. Cell. Biol. 20, 9138–9148 (2000). 59. Kops, G. J. et al. Forkhead transcription factor FOXO3a protects quiescent cells from oxidative stress. Nature 419, 316–321 (2002). 60. Schwartz, R. H. T cell anergy. Annu. Rev. Immunol. 21, 305–334 (2003).
NATURE REVIEWS | IMMUNOLOGY
61. Jackson, S. K., DeLoose, A. & Gilbert, K. M. Induction of anergy in TH1 cells associated with increased levels of cyclin-dependent kinase inhibitors p21Cip1 and p27Kip1. J. Immunol. 166, 952–958 (2001). 62. Stahl, M. et al. The forkhead transcription factor FoxO regulates transcription of p27 Kip1 and Bim in response to IL-2. J. Immunol. 168, 5024–5031 (2002). 63. Appleman, L. J., van Puijenbroek, A. A., Shu, K. M., Nadler, L. M. & Boussiotis, V. A. CD28 costimulation mediates down-regulation of p27 kip1 and cell cycle progression by activation of the PI3K/PKB signaling pathway in primary human T cells. J. Immunol. 168, 2729–2736 (2002). 64. Okkenhaug, K., Bilancio, A., Emery, J. L. & Vanhaesebroeck, B. Phosphoinositide 3-kinase in T cell activation and survival. Biochem. Soc. Trans. 32, 332–335 (2004). 65. Martinez-Gac, L., Marques, M., Garcia, Z., Campanero, M. R. & Carrera, A. C. Control of cyclin G2 mRNA expression by forkhead transcription factors: mechanism for cell cycle control by phosphoinositide 3-kinase and forkhead. Mol. Cell. Biol. 24, 2181–2189 (2004). 66. Horne, M. C. et al. Cyclin G2 is up-regulated during growth inhibition and B cell antigen receptor-mediated cell cycle arrest. J. Biol. Chem. 272, 12650–12661 (1997). 67. Dijkers, P. F. et al. FKHR-L1 can act as a critical effector of cell death induced by cytokine withdrawal: protein kinase B-enhanced cell survival through maintenance of mitochondrial integrity. J. Cell Biol. 156, 531–542 (2002). This paper showed that activation of FOXO3A alone can recapitulate all known elements of the apoptotic programme that are normally induced in lymphocytes by cytokine withdrawal. 68. Pandiyan, P. et al. CD152 (CTLA-4) determines the unequal resistance of TH1 and TH2 cells against activation-induced cell death by a mechanism requiring PI3 kinase function. J. Exp. Med. 199, 831–842 (2004). 69. Jones, R. G. et al. Protein kinase B regulates T lymphocyte survival, nuclear factor κB activation, and Bcl-XL levels in vivo. J. Exp. Med. 191, 1721–1734 (2000). 70. Tang, T. T. et al. The forkhead transcription factor AFX activates apoptosis by induction of the BCL-6 transcriptional repressor. J. Biol. Chem. 277, 14255–14265 (2002). 71. Hu, M. C. et al. IκB kinase promotes tumorigenesis through inhibition of forkhead FOXO3a. Cell 117, 225–237 (2004). 72. Pasparakis, M., Schmidt-Supprian, M. & Rajewsky, K. IκB kinase signaling is essential for maintenance of mature B cells. J. Exp. Med. 196, 743–752 (2002). 73. Yusuf, I., Zhu, X., Kharas, M. G., Chen, J. & Fruman, D. A. Optimal B-cell proliferation requires phosphoinositide 3-kinase-dependent inactivation of FOXO transcription factor. Blood 104, 784–787 (2004). 74. Bakker, W. J. et al. FoxO3a regulates erythroid differentiation and induces BTG1, an activator of protein arginine methyl transferase 1. J. Cell Biol. 164, 175–184 (2004). 75. Frank, J. et al. Exposing the human nude phenotype. Nature 398, 473–474 (1999). 76. Scheijen, B., Ngo, H. T., Kang, H. & Griffin, J. D. FLT3 receptors with internal tandem duplications promote cell viability and proliferation by signaling through Foxo proteins. Oncogene 23, 3338–3349 (2004). 77. Komatsu, N. et al. A member of Forkhead transcription factor FKHRL1 is a downstream effector of STI571-induced cell cycle arrest in BCR-ABL-expressing cells. J. Biol. Chem. 278, 6411–6419 (2003). 78. Ramaswamy, S., Nakamura, N., Sansal, I., Bergeron, L. & Sellers, W. R. A novel mechanism of gene regulation and tumor suppression by the transcription factor FKHR. Cancer Cell 2, 81–91 (2002). 79. Carlsson, P. & Mahlapuu, M. Forkhead transcription factors: key players in development and metabolism. Dev. Biol. 250, 1–23 (2002).
Competing interests statement The authors declare no competing financial interests.
Online links DATABASES The following terms in this article are linked online to: Entrez Gene: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=gene FOXJ1 | FOXN1 | FOXO3A | FOXP3 | IKK | PKB | PI3K | SMAD7 | TGF-β FURTHER INFORMATION Paul Coffer’s laboratory: http://www.pulmoscience.org Boudewijn Burgering’s laboratory: http://www.ruummc.med.uu.nl/ Access to this interactive links box is free online.
VOLUME 4 | NOVEMBER 2004 | 8 9 9 ©2004 Nature Publishing Group
REVIEWS
RECENT DEVELOPMENTS IN THE TRANSCRIPTIONAL REGULATION OF CYTOLYTIC EFFECTOR CELLS Laurie H. Glimcher*‡, Michael J. Townsend*, Brandon M. Sullivan* and Graham M. Lord* Abstract | Transcription factors have a profound influence on both the differentiation and effector function of cells of the immune system. T-bet controls the cytotoxicity of CD8+ T cells and the production of interferon-γ, and it also affects the development and function of natural killer cells and natural killer T cells. Other factors such as eomesodermin, MEF, ETS1 and members of the interferon-regulatory factor family also contribute to the effector function of immune cells. In this review, we focus on recent studies that have shed light on the transcriptional mechanisms that regulate cellular effector function in the immune system. TRANSACTIVATION
The process by which a transcription factor binds the promoter region of a gene and induces its transcription. REPRESSION
The process by which a transcription factor binds the promoter region of a gene and inhibits its transcription.
*Department of Immunology and Infectious Diseases, Harvard School of Public Health, Boston, Massachusetts 02115, USA. ‡ Department of Medicine, Harvard Medical School, Boston, Massachusetts 02115, USA. Correspondence to L.H.G. e-mail:
[email protected] doi:10.1038/nri1490
900
The ability of each cell to have specified functions is determined by its capacity to selectively express a finite set of genes. Furthermore, this coordinated gene expression must be regulated and rapidly induced, even in the context of the large amount of genomic material that is present in all cell types. So, extensive networks have evolved to control gene transcription, providing for both basal and tissue-restricted activity. The elements that construct the framework for transcription include genomic architecture, distinct stretches of DNA sequence, and general and site-specific transcription factors (that is, modular DNA-binding proteins that bind specific elements in the promoter of a gene and control transcription by recruiting other binding partners and RNA polymerase II). Coordinating the specificity of this extensive network depends on transcription factors that can recognize DNA in a sequence-dependent manner1,2. Furthermore, tissuerestricted gene expression relies on lineage-specific transcription factors, some of which can enforce an entire genetic programme in the context of lineage commitment (TABLE 1). Typically, transcription factors are modular proteins, providing multiple functions and points of regulation. Motifs that recognize distinct DNA sequences provide target-gene specificity and are often highly conserved, thereby designating families of transcription factors1. TRANSACTIVATION domains, REPRESSION
| NOVEMBER 2004 | VOLUME 4
domains and protein-interacting domains often flank DNA-binding motifs; these have minimal sequence conservation and are responsible for modulating the activity of the transcription factor, typically as targets of numerous regulatory mechanisms. Lineage commitment of CD4+ and CD8+ T cells is transcriptionally regulated, often by the same factors that mediate T-cell effector function. Indeed, effector function and cytokine polarization often overlap, as is the case for interferon-γ (IFN-γ), particularly for CD4+ T cells. The tissue-specific T-BOX (TBX) TRANSCRIPTION FACTORS T-bet (also known as TBX21) and eomesodermin (EOMES) have recently been shown to regulate the cytolytic effector mechanisms of CD8+ T cells, natural killer (NK) cells and natural killer T (NKT) cells3,4. T-bet was previously implicated in the T HELPER 1 (T 1)-CELL lineage commitment of CD4+ T cells, partly as a result of its transactivation of the hallmark TH1-type cytokine IFN-γ 5. CD4+ T cells have minimal, if any, effector function in the absence of TH1- or TH2-cell polarization, and this subject has recently been comprehensively reviewed6. By contrast, NK, NKT and CD8+ T cells do not require such polarization to initiate their effector mechanism, which is mainly cell killing. NK and NKT cells form part of the innate immune system’s first line of defence against an invading pathogen, and they also function as a bridge to instruct cells of the adaptive H
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS
Table 1 | Transcription factors involved in CD8+ T-cell, NK-cell and NKT-cell function Transcription factor
Effect
References
CD8+ T cells T-bet
IFN-γ production, cytotoxicity generation, and granzyme and perforin production
24, unpublished
EOMES
IFN-γ production, cytotoxicity generation, and granzyme and perforin production
3
STAT4
IFN-γ production (through IL-12) and possibly CTL activity
27
STAT1
IFN-γ production
28
RUNX3
CTL proliferation
29
REL
CTL cytotoxicity and proliferation (driven by IL-2)
30
NK cells T-bet
Terminal maturation, and regulation of perforin and granzyme B expression
4
ETS1
Cell development, cytotoxicity generation, and perforin and IFN-γ production
MEF
Cell development, cytotoxicity generation, and perforin and IFN-γ production
42
IRF2
Cell development and cytotoxicity generation
43
MITF
Cytotoxicity generation and perforin expression
47
C/EBP-γ
Cytotoxicity generation and IFN-γ production
48
41
Vα14i NKT cells T-bet
Terminal maturation and IFN-γ production
MEF
Cell development
42
4
ETS
Cell development and IL-4 production
54
C/EBP-γ, CCAAT/enhancer-binding protein-γ; CTL, cytotoxic T lymphocyte; EOMES, eomesodermin; IFN-γ, interferon-γ; IL, interleukin; IRF2, IFN-regulatory factor 2; MEF, myeloid ELF1 (E74-like factor 1)-like factor; MITF, microphthalmia-associated transcription factor; NK cell, natural killer cell; NKT cell, natural killer T cell; RUNX3, runt-related transcription factor 3; STAT, signal transducer and activator of transcription.
immune system. CD8+ T cells can rapidly mount a strong defence against pathogens and can directly lyse infected cells that display cell-surface pathogen-derived peptide in the context of MHC class I molecules. These cytolytic cell types have common killing mechanisms, which have overlapping programmes of transcriptional control. Some of these transcription factors are expressed in a tissue-specific manner, whereas the expresssion of others is less restricted. The control of the expression of genes involved in cytolytic effector-cell function is the focus of this review. CD8+ T-cell function
T-BOX (TBX) TRANSCRIPTION FACTORS
A family of transcription factors that each contains a DNAbinding domain of 200 amino acids known as the T-box. These factors are usually involved in developmental programmes, and the founding member of the family is Brachyury. T-bet and eomesodermin are members of this family. T HELPER 1 (TH1)-CELL
The definition of a CD4+ T cell that has differentiated into a cell that produces the cytokines interferon-γ and tumournecrosis factor.
In addition to the differentiation of CD4 + effector TH1 cells, adaptive TH1-type immune responses rely on the generation of a CD8+ effector T-cell pool. Following activation, naive CD8+ T cells undergo antigen-driven terminal differentiation in the periphery. Recent studies have shown that the transitions from naive to effector and effector to memory CD8+ T-cell populations are associated with marked changes in gene expression7,8. The naive, antigen-inexperienced CD8+ T-cell precursors undergo genetic remodelling that results in the expression of signature genes central to CD8+ effector T-cell function, including genes that encode cytokines and chemokines, and genes associated with cytolysis. IFN-γ and tumour-necrosis factor (TNF) are the main cytokines produced by differentiated CD8+ effector T cells, and IFN-γ has been shown to have a fundamental role in CD8+ T-cell-mediated immunity. Genetic ablation of the IFN-γ receptor (IFN-γR) results in susceptibility to infection with Listeria monocytogenes and vaccinia virus, even though the activity of cytotoxic
NATURE REVIEWS | IMMUNOLOGY
T lymphocytes (CTLs) is normal9,10. In the absence of STAT1 (signal transducer and activator of transcription 1), which is the intracellular mediator of signalling through the IFN-γR, mice also succumb to infection with L. monocytogenes. However, interpreting these results in the context of CD8+ T-cell function can be difficult, given both the pleiotropic effects of IFN-γ and the involvement of STAT1 in signalling downstream of other IFNs11,12. T-bet and EOMES are the principal transcription factors in the tissue-specific regulation of IFN-γ production by CD8+ T cells (FIG. 1). Other more ubiquitous transcription factors that are involved in the production of IFN-γ are CREB1 (cyclic AMP-responsive-element-binding protein 1), ATF1 (activating transcription factor 1), ATF2, JUN and OCT1 (REF. 13). These factors bind a region of proximal promoter (from –73 to –48 base pairs, bp) in the gene encoding IFN-γ, and when this region is dimerized, gene expression can be induced by phorbol ester and ionomycin and inhibited by cyclosporin A in vitro, indicating a calcineurin-dependent mechanism and therefore the involvement of NFAT (nuclear factor of activated T cells). Additional NFAT-binding sites exist outside this proximal regulatory element, and these can be bound and transactivated by NFAT proteins in vitro14–16. Nuclear factor-κB (NF-κB) proteins can transactivate the gene encoding IFN-γ by binding a region in the first intron17, and inhibition of NF-κB activity by transgenic expression of the inhibitor IκBα (α-subunit of inhibitor of NF-κB) leads to a substantial reduction in the production of IFN-γ 18. Although robust cytokine production is an important effector function of CD8+ T cells, the overriding
VOLUME 4 | NOVEMBER 2004 | 9 0 1 ©2004 Nature Publishing Group
REVIEWS
APC
CD80/ CD86
MHC class II IFN-γ
? TCR
CD28
IFN-γR
CD8+ T cell
Granzyme B NFAT
T-bet ?
Perforin
EOMES IFN-
EGR2, EGR3 FASL Nucleus
Figure 1 | Main transcriptional pathways in CD8+ effector T cells. T-bet expression is induced by signalling through the T-cell receptor (TCR) and the interferon-γ (IFN-γ) receptor (IFN-γR). T-bet then induces the expression of the effector molecules IFN-γ, perforin and granzyme B. Eomesodermin (EOMES) also induces the expression of IFN-γ and perforin. There is a CD28dependent pathway that activates the expression of FAS ligand (FASL) either directly or through the transcription factors NFAT (nuclear factor of activated T cells), EGR2 (early growth response 2) and EGR3. At present, the signals that induce expression of EOMES and the possible interactions between T-bet and EOMES are unclear. APC, antigen-presenting cell.
T-BOX DNA-BINDING DOMAIN
A 200-amino-acid DNAbinding domain found in all members of the T-box family. This domain binds a consensus sequence found in the promoter regions of genes. OT-1 TCR-TRANSGENIC MICE
Transgenic mice that have a T-cell receptor specific for an MHC-class-I-restricted peptide derived from ovalbumin. T cells from these mice can be activated in an antigen-specific manner either in vitro or in vivo. CHROMIUM-RELEASE ASSAY
An assay that determines the activity of cytotoxic cells on the basis of their ability to lyse target cells labelled with radioactive chromium. The amount of radioactivity released is proportional to the number of target cells that are killed by the cytolytic cells added to the culture.
902
function of the activated CD8+ T-cell subset is the lysis of infected cells. NK cells and CD8+ CTLs are the two main cytotoxic subsets of the immune system, and although they respond to different signals, they use similar mechanisms of action19. Transcription factors implicated in CTL function
T-bet. T-bet is a protein of 530 amino acids, which contains a classical T-BOX DNA-BINDING DOMAIN, flanked by two potent transcriptional-activation domains. It was initially cloned in our laboratory from a TH1-cell cDNA library and was shown to potently transactivate the gene encoding IFN-γ5. Analysis of the IFN-γ gene reveals three consensus T-box-binding sites, two in the proximal promoter region and one in the third intron. However, recent work has shown that the predicted consensus T-box sites are dispensable for T-bet-mediated transactivation20. Instead, multiple T-box half-sites within 300 bp of the transcriptional start site can provide T-bet-specific activity. Furthermore, a highly conserved region 5 kb upstream of the transcription-initiation site, known as the IFN-γ 5′-conserved non-coding sequence, has been identified, and it contains several consensus T-box sites21. T-bet and activator protein 1 (AP1) transcription factors interact with the IFN-γ 5′-conserved non-coding
| NOVEMBER 2004 | VOLUME 4
sequence, and inclusion of this enhancer region increases IFN-γ-promoter transactivation. T-bet expression is rapidly induced in CD8+ T cells by signalling through the T-cell receptor (TCR) and the IFN-γR, and T-bet functions downstream of STAT1 (REFS 22,23). In the context of antigen-specific activation, T-bet is required for the differentiation of naive CD8+ T cells into effector CTLs. In the absence of T-bet, CD8+ T cells fail to acquire a normal profile of effector cell-surface markers. Specifically, a failure to downregulate expression of CD62L and upregulate expression of LY6C (lymphocyte antigen 6C) — surface markers that are associated with differentiation and activation — is observed. Interestingly, these changes are masked if the T cells are activated in an antigen non-specific manner24, indicating that the strength of the signal transduced by the TCR might have an important role. Similar to the requirements of CD4+ T cells, T-bet is required for antigen-specific production of IFN-γ by CD8+ T cells. As we have discussed, IFN-γ is the hallmark functional cytokine of an effector CD8+ T cell. Naive CD8+ T cells produce little IFN-γ, but effector CD8+ T cells rapidly secrete large amounts of IFN-γ after activation. This production is important for the clearance of viral infections. CD8+ T cells from T-bet –/– mice crossed with OT-1 TCR-TRANSGENIC MICE produce considerably less IFN-γ than wild-type mice when activated with specific peptide displayed on the surface of an antigen-presenting cell (APC) in the context of MHC molecules. In addition, higher amounts and proportions of cells expressing interleukin-4 (IL-4), IL-5 and IL-10 are observed, indicating that in the absence of T-bet, cytokines that are more characteristic of a TH2-type immune response are produced24. In vitro cytotoxicity is also impaired in the absence of T-bet. CD8+ T cells from T-bet –/– OT-1 TCR+ mice are markedly less efficient at killing peptide-loaded cells than T-bet-sufficient cells, as determined using a 24 CHROMIUM-RELEASE ASSAY . Analogously, in an in vivo crosspriming assay, T-bet –/– OT-1 TCR+ CD8+ T cells showed a marked reduction in specific lysis compared with wild-type controls. However, the cell killing observed in both the in vitro and in vivo assays was not completely eliminated in the absence of T-bet, indicating that other factors might have a role in this setting. These defects in cell killing are independent of the reductions in the amount of IFN-γ produced, because the level of lysis of antigen-loaded targets is similar for CD8+ T cells from wild-type mice and Ifn-γ –/– mice. The physiological significance of these findings is underscored by the fact that T-bet –/– mice are more susceptible to infection with lymphocytic choriomeningitis virus (LCMV), a model in which clearance of virus occurs mainly through CD8+ T-cell-mediated mechanisms. DNA vaccination that induces strong protective CD8+ T-cell responses in wild-type mice fails to protect T-bet –/– mice, with their mortality being similar to naive infected wild-type mice24. This susceptibility seems to be a function of the immunization status of these mice rather than a difference in primary resistance, because viral clearance of LCMV is normal in the absence of T-bet25.
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS
RAT INSULIN PROMOTER (RIP)–LCMV TRANSGENIC MODEL
A transgenic mouse model of type 1 diabetes in which peptides derived from lymphocytic choriomeningitis (LCMV) are expressed in the pancreas under the control of RIP. Infection of the mouse with LCMV leads to the development of diabetes as a result of infiltrating CD8+ effector T cells. CHROMATIN IMMUNOPRECIPITATION
An experimental technique that analyses direct binding of an endogenous transcription factor to chromatin by fixation with formaldehyde followed by immunoprecipitation with a transcription-factor-specific antibody. Gene-specific enrichment is then assessed by polymerase chain reaction analysis of the immunoprecipitated DNA. Vα14i NKT CELLS
The most abundant subset of natural killer T (NKT) cells. They have a rearrangement of the T-cell receptor (TCR) variable-gene segment Vα14 to the joining-region segment Jα18 to form an invariant complementarity-determining region. The resulting TCR is known as Vα14 invariant (Vα14i). This TCR is autoreactive to CD1d, and Vα14i NKT cells respond strongly to α-galactosylceramide (α-GalCer) presented in the context of CD1d.
Furthermore, T-bet has been shown to control CD8+ T-cell responses in the RAT INSULIN PROMOTER (RIP)–LCMV TRANSGENIC MODEL of type 1 diabetes. In the absence of T-bet, the number of LCMV-specific CD8+ effector T cells was reduced, and these cells produced less IFN-γ. In this model, the overall effect of T-bet deficiency was to confer almost complete protection against the development of diabetes, probably through a combination of impaired effector function and reduced IFN-γ production25. EOMES. Recently, it was shown that the T-box gene encoding EOMES is expressed specifically by activated CD8+, and not CD4+, T cells3. An EOMES retroviralexpression construct fused to the potent transcriptional repressor known as engrailed was used to inhibit EOMES activity, resulting in reduced IFN-γ production and cytolysis by CD8+ effector T cells. Furthermore, ectopic expression of EOMES induced expression of the CTL-associated genes encoding perforin and granzyme B by polarized TH2 cells, a cell type that is normally devoid of cytolytic activity3. Inhibition of EOMES results in reduced CTL cytotoxicity and IFN-γ production3. Overexpression of EOMES by T-bet –/– CD8+ T cells can restore IFN-γ production, indicating that the regulation of Ifn-γ transcription might be under the control of both EOMES and T-bet. Analysis of CD8+ T cells from haploinsufficient Eomes+/– mice failed to reveal any defect in IFN-γ production3, which might be explained by the presence of normal levels of T-bet in these cells. It has not been reported whether CD4+ T cells from these mice produce less IFN-γ. Determining the full biological importance of EOMES for T-cell cytokine production and effector function awaits analysis of conditional gene-knockout mice or at least analysis of immune cells from mice that are both haploinsufficient in EOMES and deficient in T-bet. Similarly, the relative importance of T-bet and EOMES remains to be determined. T-box-target genes. Almost all of the known T-boxfamily members function as homodimers. Indeed, the crystal structure of the T-box domain in contact with DNA illustrates the importance of dimer formation 26. Gene-expression profiling of CD8 + T cells from T-bet –/– mice has revealed that the expression of genes that are important for cytotoxicity, such as those encoding perforin and granzyme B, is reduced (B.M.S., unpublished observations). Interestingly, the expression of these genes was also induced (similar to the gene encoding IFN-γ) by overexpression of EOMES. Moreover, their expression can be reduced using a dominant-negative EOMES construct, and perforin expression is reduced in CD8+ T cells from Eomes+/– mice3. Whether these genes are direct targets for T-bet and/or EOMES in CD8+ T cells is unclear, although CHROMATIN-IMMUNOPRECIPITATION experiments carried out using NK cells indicate that this might be the case4. In addition, it will be important to determine what, if any, functional interaction occurs between T-bet and EOMES.
NATURE REVIEWS | IMMUNOLOGY
Non-T-box transcription factors. Other transcription factors have been identified that might have a role in CD8+ T-cell effector function. Dependent on the stimulation conditions, Murphy and Carter observed reduced IFN-γ production by STAT4-deficient CD8+ T cells27. Activation of STAT4-deficient CD8+ T cells with antibodies specific for CD3 and CD28 resulted in normal levels of IFN-γ, whereas stimulation with allogeneic splenocytes resulted in reduced IFN-γ secretion. This study did not assess cytolytic activity, so the role of STAT4 in CTL effector function has not been completely defined. STAT1 is required for IFN-γ-mediated signalling, and a deficiency in this transcription factor leads to increased susceptibility to infection with intracellular pathogens28. Littman and colleagues showed that, in the absence of RUNX3 (runt-related transcription factor 3), CD8+ effector T-cell cytotoxicity was ablated29. On closer examination, RUNX3 was found to drive the proliferation, not the cytolytic mechanisms, of CD8+ T cells. Interestingly, CD4+ T cells do not have the same proliferative requirement for RUNX3 as CD8+ T cells. Similarly, REL-deficient T cells showed marked proliferative defects, and as a result, impaired effector function in terms of both cytokine production and CTL activity. However, the addition of exogenous IL-2 was sufficient to restore the functions affected by the absence of REL30. The transcription factors involved in CTL function are summarized in TABLE 1. α14i NKT cells NK and Vα
NK and NKT cells are two important effector-cell lineages in the immune system. NK cells are crucial cellular mediators of innate defence, and they can rapidly recognize and kill infected or tumorigenic cells, as well as produce effector cytokines and chemokines that modulate other immune cells. These cells have a heterogeneous repertoire of cell-surface activating and inhibitory receptors, and unlike CD8+ T cells, they do not require activation or previous exposure to antigen to mediate cell killing. NKT cells have properties of both NK cells and T cells in that they express a TCR, which is CD1d restricted, but they also express NK-cell activating and inhibitory receptors. In mice, the best-defined population of NKT cells express the Vα14 to Jα18 TCR rearrangement to form an invariant complementaritydetermining region, which forms part of the Vα14 invariant (Vα14i) TCR; these cells are known as Vα14i NKT CELLS. They have cytolytic activity and also rapidly produce large quantities of immunoregulatory cytokines, such as IL-4 and IFN-γ, after TCR stimulation; therefore, they are important for the initiation and regulation of immune responses. So, both NK and Vα14i NKT cells respond rapidly to immune challenge, have cytolytic activity and express cytokines that modulate downstream immune functions. Transcriptional control in NK-cell effector function. The development and function of NK cells have recently been the subject of several excellent reviews31,32. Mature NK cells are activated by stimulation with immunomodulatory cytokines and through many
VOLUME 4 | NOVEMBER 2004 | 9 0 3 ©2004 Nature Publishing Group
REVIEWS
CD16 LY49
NFAT
IL-15R
IL-12R
IL-18R
STAT1 STAT3 STAT5
STAT4
NF-κB
EOMES
T-bet ?
NK-cell development and proliferation
EGR2 EGR3
ETS1 AP1 SP1 NF-κB
FASL
Transcriptionfactor activation and cytotoxicity
SYK ZAP70 (mouse only) JAK2 STAT5 PI3K ERK
NFAT SP1 AP1 ETS1 MEF NF-κB
NFAT Ikaros AP1 MITF
Granzyme B
SYK ZAP70
NKG2D
Perforin
IFN-
Nucleus
Figure 2 | Overview of the transcriptional pathways in natural killer cells. Natural killer (NK)-cell development, growth and activation are mediated by various cell-surface receptors. Signalling through the interleukin-15 (IL-15) receptor (IL-15R) is crucial for NK-cell development and proliferation, whereas signals from the IL-12R and IL-18R synergize to activate transcription of the interferon-γ (IFN-γ ) gene and the genes involved in the granule-exocytosis pathway (the perforin and granzyme B genes). The NK-cell receptors LY49 (lymphocyte antigen 49) and NKG2D (NK group 2, member D) also mediate NK-cell activation, although the signalling pathways downstream of NKG2D vary between mice and humans, as indicated. Ligation of CD16 (the low-affinity Fc receptor for IgG, FcγRIII) results in activation of NFAT (nuclear factor of activated T cells) and expression of FAS ligand (FASL), as well as production of cytokines and degranulation. AP1, activator protein 1; EGR, early growth response 1; EOMES, eomesodermin; ERK, extracellular signal-regulated kinase; JAK2, Janus activated kinase 2; MEF, myeloid ELF1 (E74-like factor 1)-like factor; MITF, microphthalmiaassociated transcription factor; NF-κB, nuclear factor-κB; PI3K, phosphatidylinositol 3-kinase; SP1, SP1 transcription factor; STAT, signal transducer and activator of transcription; SYK, spleen tyrosine kinase; ZAP70, ζ-chain-associated protein kinase of 70 kDa.
types of cell-surface-expressed activating receptor31,33. Importantly, in contrast to naive CD8+ T cells, which require prolonged stimulation through the TCR to initiate the transcription of cell-killing machinery components, mature NK cells contain pre-formed lytic granules that can kill a target cell within 20–30 minutes of NK-cell contact34. This indicates that the cell-killing machinery of NK cells can be activated without de novo transcription. This rapid degranulation of NK cells in response to the presence of target cells is mediated by multiple pathways (FIG. 2). The tyrosine kinases SYK (spleen tyrosine kinase) and ZAP70 (ζ-chain-associated protein kinase of 70 kDa) function downstream of NK-cell activating receptors that contain ITAMs (immunoreceptor tyrosine-based activation motifs) to phosphorylate and mobilize multiple downstream signalling proteins, which results in the activation of NK cells and the initiation of granule exocytosis33. When NK cells bind target cells, phosphatidylinositol 3-kinase (PI3K) is rapidly activated by SYK, and this triggers a signalling cascade that results in the mobilization of lytic granules that move towards the bound target and degranulate to release perforin and granzymes35,36. In addition, PI3K has been shown to be required for the production of IFN-γ by human NK cells following
904
| NOVEMBER 2004 | VOLUME 4
activation through the receptor NKG2D (NK group 2, member D)37. In addition, it has been shown that PI3K expressed by human NK cells is activated downstream of NKG2D in a SYK- and ITAM-independent manner, leading to granule-dependent cytotoxicity38. This rapid activation of the cell-killing machinery of NK cells allows them to immediately respond to challenges by pathogens or tumour cells, without the time-consuming constraint of de novo synthesis of cytotoxic molecules. Clearly, however, the initial generation and the prolonged release of these cytotoxic proteins requires precise transcriptional control. It has recently been suggested that the effector functions of NK cells — namely, the granule-exocytosis pathway, the expression of FAS ligand (FASL; also known as CD95L) and the production of effector cytokines such as IFN-γ — are closely linked to lineage maturation39. For example, NK cells from mice deficient in Ikaros, ETS1, MEF (myeloid ELF1 (E74-like factor 1)-like factor; also known as ELF4), IRF2 (interferon-regulatory factor 2) or T-bet all have decreased NK-cell cytotoxicity, but they also have fewer peripheral NK cells4,40–43. However, it is unclear whether these transcription factors have roles in both NK cytotoxicity and development or whether the decreased effector
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS function observed is solely a result of defective NK-cell maturation. In some cases, however, our knowledge of the target genes of these transcription factors allows us to make this distinction. There are also examples of transcription-factor deficiencies in which NK-cell development is defective but, on a per cell basis, the NK cells show normal cytotoxicity: namely, deficiencies in PU.1 (also known as SPI1), ID2 (inhibitor of DNA binding 2) and GATA3 (GATA-binding protein 3)44–46. Therefore, it remains to be determined precisely how closely NK-cell maturation is linked to effector capability. The ETS-family member MEF has been shown to directly bind and regulate the expression of the perforin gene, and in its absence, both NK-cell cytolytic function and development are severely impaired42. Ikaros and ETS factors regulate the expression of granzyme B (discussed later), and NK-cell development and effector function are markedly impaired in their absence40,41. T-bet has been shown to regulate both NK-cell maturation and cytotoxic function in a stem-cell intrinsic manner4. T-bet-deficient NK cells have decreased terminal maturation and moderately impaired cytotoxicity. Furthermore, the genes encoding perforin and granzyme B were shown to be T-bet target genes, using chromatin immunoprecipitation. Intriguingly, the expression of EOMES — another T-box-family transcription factor that is highly expressed by NK and CD8+ T cells and drives the expression of IFN-γ, perforin and granzymes3 — was shown to be normal in T-bet –/– NK cells. This indicates that there is a compensatory relationship between EOMES and T-bet in the regulation of expression of IFN-γ, perforin and granzymes. This is consistent with the fact that T-bet –/– NK cells have relatively minor cytotoxic defects and normal early production of IFN-γ. The study of EOMES deficiency in combination with T-bet deficiency should help elucidate the roles of these factors in NK-cell development, IFN-γ production, and perforin and granzyme expression. In contrast to the previously mentioned factors, NK cells deficient in the transcription factors MITF (microphthalmia-associated transcription factor), C/EBP-γ (CCAAT/enhancer-binding protein-γ) or NEMO (NF-κB essential modulator; also known as IKK-γ) have normal NK-cell development but decreased cytotoxicity. MITF regulates the production of granzyme B by mast cells, but the basis of the cell-killing defect of MITF-deficient mice was reported to be decreased expression of perforin and not granzyme47. C/EBP-γ-deficient NK cells show decreased cytotoxicity and IFN-γ production, although the identities of the target genes of this factor are not clear48. NEMO-deficient NK cells have defective cytotoxic function owing to impaired activation of NF-κB49, which is consistent with the known roles for the NF-κB pathway in the regulation of perforin and FASL expression. These studies underline the fact that NK-cell development and effector function are closely linked and are controlled by multiple transcriptional pathways. Much work still needs to be done to determine all of the target genes of these factors and to elucidate the network of transcription factors that operate in the NK-cell lineage.
NATURE REVIEWS | IMMUNOLOGY
Transcriptional control in NKT-cell effector function. Vα14i NKT cells restricted by α-galactosylceramide have cytolytic activity and rapidly produce the cytokines IL-4 and IFN-γ after stimulation through their restricted TCR50–52. Recent advances in our understanding of the development of NKT cells have shown that immature Vα14i NKT cells mainly produce IL-4 but switch to the production of IFN-γ after maturation53. Deficiencies in the ETS-family transcription factors ETS1 and MEF have been shown to result in defective development of, and cytokine production by, Vα14i NKT cells42,54. Recently, T-bet was shown to be a crucial transcription factor for the development of this lineage. T-bet deficiency blocked the terminal maturation of Vα14i NKT cells at the Vα14i TCR+CD44hiNK1.1– LY49–CD122– stage, markedly decreased peripheral cell numbers and resulted in a complete absence of production of IFN-γ 4. Therefore, as shown for the CD4+ T-cell lineage55, T-bet is a crucial regulator of IFN-γ production in the Vα14i NKT-cell compartment. It has also been shown that EOMES is not expressed by NKT cells, so EOMES does not contribute to the expression of IFN-γ and cytolytic molecules as is postulated for NK cells. Similar to NK cells, deficiency in either of the transcription factors T-bet or ETS1 results in a block in both development and effector function4,54. It is therefore clear that these factors have multiple downstream target genes and that acquisition of effector function is part of the developmental machinery and is regulated by a complex network of transcription factors. Transcriptional control of NK and CTL cytotoxicity
Lymphocyte-mediated cytotoxicity is a central effector function of both the innate (NK cells) and adaptive (CD8+ T cells) immune systems. The induction of targetcell death can be divided into two general mechanisms: initiation of apoptosis by activation of FADD (FASassociated via death domain) through the ligation of a receptor expressed by the target cell, and the killing of target cells by localized exocytosis of perforin and cytolytic granules containing granzymes19,56. The first mechanism, the FAS–FASL pathway, is the best-characterized ‘deathligand’ system, although other members of the TNF family — such as TRAIL (TNF-related apoptosis-inducing ligand), in its interaction with the TRAIL receptor — have also been shown to use FADD in a similar way57,58. The second mechanism is the use of perforin to potentiate the traffic of granzyme proteins into target cells, where they subsequently initiate cell death through various mechanisms, including activation of caspase-independent mitochondrial and nuclear celldeath pathways, disruption of the plasma membrane and damaging the DNA. Recent studies using gene-targeted mice indicate that the granule-exocytosis pathway is the crucial pathway for mediating cellular cytotoxicity against intracellular pathogens and tumours, although it is clear that FADD-activating receptors and the secretion of cytokines are also important59,60. The detailed mechanisms of action of these pathways have been the subject
VOLUME 4 | NOVEMBER 2004 | 9 0 5 ©2004 Nature Publishing Group
MEF
SP1
SP1
MEF
ETS1
AP1
NF-κB
STAT3, -4, -5A, -5B
NFAT
STAT5
REVIEWS
Perforin –15 kb
–1200 bp
–1000 bp
–700 bp
–500 bp
–100 bp
ETS1
RUNX1 CREB1 NFAT
AP1
RUNX1
Ikaros
AP1
MITF MITF MITF
1.1-kb tissue-specific promoter
Granzyme B –100 bp
–700 bp
+60 bp
MYC
NFAT
EGR2, EGR3
AP1
ETS1 SP1 NFAT
AP1
NF-κB
208-bp T-cell-inducible element
FASL –1200 bp
–900 bp
–300 bp
CD28RE
–100 bp
Figure 3 | General organization of the perforin, granzyme B and FAS ligand gene promoters. A generalized schema of the 5′ promoters of the perforin, granzyme B and FAS ligand (FASL) genes, showing the approximate positions of reported transcription factors bound to their DNA elements. Factors such as T-bet and eomesodermin are also implicated in the regulation of expression of perforin and granzyme B, but their T-box recognition motif(s) in the promoters have not yet been identified. The identity of the factor that binds the CD28 response element (CD28RE) is unknown, and therefore it is depicted differently (triangle). AP1, activator protein 1; bp, base pairs; CREB1, cyclic AMP-responsive-element-binding protein 1; EGR, early growth response; MEF, myeloid ELF1 (E74-like factor 1)-like factor; MITF, microphthalmia-associated transcription factor; NFAT, nuclear factor of activated T cells; NF-κB, nuclear factor-κB; RUNX1, runt-related transcription factor 1; SP1, SP1 transcription factor; STAT, signal transducer and activator of transcription.
of recent comprehensive reviews19,56 and are not reiterated here. Instead, we focus on what is known at present about the transcriptional regulation of these pathways.
DNASE I FOOTPRINTING ANALYSIS
An in vitro experimental technique for identifying the DNA sequence to which a transcription factor binds. A short end-labelled fragment of the DNA sequence of interest is incubated with nuclear extract and then digested with a low concentration of DNase I. The digested DNA is then recovered from the reaction and resolved on a polyacrylamide gel, together with a sequencing reaction using the same DNA fragment as the template. The regions bound by proteins are protected from DNase I digestion and appear as blank areas on the gel, and the exact protein-bound sequence can be determined by comparing the location of the blank areas with that of the sequencing reaction.
906
Transcriptional regulation of perforin gene expression. More than 10 years ago, it was shown that 5 kb of the 5′ proximal region of the mouse perforin gene contains promoter elements that are sufficient to confer tissuespecific expression of perforin61. Whereas a 122-bp proximal core promoter element was constitutively active in all cell lines tested, 5 kb of upstream promoter sequence was shown to be strongly activated in perforin-expressing CTL lines (CTLL-2 cells) but repressed in perforindeficient cells, presumably owing to the presence of upstream positive- and negative-regulatory elements. Using mice transgenic for 5.1 kb of the 5′ flanking and promoter region of the mouse perforin gene, it was confirmed that this region conferred tissue specificity62. Another study indicated that tissue-specific expression of the perforin gene could be conferred by a promoter element encompassing 1.1 kb of upstream sequence63. By contrast, there is evidence that the perforin gene locus is silenced by an intronic element containing a DNase I hypersensitive site64, which is another level of control of perforin gene expression. A motif homologous to the ETS-binding site was identified in a 32-bp element located at –491 to –522 bp of the mouse promoter that enhanced perforin
| NOVEMBER 2004 | VOLUME 4
transcription in cytotoxic cells65. This study also identified two DNA-binding proteins that interact with this element, which were postulated to be members of the ETS family of transcription factors. Subsequently, DNASE I FOOTPRINTING ANALYSIS of the perforin promoter, together with electromobility-shift assays, identified several transcription factors and their binding sites on the perforin promoter, including AP1, SP1 (SP1 transcription factor) and one positive- and two negative-regulatory transcription factors62. The positive-regulatory protein was shown to bind an ETS consensus sequence and was expressed in a killer-cell-specific manner, whereas the two negative-regulatory proteins — one of which was an ETS-family member — were not killer-cell specific. The positive-regulatory ETS-domain-binding protein was postulated to be one of the proteins identified by Koizumi et al. that bound the perforin promoter at a similar location65. So, it seems that ETS-family members have opposing roles as transactivators and repressors of perforin expression. Another study used mutation of binding elements to show a crucial role for SP1 and ETS-family transcription factors in the maintenance of perforin expression by perforin-expressing cell lines, thereby confirming the involvement of these transcription factors in the regulation of the perforin gene66. Recently, the ETS-family transcription factor MEF was shown to directly regulate perforin expression through its binding of two consensus ETS-binding sites in the
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS perforin promoter, one of which corresponds to the previously reported ETS-binding site at approximately –500 bp42,62,65. Therefore, it is probable that one of the killer-cell-specific ETS-family transcription factors is MEF. Indeed, MEF-deficient CD8+ T cells and NK cells were shown to be severely deficient in cytotoxic function42. Recent data have implicated the T-box transcription factors T-bet and EOMES in the regulation of perforin gene expression. T-bet was shown by chromatin immunoprecipitation to directly bind the perforin promoter, and T-bet-deficient NK and CD8+ T cells have decreased levels of perforin expression and cytotoxicity4,24. EOMES — which was cloned on the basis of its expression by activated CD8+ T cells — has been shown to drive the expression of perforin by primary TH2 cells when ectopically expressed by retroviral transduction3. Further work is required to examine whether EOMES directly binds the perforin promoter, to determine the cytotoxic phenotype of EOMES-deficient NK and CD8+ T cells and to determine whether EOMES synergizes with or has a compensatory relationship with T-bet in regulating perforin gene expression. The cytokine-induced expression of perforin by NK and CD8+ T cells has been mapped to an enhancer region in the human perforin promoter that is located between –1136 and –1076 bp, within which a STATbinding site has been identified67 (FIG. 3). STAT proteins are a family of transcription factors that are recruited to the cytoplasmic domains of cytokine receptors when ligand binds; they subsequently undergo phosphorylation, dimerization and translocation to the nucleus, where they mediate the expression of target genes. The STAT-binding site in the human perforin promoter was shown to be bound not only by STAT3 (in the human NK cell line YT) but also by IL-2-induced STAT5A and/or STAT5B and IL-6-induced STAT1 (in primary human NK cells). A further study confirmed the presence of this IL-2-responsive STAT-binding element ∼1 kb upstream in the perforin promoter, and another STAT-binding element was also identified, located ∼15 kb upstream of the promoter68. STAT5-mediated responsiveness to IL-2 stimulation was shown to result from two identically spaced STAT-binding-like elements in these enhancer regions. The most distal enhancer (at ∼15 kb upstream) also responded to signals from the TCR, as shown by its response to phorbol ester and ionomycin in a cyclosporin-sensitive manner, which was probably mediated through an NFAT-like element. The expression of perforin in response to signalling through the TCR is consistent with early reports that show IL-2independent induction of perforin expression in stimulated CD8+ T cells69. The more-proximal STAT-binding element (∼1 kb upstream) has also been shown to bind IL-12-activated STAT4, as well as IL-2-activated STAT5 (REF. 70), providing an explanation for IL-12-mediated activation of perforin gene expression. Interestingly, the proximal enhancer element at approximately –1 kb is activated by NF-κB components that bind a highly conserved NF-κB-binding consensus site in the enhancer71. In this study, NF-κB activation occurred through
NATURE REVIEWS | IMMUNOLOGY
signalling from the IL-2 receptor (IL-2R), showing that there is an IL-2-dependent pathway for activating perforin gene expression that does not involve STAT5. Transcriptional regulation of granzyme gene expression. Granzymes (granule enzymes) are encoded by a family of distinct genes in humans and mice56. Both of these species have distinct chromosomal clusters of granzyme genes, each encoding different homologous duplications. Granzymes A and B are the best-studied proteins in humans and mice, whereas less is known about granzymes C, D, E, F, G (mouse), H (human), K (mouse and human) and M (human). There is relatively little known about the regulation of these genes, although expression studies have indicated that they are differentially regulated during lymphoid ontogeny, and a distinct and restricted repertoire of granzymes is expressed by individual CD8+ T cells72,73. Furthermore, whereas granzymes A and B are the most abundant granzymes in humans and mice and are expressed by both CD8+ T cells and NK cells, mouse granzymes C, D and F and human granzyme H are expressed at a low level by CD8+ T cells and are preferentially expressed by the NK-cell compartment56,74,75. Granzyme B is not expressed by resting T cells, but its expression is induced following stimulation76. This inducibility was mapped to a 200-bp element of the granzyme B proximal promoter located at –148 to +60 bp, which contains a binding site for NFAT77 (FIG. 3). This element was also shown to contain binding sites for the transcription factors Ikaros, RUNX1 (also known as CBF), ETS, CREB1 and AP1 (REFS 78–81). Mutational analysis showed that these binding sites were functional and contributed to granzyme B expression. It was also shown that interference with the binding of Ikaros and AP1 to this element, using treatment with dexamethasone, decreased the expression of granzyme B by phytohaemagglutinin-activated peripheral-blood mononuclear cells82. Importantly, a strong DNase I hypersensitivity site was present in this proximal promoter element in activated CD8+ T cells, but it was absent in resting T cells and non-cytolytic cell lines, thereby providing an explanation at the level of chromatin structure for the tissue-specific and activation/differentiation-state-specific expression of granzyme B79. The presence or absence of DNase I hypersensitivity sites at a particular location in a DNA sequence is indicative of open or closed chromatin conformations, respectively. An open chromatin conformation allows access to the DNA and binding of transcription factors, thereby facilitating gene expression. It has also been shown that MITF (a basic helixloop-helix leucine-zipper transcription factor) binds three sites in the mouse granzyme B promoter between –574 and –515 bp83. However, studies using mice that have mutant Mitf (Mi) alleles yield conflicting results: Mi/Mi mast cells have decreased expression of granzyme B, but expression of granzyme B by Mi/Mi cells was normal47,84. So, the requirement for MITF in the regulation of granzyme B expression varies depending on the cell type.
VOLUME 4 | NOVEMBER 2004 | 9 0 7 ©2004 Nature Publishing Group
REVIEWS
SV40 TAG REPORTER
A reporter gene consisting of the SV40 virus large T antigen — a multifunctional 85 kDa protein that is the sole viral protein required for SV40 replication and causes malignant transformation of susceptible cells.
Similar to perforin, the T-box transcription factors T-bet and EOMES have recently been implicated in the regulation of granzyme B expression. T-bet was shown to bind the granzyme B promoter (using chromatin immunoprecipitation), and T-bet –/– NK cells were shown to have decreased levels of granzyme B expression and cytotoxicity4. Retroviral overexpression of EOMES by TH2 cells results in the induction of granzyme B expression, although it has not yet been shown that EOMES binds directly to the granzyme B promoter3. The regulation of human granzyme H, which is expressed by NK cells but not T cells, has received some attention. Using transgenic mice, it was shown that 1.2 kb of proximal human granzyme H promoter driving an SV40 TAG REPORTER was sufficient to confer restricted expression of the SV40 antigen to the NK-cell compartment74. Presumably, this 1.2 kb of promoter contains the cis regulatory sequences that are required for NK-cell-specific expression, although more work is required to define these sequences and the transcription factors that bind them. Therefore, despite the fact that the granzyme B and granzyme H genes are tightly linked (because they are located close to one another on human chromosome 14q11.2), their expression by CD8+ T cells is distinct. The expression of granzymes has been shown to be induced and increased by cytokines such as IL-2, IL-12 and IL-15, drawing parallels with the activation of perforin by these cytokines. IL-2 by itself activates the expression of granzymes A and B, and synergizes with IL-12 to further induce the expression of these genes85,86. Stimulation of splenic T-cell-enriched lymphocytes with IL-15 induces the expression of granzymes A and B, together with perforin and FASL87. However, the precise signal-transduction mechanisms that take place — from the cytokine receptors to the initiation of granzyme gene expression — remain to be thoroughly explored. Regulation of FASL gene expression. Expression of the TNF-family member FASL by NK and T cells is an alternative cytotoxic mechanism distinct from the granuleexocytosis pathway, which triggers apoptotic cell death of target cells expressing the cognate receptor FAS. Other members of the TNF family, such as TRAIL, are used in a similar way by these cells. The mechanisms of action of these TNF-family members have been the subject of comprehensive reviews88,89. Immature human NK cells have been shown to mediate cytotoxicity through the expression of TRAIL, whereas mature NK cells acquire expression of FASL, together with functional granule-exocytosis machinery90,91. In the T-cell compartment, FASL is induced following activation of CD4+ and CD8+ T cells through the TCR92,93, and it is crucial for the effector function of both of these cell types. Furthermore, secondary stimulation of mouse CD4+ T cells results in increased expression of FASL compared with primary stimulation94. The transcriptional regulation of FASL has received considerable attention in the past few years. It has been
908
| NOVEMBER 2004 | VOLUME 4
shown that 1.2 kb of proximal FASL promoter can be bound and transactivated by NF-κB and AP1 (REFS 95,96), and NF-κB could be a factor downstream of TCR activation that regulates FASL expression97. Indeed, perturbation of the NF-κB-signalling pathway results in decreased FASL expression by T-cell hybridomas98. Members of the early growth response (EGR) family of transcription factors have been shown to be important regulators of the FASL gene. Overexpression of EGR2 or EGR3 was shown to be sufficient to drive a reporter under the control of the FASL promoter and to increase endogenous expression of mRNA encoding FASL. It was also shown that these factors bind a regulatory element located at –214 to –207 bp in the proximal promoter99,100. More recently, these transcription factors have also been implicated in the regulation of TNF and TRAIL expression, at least in non-lymphoid cells101. It has been suggested that the NFAT family of transcription factors might directly regulate FASL transcription. When FASL expression was induced by activation of calcineurin (a phosphatase involved in the nuclear translocation of NFAT), through TCR signalling in a cyclosporin-A-sensitive manner, two NFAT-binding sites (one proximal and one distal) were identified in the –300-bp region of the proximal promoter, using DNase I footprinting; moreover, NFAT proteins were shown to bind and transactivate the FASL promoter in Jurkat T cells93,102. Also, mice deficient in NFATP (also known as NFATC2) and/or NFAT4 (also known as NFATC2) have impaired expression of FASL103,104. However, it was subsequently shown that EGR2 and EGR3 are NFAT target genes and are also controlled by NFAT-interacting protein 45 (NIP45; also known as NFATC2IP)105. Therefore, although NFAT proteins might directly bind and regulate the FASL promoter, the induction of EGR2 and EGR3 by NFATs is also crucial for optimal FASL gene expression. Indeed, the transcription factor MHC class II transactivator (CIITA; also known as MHC2TA) has been shown to repress FASL expression through its competitive interaction with the NFATs106,107. FASL expression has been shown to result from a calcineurin–NFAT pathway triggered by ligation of CD16 (the low-affinity Fc receptor for IgG, FcγRIII) at the surface of human NK cells108,109. Therefore, antibody-mediated activation of NK cells is another mechanism through which FASL expression is indicated by the activation of NFATs. Although the NFATs are involved in the main pathway downstream of TCR signalling that regulates FASL expression, a recent report has also identified a CD28 response element (CD28RE) located at –210 to –201 bp of the FASL promoter that increases transcription of FASL following ligation of CD28 (REF. 110). This report also identified an AP1-binding transcriptional repressor element located 60 bp upstream of CD28RE. The existence of this response element is consistent with an earlier report showing that maximal induction of a FASL promoter–reporter construct in transgenic mice was dependent on both prolonged TCR stimulation and CD28 ligation 111. The distal
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS
Table 2 | Lessons from gene-knockout mice and disease models Gene
Phenotype of knockout mouse
Perforin
Increased susceptibility to lung tumour metastases and tumour surveillance
References 119,120
Granzyme A
Increased susceptibility to ectromelia poxvirus infection, but normal control of tumour rejection
121,122
Granzyme B
Defective target-cell DNA fragmentation, but normal control of tumour rejection
122,123
FAS ligand
Development of lymphoproliferative disease Increased susceptibility to tumour metastases in the lungs
Ifn-γ
Increased susceptibility to Mycobacterium bovis infection Increased susceptibility to Mycobacterium tuberculosis infection Iincreased susceptibility to experimental allergic encephalomyelitis
124 125 126 127,128 129,130
Ifn-γ r
Increased susceptibility to Listeria monocytogenes and vaccinia virus infections
Stat1
Increased susceptibility to Listeria monocytogenes and vesicular stomatitis virus infections Increased susceptibility to experimental allergic encephalomyelitis
Stat4
Increased susceptibility to Leishmania major infection
Eomes
Heterozygote not analysed in disease model; Eomes knockout is embryonic lethal
T-bet
Increased susceptibility to Leishmania major infection Increased susceptibility to lymphocytic choriomeningitis virus infection Resistant to insulin-dependent diabetes Resistant to inflammatory bowel disease Resistant to experimental allergic encephalomyelitis
9,10 28 131,132 133 3 55 25 25 134 131
Eomes, eomesodermin; Ifn-γ, interferon-γ; Ifn-γ r, IFN-γ receptor; Stat, signal transducer and activator of transcription.
NFAT-binding site in the FASL proximal promoter was shown to bind both SP1 and NFAT proteins, and FASL expression could be induced by signalling downstream of the IL-2R112. It has also been reported that transactivation of this element by SP1 is responsible for the constitutive expression of FASL by Sertoli cells and that ETS1 molecules binding near this NFAT/SP1-recognition element cooperate with SP1 in driving the transcription of FASL in vascular smooth muscle cells113,114. Recent studies have shown that MYC and the cyclin B1–cyclin-dependent kinase 1 (CDK1) complex — both regulators of the cell cycle — regulate the induction of FASL expression by activated T cells115–117. MYC directly binds a putative element located at –120 bp in the FASL promoter and drives expression, whereas the cyclin B1–CDK1 complex potentiates FASL expression through its control of NF-κB activation. Therefore, transcriptional control of FASL expression is also regulated at the level of cell-cycle progression. Concluding remarks
The considerable body of work carried out in the past few years shows that there are many transcription factors involved in the regulation of cytotoxic and cytokine genes in cytolytic effector cells. However, many of these genes — for example, the NF-κB-family members — are expressed by many cell lineages, so this cannot explain the tissue specificity of the expression of genes such as those encoding perforin, the granzymes and IFN-γ by the effector cells of the immune system. Understanding the molecular basis of this tissue specificity and how cell-lineage decisions are made is therefore an ongoing challenge in this field. Recent progress has been made in the identification of transcription factors that are responsible for the development of the cytolytic effector-cell lineages and for
NATURE REVIEWS | IMMUNOLOGY
the tightly controlled expression of cytotoxic molecules and immunoregulatory cytokines. In particular, insights into the transcriptional control of cytolytic immune effector cells have been gained by the description of the roles of the T-box-family members T-bet and EOMES in NK and CD8+ T cells. These studies reveal that the transcriptional pathways of these cell types are similar, and both factors regulate effector cytokine production and cytotoxicity with a degree of redundancy and overlap. In NK cells, the cytolytic molecules perforin and the granzymes are direct transcriptional targets of T-bet; in CD8+ T cells, the expression of these genes is also controlled by T-bet, as well as by EOMES. Identification of more novel tissuespecific transcription factors and further investigation of the chromatin structure, promoter elements and locus control regions of cytolytic and effector cytokine genes will allow a deeper understanding of the tightly regulated transcriptional control of cytolytic effector function. Recently, an intriguing study has indicated that NK cells, in addition to being immediately capable of directly lysing infected target cells, can also activate cytolytic CD8+ T cells and TH1 cells by providing help to the dendritic cells that are instructing the development of these effectors118. Therefore, through their influence on dendritic-cell function, NK cells could have a role as an activating bridge between the rapidly responding but transient innate immune system and the slower responding but long-term adaptive immune system. This putative close interaction of NK and CD8 + T cells is reflected by the fact that the transcriptional mechanisms by which these different cell types achieve their effector function seem to be very similar. Detailed knowledge of these transcriptional pathways might provide attractive targets (TABLE 2) for the treatment of autoimmune, infectious and neoplastic diseases.
VOLUME 4 | NOVEMBER 2004 | 9 0 9 ©2004 Nature Publishing Group
REVIEWS 1.
2. 3.
4.
5. 6.
7.
8.
9. 10.
11.
12. 13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
910
Kadonaga, J. T. Regulation of RNA polymerase II transcription by sequence-specific DNA binding factors. Cell 116, 247–257 (2004). This review provides a general overview of the control of gene expression by transcription factors. Latchman, D. S. Eukaryotic Transcription Factors (Academic Press, Boston, 1998). Pearce, E. L. et al. Control of effector CD8+ T cell function by the transcription factor Eomesodermin. Science 302, 1041–1043 (2003). This paper describes the identification of EOMES as an important transcription factor in CD8+ T cells, driving effector function and IFN-γ production. Townsend, M. J. et al. T-bet regulates the terminal maturation and homeostasis of NK and Vα14i NKT Cells. Immunity 20, 477–494 (2004). This paper describes NK- and Vα14i NKT-cell defects in T-bet –/– mice. The NK-cell defect was partial and mostly developmental, whereas the Vα14i NKT-cell defect was more profound. This difference might be explained by the expression of EOMES by NK cells but not Vα14i NKT cells. Szabo, S. J. et al. A novel transcription factor, T-bet, directs TH1 lineage commitment. Cell 100, 655–669 (2000). Szabo, S. J., Sullivan, B. M., Peng, S. L. & Glimcher, L. H. Molecular mechanisms regulating TH1 immune responses. Annu. Rev. Immunol. 21, 713–758 (2003). Kaech, S. M., Hemby, S., Kersh, E. & Ahmed, R. Molecular and functional profiling of memory CD8 T cell differentiation. Cell 111, 837–851 (2002). This paper provides a comprehensive analysis of the changes in gene expression in CD8+ T cells during differentiation and memory-cell formation. Harty, J. T., Tvinnereim, A. R. & White, D. W. CD8+ T cell effector mechanisms in resistance to infection. Annu. Rev. Immunol. 18, 275–308 (2000). Huang, S. et al. Immune response in mice that lack the interferon-γ receptor. Science 259, 1742–1745 (1993). Lu, B. et al. Targeted disruption of the interferon-γ receptor 2 gene results in severe immune defects in mice. Proc. Natl Acad. Sci. USA 95, 8233–8238 (1998). Boehm, U., Klamp, T., Groot, M. & Howard, J. C. Cellular responses to interferon-γ. Annu. Rev. Immunol. 15, 749–795 (1997). Leonard, W. J. & O’Shea, J. J. Jaks and STATs: biological implications. Annu. Rev. Immunol. 16, 293–322 (1998). Penix, L. A. et al. The proximal regulatory element of the interferon-γ promoter mediates selective expression in T cells. J. Biol. Chem. 271, 31964–31972 (1996). Aune, T. M., Penix, L. A., Rincon, M. R. & Flavell, R. A. Differential transcription directed by discrete γ interferon promoter elements in naive and memory (effector) CD4 T cells and CD8 T cells. Mol. Cell. Biol. 17, 199–208 (1997). Campbell, P. M., Pimm, J., Ramassar, V. & Halloran, P. F. Identification of a calcium-inducible, cyclosporine sensitive element in the IFN-γ promoter that is a potential NFAT binding site. Transplantation 61, 933–939 (1996). Sweetser, M. T. et al. The roles of nuclear factor of activated T cells and ying-yang 1 in activation-induced expression of the interferon-γ promoter in T cells. J. Biol. Chem. 273, 34775–34783 (1998). Sica, A. et al. Interaction of NF-κB and NFAT with the interferon-γ promoter. J. Biol. Chem. 272, 30412–30420 (1997). Aronica, M. A. et al. Preferential role for NF-κB/Rel signaling in the type 1 but not type 2 T cell-dependent immune response in vivo. J. Immunol. 163, 5116–5124 (1999). Russell, J. H. & Ley, T. J. Lymphocyte-mediated cytotoxicity. Annu. Rev. Immunol. 20, 323–370 (2002). This is an excellent review that provides an overview of all known cytotoxic mechanisms. Cho, J. Y., Grigura, V., Murphy, T. L. & Murphy, K. Identification of cooperative monomeric Brachyury sites conferring T-bet responsiveness to the proximal IFN-γ promoter. Int. Immunol. 15, 1149–1160 (2003). Lee, D. U., Avni, O., Chen, L. & Rao, A. A distal enhancer in the interferon-γ (IFN-γ ) locus revealed by genome sequence comparison. J. Biol. Chem. 279, 4802–4810 (2004). Lighvani, A. A. et al. T-bet is rapidly induced by interferon-γ in lymphoid and myeloid cells. Proc. Natl Acad. Sci. USA 98, 15137–15142 (2001). Afkarian, M. et al. T-bet is a STAT1-induced regulator of IL-12R expression in naive CD4+ T cells. Nature Immunol. 3, 549–557 (2002). Sullivan, B. M., Juedes, A., Szabo, S. J., von Herrath, M. & Glimcher, L. H. Antigen-driven effector CD8 T cell function regulated by T-bet. Proc. Natl Acad. Sci. USA 100, 15818–15823 (2003).
25.
26.
27.
28.
29.
30.
31.
32.
33. 34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
This paper was the first to describe a defect in CD8+ T-cell effector function, together with impaired production of IFN-γ, in the absence of T-bet. Increased mortality after infection with LCMV was also observed in T-bet –/– mice. Juedes, A. E., Rodrigo, E., Togher, L., Glimcher, L. H. & von Herrath, M. G. T-bet controls autoaggressive CD8 lymphocyte responses in type 1 diabetes. J. Exp. Med. 199, 1153–1162 (2004). Muller, C. W. & Herrmann, B. G. Crystallographic structure of the T domain–DNA complex of the Brachyury transcription factor. Nature 389, 884–888 (1997). Carter, L. L. & Murphy, K. M. Lineage-specific requirement for signal transducer and activator of transcription (Stat)4 in interferon γ production from CD4+ versus CD8+ T cells. J. Exp. Med. 189, 1355–1360 (1999). Meraz, M. A. et al. Targeted disruption of the Stat1 gene in mice reveals unexpected physiologic specificity in the JAK–STAT signaling pathway. Cell 84, 431–442 (1996). Taniuchi, I. et al. Differential requirements for Runx proteins in CD4 repression and epigenetic silencing during T lymphocyte development. Cell 111, 621–633 (2002). Liou, H. C. et al. c-Rel is crucial for lymphocyte proliferation but dispensable for T cell effector function. Int. Immunol. 11, 361–371 (1999). Colucci, F., Caligiuri, M. A. & Di Santo, J. P. What does it take to make a natural killer? Nature Rev. Immunol. 3, 413–425 (2003). Biron, C. A., Nguyen, K. B., Pien, G. C., Cousens, L. P. & Salazar-Mather, T. P. Natural killer cells in antiviral defense: function and regulation by innate cytokines. Annu. Rev. Immunol. 17, 189–220 (1999). Lanier, L. L. Natural killer cell receptor signaling. Curr. Opin. Immunol. 15, 308–314 (2003). Eriksson, M. et al. Inhibitory receptors alter natural killer cell interactions with target cells yet allow simultaneous killing of susceptible targets. J. Exp. Med. 190, 1005–1012 (1999). Jiang, K. et al. Pivotal role of phosphoinositide-3 kinase in regulation of cytotoxicity in natural killer cells. Nature Immunol. 1, 419–425 (2000). Jiang, K. et al. Syk regulation of phosphoinositide 3-kinasedependent NK cell function. J. Immunol. 168, 3155–3164 (2002). Sutherland, C. L. et al. UL16-binding proteins, novel MHC class I-related proteins, bind to NKG2D and activate multiple signaling pathways in primary NK cells. J. Immunol. 168, 671–679 (2002). Billadeau, D. D., Upshaw, J. L., Schoon, R. A., Dick, C. J. & Leibson, P. J. NKG2D–DAP10 triggers human NK cellmediated killing via a Syk-independent regulatory pathway. Nature Immunol. 4, 557–564 (2003). Kim, S. et al. In vivo developmental stages in murine natural killer cell maturation. Nature Immunol. 3, 523–528 (2002). Wang, J. H. et al. Selective defects in the development of the fetal and adult lymphoid system in mice with an Ikaros null mutation. Immunity 5, 537–549 (1996). Barton, K. et al. The Ets-1 transcription factor is required for the development of natural killer cells in mice. Immunity 9, 555–563 (1998). Lacorazza, H. D. et al. The ETS protein MEF plays a critical role in perforin gene expression and the development of natural killer and NK-T cells. Immunity 17, 437–449 (2002). This paper shows that the ETS-family transcription factor MEF directly regulates the perforin gene, in addition to being an important factor for NK- and NKT-cell development. Lohoff, M. et al. Deficiency in the transcription factor interferon regulatory factor (IRF)-2 leads to severely compromised development of natural killer and T helper type 1 cells. J. Exp. Med. 192, 325–336 (2000). Colucci, F. et al. Differential requirement for the transcription factor PU.1 in the generation of natural killer cells versus B and T cells. Blood 97, 2625–2632 (2001). Samson, S. I. et al. GATA-3 promotes maturation, IFN-γ production, and liver-specific homing of NK Cells. Immunity 19, 701–711 (2003). Yokota, Y. et al. Development of peripheral lymphoid organs and natural killer cells depends on the helix-loop-helix inhibitor Id2. Nature 397, 702–706 (1999). Ito, A. et al. Inhibitory effect on natural killer activity of microphthalmia transcription factor encoded by the mutant mi allele of mice. Blood 97, 2075–2083 (2001). Kaisho, T. et al. Impairment of natural killer cytotoxic activity and interferon γ production in CCAAT/enhancer binding protein γ-deficient mice. J. Exp. Med. 190, 1573–1582 (1999). Orange, J. S. et al. Deficient natural killer cell cytotoxicity in patients with IKK-γ/NEMO mutations. J. Clin. Invest. 109, 1501–1509 (2002).
| NOVEMBER 2004 | VOLUME 4
50. Metelitsa, L. S. et al. Human NKT cells mediate antitumor cytotoxicity directly by recognizing target cell CD1d with bound ligand or indirectly by producing IL-2 to activate NK cells. J. Immunol. 167, 3114–3122 (2001). 51. Cui, J. et al. Requirement for Vα14 NKT cells in IL-12mediated rejection of tumors. Science 278, 1623–1626 (1997). 52. Kronenberg, M. & Gapin, L. The unconventional lifestyle of NKT cells. Nature Rev. Immunol. 2, 557–568 (2002). 53. Benlagha, K., Kyin, T., Beavis, A., Teyton, L. & Bendelac, A. A thymic precursor to the NK T cell lineage. Science 296, 553–555 (2002). 54. Walunas, T. L., Wang, B., Wang, C. R. & Leiden, J. M. The Ets1 transcription factor is required for the development of NK T cells in mice. J. Immunol. 164, 2857–2860 (2000). 55. Szabo, S. J. et al. Distinct effects of T-bet in TH1 lineage commitment and IFN-γ production in CD4 and CD8 T cells. Science 295, 338–342 (2002). 56. Lieberman, J. The ABCs of granule-mediated cytotoxicity: new weapons in the arsenal. Nature Rev. Immunol. 3, 361–370 (2003). This is a comprehensive review of cell-killing mechanisms, with a focus on granzyme- and perforindependent pathways. 57. Kischkel, F. C. et al. Apo2L/TRAIL-dependent recruitment of endogenous FADD and caspase-8 to death receptors 4 and 5. Immunity 12, 611–620 (2000). 58. Sprick, M. R. et al. FADD/MORT1 and caspase-8 are recruited to TRAIL receptors 1 and 2 and are essential for apoptosis mediated by TRAIL receptor 2. Immunity 12, 599–609 (2000). 59. Balkow, S. et al. Concerted action of the FasL/Fas and perforin/granzyme A and B pathways is mandatory for the development of early viral hepatitis but not for recovery from viral infection. J. Virol. 75, 8781–8791 (2001). 60. Muller, U. et al. Concerted action of perforin and granzymes is critical for the elimination of Trypanosoma cruzi from mouse tissues, but prevention of early host death is in addition dependent on the FasL/Fas pathway. Eur. J. Immunol. 33, 70–78 (2003). References 59 and 60 describe studies that used gene-deficient mice to determine the relative contributions to cytotoxicity of the FAS–FASL pathway and the granule-exocytosis pathway. 61. Lichtenheld, M. G. & Podack, E. R. Structure and function of the murine perforin promoter and upstream region. Reciprocal gene activation or silencing in perforin positive and negative cells. J. Immunol. 149, 2619–2626 (1992). 62. Zhang, Y. & Lichtenheld, M. G. Non-killer cell-specific transcription factors silence the perforin promoter. J. Immunol. 158, 1734–1741 (1997). 63. Smyth, M. J., Kershaw, M. H., Hulett, M. D., McKenzie, I. F. & Trapani, J. A. Use of the 5′-flanking region of the mouse perforin gene to express human Fcγ receptor I in cytotoxic T lymphocytes. J. Leukoc. Biol. 55, 514–522 (1994). 64. Youn, B. S., Lim, C. L., Shin, M. K., Hill, J. M. & Kwon, B. S. An intronic silencer of the mouse perforin gene. Mol. Cells 13, 61–68 (2002). 65. Koizumi, H. et al. Identification of a killer cell-specific regulatory element of the mouse perforin gene: an Etsbinding site-homologous motif that interacts with Ets-related proteins. Mol. Cell. Biol. 13, 6690–6701 (1993). 66. Youn, B. S., Kim, K. K. & Kwon, B. S. A critical role of Sp1and Ets-related transcription factors in maintaining CTLspecific expression of the mouse perforin gene. J. Immunol. 157, 3499–3509 (1996). 67. Yu, C. R. et al. Role of a STAT binding site in the regulation of the human perforin promoter. J. Immunol. 162, 2785–2790 (1999). 68. Zhang, J., Scordi, I., Smyth, M. J. & Lichtenheld, M. G. Interleukin 2 receptor signaling regulates the perforin gene through signal transducer and activator of transcription (Stat)5 activation of two enhancers. J. Exp. Med. 190, 1297–1308 (1999). 69. Lu, P., Garcia-Sanz, J. A., Lichtenheld, M. G. & Podack, E. R. Perforin expression in human peripheral blood mononuclear cells. Definition of an IL-2-independent pathway of perforin induction in CD8+ T cells. J. Immunol. 148, 3354–3360 (1992). 70. Yamamoto, K., Shibata, F., Miyasaka, N. & Miura, O. The human perforin gene is a direct target of STAT4 activated by IL-12 in NK cells. Biochem. Biophys. Res. Commun. 297, 1245–1252 (2002). 71. Zhou, J., Zhang, J., Lichtenheld, M. G. & Meadows, G. G. A role for NF-κB activation in perforin expression of NK cells upon IL-2 receptor signaling. J. Immunol. 169, 1319–1325 (2002). 72. Kelso, A. et al. The genes for perforin, granzymes A–C and IFN-γ are differentially expressed in single CD8+ T cells during primary activation. Int. Immunol. 14, 605–613 (2002).
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
REVIEWS 73. Grossman, W. J. et al. Differential expression of granzymes A and B in human cytotoxic lymphocyte subsets and T regulatory cells. Blood 6 July 2004 (doi:10.1182/blood-2004-03-0859). 74. MacIvor, D. M., Pham, C. T. & Ley, T. J. The 5′ flanking region of the human granzyme H gene directs expression to T/natural killer cell progenitors and lymphokineactivated killer cells in transgenic mice. Blood 93, 963–973 (1999). 75. Pham, C. T., MacIvor, D. M., Hug, B. A., Heusel, J. W. & Ley, T. J. Long-range disruption of gene expression by a selectable marker cassette. Proc. Natl Acad. Sci. USA 93, 13090–13095 (1996). 76. Heusel, J. W., Hanson, R. D., Silverman, G. A. & Ley, T. J. Structure and expression of a cluster of human hematopoietic serine protease genes found on chromosome 14q11.2. J. Biol. Chem. 266, 6152–6158 (1991). 77. Haddad, P., Wargnier, A., Bourge, J. F., Sasportes, M. & Paul, P. A promoter element of the human serine esterase granzyme B gene controls specific transcription in activated T cells. Eur. J. Immunol. 23, 625–629 (1993). 78. Wargnier, A. et al. Identification of human granzyme B promoter regulatory elements interacting with activated T-cell-specific proteins: implication of Ikaros and CBF binding sites in promoter activation. Proc. Natl Acad. Sci. USA 92, 6930–6934 (1995). 79. Babichuk, C. K., Duggan, B. L. & Bleackley, R. C. In vivo regulation of murine granzyme B gene transcription in activated primary T cells. J. Biol. Chem. 271, 16485–16493 (1996). 80. Babichuk, C. K. & Bleackley, R. C. Mutational analysis of the murine granzyme B gene promoter in primary T cells and a T cell clone. J. Biol. Chem. 272, 18564–18571 (1997). 81. Hanson, R. D., Grisolano, J. L. & Ley, T. J. Consensus AP-1 and CRE motifs upstream from the human cytotoxic serine protease B (CSP-B/CGL-1) gene synergize to activate transcription. Blood 82, 2749–2757 (1993). 82. Wargnier, A. et al. Down-regulation of human granzyme B expression by glucocorticoids. Dexamethasone inhibits binding to the Ikaros and AP-1 regulatory elements of the granzyme B promoter. J. Biol. Chem. 273, 35326–35331 (1998). 83. Ito, A. et al. Systematic method to obtain novel genes that are regulated by mi transcription factor: impaired expression of granzyme B and tryptophan hydroxylase in mi/mi cultured mast cells. Blood 91, 3210–3221 (1998). 84. Kim, D. K. et al. Different effect of various mutant MITF encoded by mi, Mior, or Miwh allele on phenotype of murine mast cells. Blood 93, 4179–4186 (1999). 85. Manyak, C. L. et al. IL-2 induces expression of serine protease enzymes and genes in natural killer and nonspecific T killer cells. J. Immunol. 142, 3707–3713 (1989). 86. DeBlaker-Hohe, D. F., Yamauchi, A., Yu, C. R., HorvathArcidiacono, J. A. & Bloom, E. T. IL-12 synergizes with IL-2 to induce lymphokine-activated cytotoxicity and perforin and granzyme gene expression in fresh human NK cells. Cell. Immunol. 165, 33–43 (1995). 87. Ye, W., Young, J. D. & Liu, C. C. Interleukin-15 induces the expression of mRNAs of cytolytic mediators and augments cytotoxic activities in primary murine lymphocytes. Cell. Immunol. 174, 54–62 (1996). 88. Wallach, D. et al. Tumor necrosis factor receptor and Fas signaling mechanisms. Annu. Rev. Immunol. 17, 331–367 (1999). 89. Nagata, S. Fas ligand-induced apoptosis. Annu. Rev. Genet. 33, 29–55 (1999). 90. Bennett, I. M. et al. Definition of a natural killer NKR-P1A+/CD56–/CD16– functionally immature human NK cell subset that differentiates in vitro in the presence of interleukin 12. J. Exp. Med. 184, 1845–1856 (1996). 91. Zamai, L. et al. Natural killer (NK) cell-mediated cytotoxicity: differential use of TRAIL and Fas ligand by immature and mature primary human NK cells. J. Exp. Med. 188, 2375–2380 (1998). 92. Suda, T. et al. Expression of the Fas ligand in cells of T cell lineage. J. Immunol. 154, 3806–3813 (1995). 93. Latinis, K. M. et al. Regulation of CD95 (Fas) ligand expression by TCR-mediated signaling events. J. Immunol. 158, 4602–4611 (1997). 94. Wang, J. K., Zhu, B., Ju, S. T., Tschopp, J. & Marshak-Rothstein, A. CD4+ T cells reactivated with superantigen are both more sensitive to FasL-mediated killing and express a higher level of FasL. Cell. Immunol. 179, 153–164 (1997).
95. Matsui, K., Fine, A., Zhu, B., Marshak-Rothstein, A. & Ju, S. T. Identification of two NF-κB sites in mouse CD95 ligand (Fas ligand) promoter: functional analysis in T cell hybridoma. J. Immunol. 161, 3469–3473 (1998). 96. Kasibhatla, S. et al. DNA damaging agents induce expression of Fas ligand and subsequent apoptosis in T lymphocytes via the activation of NF-κB and AP-1. Mol. Cell 1, 543–551 (1998). 97. Kasibhatla, S., Genestier, L. & Green, D. R. Regulation of Fas-ligand expression during activation-induced cell death in T lymphocytes via nuclear factor κB. J. Biol. Chem. 274, 987–992 (1999). 98. Matsui, K. et al. Proteasome regulation of Fas ligand cytotoxicity. Eur. J. Immunol. 27, 2269–2278 (1997). 99. Mittelstadt, P. R. & Ashwell, J. D. Cyclosporin A-sensitive transcription factor Egr-3 regulates Fas ligand expression. Mol. Cell. Biol. 18, 3744–3751 (1998). 100. Mittelstadt, P. R. & Ashwell, J. D. Role of Egr-2 in upregulation of Fas ligand in normal T cells and aberrant double-negative lpr and gld T cells. J. Biol. Chem. 274, 3222–3227 (1999). 101. Droin, N. M., Pinkoski, M. J., Dejardin, E. & Green, D. R. Egr family members regulate nonlymphoid expression of Fas ligand, TRAIL, and tumor necrosis factor during immune responses. Mol. Cell. Biol. 23, 7638–7647 (2003). 102. Latinis, K. M., Norian, L. A., Eliason, S. L. & Koretzky, G. A. Two NFAT transcription factor binding sites participate in the regulation of CD95 (Fas) ligand expression in activated human T cells. J. Biol. Chem. 272, 31427–31434 (1997). 103. Hodge, M. R. et al. Hyperproliferation and dysregulation of IL-4 expression in NF-ATp-deficient mice. Immunity 4, 397–405 (1996). 104. Ranger, A. M., Oukka, M., Rengarajan, J. & Glimcher, L. H. Inhibitory function of two NFAT family members in lymphoid homeostasis and TH2 development. Immunity 9, 627–635 (1998). 105. Rengarajan, J. et al. Sequential involvement of NFAT and Egr transcription factors in FasL regulation. Immunity 12, 293–300 (2000). 106. Gourley, T. S. & Chang, C. H. The class II transactivator prevents activation-induced cell death by inhibiting Fas ligand gene expression. J. Immunol. 166, 2917–2921 (2001). 107. Gourley, T. S., Patel, D. R., Nickerson, K., Hong, S. C. & Chang, C. H. Aberrant expression of Fas ligand in mice deficient for the MHC class II transactivator. J. Immunol. 168, 4414–4419 (2002). 108. Eischen, C. M., Schilling, J. D., Lynch, D. H., Krammer, P. H. & Leibson, P. J. Fc receptor-induced expression of Fas ligand on activated NK cells facilitates cell-mediated cytotoxicity and subsequent autocrine NK cell apoptosis. J. Immunol. 156, 2693–2699 (1996). 109. Furuke, K., Shiraishi, M., Mostowski, H. S. & Bloom, E. T. Fas ligand induction in human NK cells is regulated by redox through a calcineurin–nuclear factors of activated T celldependent pathway. J. Immunol. 162, 1988–1993 (1999). 110. Crist, S. A., Griffith, T. S. & Ratliff, T. L. Structure/function analysis of the murine CD95L promoter reveals the identification of a novel transcriptional repressor and functional CD28 response element. J. Biol. Chem. 278, 35950–35958 (2003). 111. Norian, L. A. et al. The regulation of CD95 (Fas) ligand expression in primary T cells: induction of promoter activation in CD95LP–Luc transgenic mice. J. Immunol. 164, 4471–4480 (2000). 112. Xiao, S. et al. FasL promoter activation by IL-2 through SP1 and NFAT but not Egr-2 and Egr-3. Eur. J. Immunol. 29, 3456–3465 (1999). 113. McClure, R. F., Heppelmann, C. J. & Paya, C. V. Constitutive Fas ligand gene transcription in Sertoli cells is regulated by Sp1. J. Biol. Chem. 274, 7756–7762 (1999). 114. Kavurma, M. M., Bobryshev, Y. & Khachigian, L. M. Ets-1 positively regulates Fas ligand transcription via cooperative interactions with Sp1. J. Biol. Chem. 277, 36244–36252 (2002). 115. Kasibhatla, S., Beere, H. M., Brunner, T., Echeverri, F. & Green, D. R. A ‘non-canonical’ DNA-binding element mediates the response of the Fas-ligand promoter to c-Myc. Curr. Biol. 10, 1205–1208 (2000). 116. Torgler, R. et al. Regulation of activation-induced Fas (CD95/Apo-1) ligand expression in T cells by the cyclin B1/Cdk1 complex. J. Biol. Chem. 279, 37334–37342 (2004). 117. Brunner, T. et al. Expression of Fas ligand in activated T cells is regulated by c-Myc. J. Biol. Chem. 275, 9767–9772 (2000).
NATURE REVIEWS | IMMUNOLOGY
118. Mailliard, R. B. et al. Dendritic cells mediate NK cell help for TH1 and CTL responses: two-signal requirement for the induction of NK cell helper function. J. Immunol. 171, 2366–2373 (2003). 119. Smyth, M. J. et al. Perforin is a major contributor to NK cell control of tumor metastasis. J. Immunol. 162, 6658–6662 (1999). 120. van den Broek, M. E. et al. Decreased tumor surveillance in perforin-deficient mice. J. Exp. Med. 184, 1781–1790 (1996). 121. Mullbacher, A. et al. Granzyme A is critical for recovery of mice from infection with the natural cytopathic viral pathogen, ectromelia. Proc. Natl Acad. Sci. USA 93, 5783–5787 (1996). 122. Smyth, M. J., Street, S. E. & Trapani, J. A. Granzymes A and B are not essential for perforin-mediated tumor rejection. J. Immunol. 171, 515–518 (2003). 123. Heusel, J. W., Wesselschmidt, R. L., Shresta, S., Russell, J. H. & Ley, T. J. Cytotoxic lymphocytes require granzyme B for the rapid induction of DNA fragmentation and apoptosis in allogeneic target cells. Cell 76, 977–987 (1994). 124. Takahashi, T. et al. Generalized lymphoproliferative disease in mice, caused by a point mutation in the Fas ligand. Cell 76, 969–976 (1994). 125. Caldwell, S. A., Ryan, M. H., McDuffie, E. & Abrams, S. I. The Fas/Fas ligand pathway is important for optimal tumor regression in a mouse model of CTL adoptive immunotherapy of experimental CMS4 lung metastases. J. Immunol. 171, 2402–2412 (2003). 126. Dalton, D. K. et al. Multiple defects of immune cell function in mice with disrupted interferon-γ genes. Science 259, 1739–1742 (1993). 127. Flynn, J. L. et al. An essential role for interferon-γ in resistance to Mycobacterium tuberculosis infection. J. Exp. Med. 178, 2249–2254 (1993). 128. Cooper, A. M. et al. Disseminated tuberculosis in interferon γ gene-disrupted mice. J. Exp. Med. 178, 2243–2247 (1993). 129. Willenborg, D. O., Fordham, S., Bernard, C. C., Cowden, W. B. & Ramshaw, I. A. IFN-γ plays a critical down-regulatory role in the induction and effector phase of myelin oligodendrocyte glycoprotein-induced autoimmune encephalomyelitis. J. Immunol. 157, 3223–3227 (1996). 130. Willenborg, D. O., Fordham, S. A., Staykova, M. A., Ramshaw, I. A. & Cowden, W. B. IFN-γ is critical to the control of murine autoimmune encephalomyelitis and regulates both in the periphery and in the target tissue: a possible role for nitric oxide. J. Immunol. 163, 5278–5286 (1999). 131. Bettelli, E. et al. Loss of T-bet, but not STAT1, prevents the development of experimental autoimmune encephalomyelitis. J. Exp. Med. 200, 79–87 (2004). 132. Nishibori, T., Tanabe, Y., Su, L. & David, M. Impaired development of CD4+ CD25+ regulatory T Cells in the absence of STAT1: increased susceptibility to autoimmune disease. J. Exp. Med. 199, 25–34 (2004). 133. Stamm, L. M., Satoskar, A. A., Ghosh, S. K., David, J. R. & Satoskar, A. R. STAT-4 mediated IL-12 signaling pathway is critical for the development of protective immunity in cutaneous leishmaniasis. Eur. J. Immunol. 29, 2524–2529 (1999). 134. Neurath, M. F. et al. The transcription factor T-bet regulates mucosal T cell activation in experimental colitis and Crohn’s disease. J. Exp. Med. 195, 1129–1143 (2002).
Acknowledgements Support was provided by grants from the National Institutes of Health (Bethesda, United States) and the Juvenile Diabetes Research Foundation (New York, United States). M.J.T. is supported by a Postdoctoral Fellowship from the Cancer Research Institute (New York). G.M.L. is a Medical Research Council (United Kingdom) Clinician Scientist.
Competing interests statement The authors declare competing financial interests: see Web version for details.
Online links DATABASES The following terms in this article are linked online to: Entrez Gene: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=gene EOMES | FASL | granzyme B | IFN-γ | perforin | T-bet Access to this interactive links box is free online.
VOLUME 4 | NOVEMBER 2004 | 9 1 1 ©2004 Nature Publishing Group
PERSPECTIVES
SCIENCE AND SOCIETY
Complementary and alternative medicine: assessing the evidence for immunological benefits Martin H. Goldrosen and Stephen E. Straus Abstract | With words such as AIDS, allergy and autoimmunity embedded in the popular lexicon, we often equate health with the precision and the tenor of responses to allergens and microorganisms. This leads many people to seek their own solutions to sustain, restore or even boost their immune competence, hoping to live more comfortably and longer. Here, we consider the social and clinical contexts in which these promises of enhanced immunity are pursued through popular practices known as complementary and alternative medicine and the evidence that supports these.
Complementary and alternative medicine (CAM) consists of diverse clinical interventions that are popular yet not embraced by conventional medicine because there is insufficient proof that they are safe and effective. Complementary interventions are used together with conventional treatments, whereas alternative interventions are used instead of them. In 1998, the National Center for Complementary and Alternative Medicine (NCCAM) was established by the US Congress at the National Institutes of Health (Bethesda, United States) to rigorously investigate popular CAM modalities to determine which are beneficial and worthy of further consideration for mainstream practice1. Among the many CAM approaches that warrant careful investigation are those that claim to sustain, restore or boost immunity. Here, we discuss who uses CAM, the social, ethical
912
and legal issues regarding its use, the benefits versus the risks, the complexities of carrying out clinical research with CAM therapeutic modalities, and the implications of these therapies. In particular, we discuss these issues in the context of the immune system. Use of CAM
When originally developing its research strategies and priorities, NCCAM reviewed more than 800 types of CAM practice, including many with potential immunological effects. These can be grouped into five main domains: biologically based approaches, manipulative and body-based therapies, mind–body interventions, energy therapies and alternative medical systems (FIG. 1). There is little overlap between the therapeutic modalities in different domains, with the exception of the alternative medical systems, which draw on therapies from the other four domains. For example, AYURVEDA (see Glossary), the traditional system of medicine in India, addresses conditions of the mind, body and spirit through the use of diet, exercise, MEDITATION, HERBS, MASSAGE, YOGA and exposure to sunlight. The prevalence of the use of CAM throughout developed countries ranges from 9% of individuals to 65% (REFS 2,3). This variation results from the disparity in definitions of CAM, the selection of therapies assessed, the method of survey used and the period of CAM use. By one often-cited estimate, between 1990 and 1997, visits to
| NOVEMBER 2004 | VOLUME 4
CAM practitioners in the United States increased from 427 million to 629 million, thereby exceeding the number of visits to all primary-care physicians by 243 million4. In the most recent and comprehensive survey of CAM use3, CAM was most often used to treat back pain or back problems, head or chest colds, neck pain or neck problems, joint pain or stiffness, and anxiety or depression. The ten most common CAM therapies used during the previous year were prayer for one’s own health (used by 43.0% of individuals), prayer by others for one’s own health (24.4%), natural products (18.9%), deep-breathing exercises (11.6%), participation in a prayer group for one’s own health (9.6%), meditation (7.6%), CHIROPRACTIC care (7.5%), yoga (5.1%), massage (5.0%) and diet-based therapies (3.5%)3. CAM use was more prevalent among women, older adults and those with higher educational attainment (more than 16 years of schooling)3,5. There are several reasons why CAM therapies are popular. CAM practitioners aim not only to treat the physical and biochemical manifestations of illness but also to consider the nutritional, emotional, social and spiritual context in which the illness arises. The use of body-based techniques and mind– body interventions is comforting and can reduce stress. Furthermore, numerous CAM products can be purchased at health-food shops and supermarkets and, as natural substances that have often been used for centuries, they claim to be effective yet safer than drugs. CAM places patients in control of their own health. Regulation of CAM
The average consumer, however, does not understand how CAM products are regulated. In the United States, in general, DIETARY SUPPLEMENTS are regulated by the Food and Drug Administration (FDA) as foods, not as drugs, in accordance with the Dietary Supplement Health and Education Act of 1994 (DSHEA)114. Whereas drugs must be approved by the FDA as being safe and
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
PERSPECTIVES
Manipulative and body-based therapies
Biologically based approaches
Massage Chiropractic Osteopathy
Diets Herbs Vitamins
Reiki Magnets Qigong Energy therapies
Common CAM practices
Yoga Spirituality Relaxation Mind–body interventions
Homeopathy Naturopathy Ayurveda
Alternative medical systems
Many other countries regulate dietary supplements more stringently8. For example, in most countries of the European Union, HERBAL MEDICINES are sold in pharmacies as licensed non-prescription or prescription medicines. The sale of herbal medicinal products is allowed without the requirement to carry out specific clinical studies provided that their safety and efficacy are supported by common experience accrued through long-standing use. The quality aspect of the medicinal product is independent of its traditional use, so no derogation can be made with regard to the necessary physiochemical, biological and microbiological tests. Moreover, in 2002, the European Union defined the upper limits of safe levels of VITAMINS and minerals, and required products that could exceed these levels to carry explicit warnings9. There is no similar requirement in the United States. Risks of CAM use
Figure 1 | CAM domains and some of the most common examples. Biologically based complementary and alternative medicine (CAM) approaches include herbal medicines, ‘megadoses’ of vitamins and SPECIAL DIETS107, such as those proposed by Drs Atkins108 and Ornish109. Manipulative and body-based therapies include methods that involve manipulation and/or movement of the body, such as massage, chiropractic and osteopathy. Mind–body interventions use various techniques that are designed to facilitate the capacity of the mind to affect bodily function and symptoms, including yoga, prayer, meditation, spirituality and guided imagery. Energy therapies are intended to affect energy fields (biofields) that purportedly surround and penetrate the human body, using REIKI and therapeutic touch. Alternatively, energy therapies can involve the unconventional use of ELECTROMAGNETIC FIELDS, such as pulsed fields, magnetic fields, or alternating- or direct-current fields. Alternative medical systems involve complete systems of theory and practice that have evolved independently of, and often before, the conventional biomedical approach. Many of these are traditional systems of medicine that are practised by individual cultures throughout the world, such as traditional Chinese medicine (of which acupuncture is a principal component) and Ayurvedic medicine from India, but they also include the more-modern Western approaches that are applied in HOMEOPATHY and naturopathy.
effective before they can be sold to treat or prevent specific diseases, the FDA is not authorized to evaluate the safety or efficacy of dietary supplements before their marketing. However, the FDA can ban the sale of supplements that are shown to be unsafe, and it did so recently for the first time, for products containing ephedra (Ephedra sinica) — a Chinese herb used for weight loss and athletic-performance enhancement. As was evident from the prolonged efforts of the FDA to curtail the use of ephedra, it is difficult to prove that such products are unsafe. Whereas manufacturers of drugs must report adverse drug events to the FDA after a drug has been marketed, manufacturers of dietary supplements are not required to collect or file such reports. Furthermore, under the DSHEA regulations, manufacturers of dietary supplements can claim that their products maintain the normal ‘structure or function’ of bodily systems, but they cannot claim that they treat or prevent diseases. For example, a herbal manufacturer
needs no proof to claim that its products maintain a healthy immune system or healthy joints. However, a manufacturer cannot claim that its products treat states of immune deficiency or arthritis, even though the distinctions between these sets of claims could be subtle. Consumers in the United States spend nearly $20 billion each year on dietary supplements that do not require a prescription, and they must rely on the manufacturers to provide truthful and non-misleading information regarding the safety and benefit of their products. Surveys have shown that 59% of Americans believe that the FDA regulates dietary supplements in the same manner as drugs because they are often sold next to over-the-counter medications 6. Although the DSHEA legislation affords the public unfettered access to thousands of dietary supplements, some people have proposed that public health would be better served if the FDA required at least evidence of product safety before sale7.
NATURE REVIEWS | IMMUNOLOGY
Using CAM can have risks. Most importantly, it could distract patients from pursuing well-established treatments. However, the authors of a survey published in 1997 reported that only 4.4% of respondents relied solely on alternative therapies4. So, most individuals are not turning away from conventional medicine but use unconventional approaches as complements to conventional ones10. Unfortunately, until recently, only 40% of patients disclosed their use of CAM therapies to physicians4. This could have serious consequences because several herbal medicines can have harmful effects and others can markedly alter the metabolism of conventional drugs. For example, aristolochia (Aristolochia fangchi), a Chinese herb incorporated in a weight-reducing product in error, is associated with genitourinary cancers11; the anxiolytic herb kava (Piper methysticum) is associated with hepatic failure12; and ephedra is associated with strokes13. St John’s wort (Hypericum perforatum), which is widely used for the treatment of depression, upregulates the expression of several cytochrome oxidases (such as CYP3A); together, these affect the metabolism of more than half of all conventional drugs14, including the HIV-protease inhibitor indinavir15, the topoisomerase inhibitor irinotecan (used in multi-drug treatment of solid tumors)16,17 and the potent immunosuppressant drug cyclosporine (used to reduce the risk of transplant rejection)18,19. Taking St John’s wort extract in combination with a low-dose oral contraceptive increased intracyclic bleeding episodes, thereby increasing the risk of unintended pregnancies20,21.
VOLUME 4 | NOVEMBER 2004 | 9 1 3 ©2004 Nature Publishing Group
PERSPECTIVES
Box 1 | Challenges of conducting clinical trials of CAM • Use of a multiple-modality complex intervention, rather than a single intervention, in some forms of complementary and alternative medicine (CAM). • Development of a specific individualized treatment for each patient that often focuses on the symptoms of the disease rather than on one main pathology. • Finding appropriate placebos or shams for manipulative and body-based interventions. • Accruing, randomizing and retaining patients with strong opinions favouring or rejecting CAM. • Availability of standardized and well-characterized herbal preparations.
small and large randomized trials. In this article, we carefully evaluate the immune changes that are induced by these selected CAM modalities. The emphasis is mainly on dietary interventions, because the clinical data for some of these are the most compelling. For other CAM approaches that are not directly addressed here, there is minimal objective evidence for beneficial immunological or clinical effects. Dietary supplements
Clinical trials of CAM BOX 1 summarizes the complexities of carry-
ing out clinical trials to verify the safety and efficacy of CAM modalities22. Despite these difficulties, NCCAM has been able to support well-designed randomized clinical trials of CAM treatments. Ongoing trials include the following: a comparison of a NATUROPATHIC TREATMENT, a TRADITIONAL CHINESE MEDICINE treatment and a conventional treatment for women with temporomandibular-joint disease; a large study of acupuncture for the treatment of osteoarthritis; and many studies of individual dietary supplements for the prevention or treatment of diseases. The status of NCCAM’s ongoing phase III clinical trials is shown in TABLE 1. The decision to undertake each of these trials was based on the following: the importance of the disease to public health, the quality of the preliminary data, the availability of a well-characterized intervention, the extent of public use and the cost of doing the research22. It is notable that CAM approaches for which the main target is the immune system are not represented on this list, because the existing data for such interventions do not currently warrant phase III trials. Here, we summarize what is known about these approaches from the limited studies that have been carried out. CAM and immunity
Among the many mechanisms by which CAM approaches are claimed to function, their purported effects on immunity resonate
with the contemporary appreciation that health depends considerably on immune competence. Even before the concept of an immune system was articulated, vitalistic practitioners, such as naturopathic physicians, maintained that disease should be treated by stimulating the ability of the body to heal itself rather than by treating symptoms. Now, we know that infections and cancers can result from a loss of ‘optimal’ immune surveillance. In addition, at present, although the aetiologies for some disorders, such as multiple sclerosis and sarcoidosis, remain unproven, they are nonetheless best understood in immunological terms. Moreover, popular literature is rife with speculations that debilitating fatigue, behavioural disorders and many other disorders could be immunologically mediated: for examples, see the Chronic Fatigue and Immune Dysfunction Syndrome Association of America and The Autism Autoimmunity Project in the Online links box. So, the appeal of CAM approaches that claim to alter immunity can be readily understood. But, do these claims engender unjustified optimism or might some of them have a basis in fact? There are numerous publications describing the therapeutic efficacy of CAM modalities that are thought to mediate their effects through the immune system, all reporting varying degrees of evidence. In FIG. 2, the hierarchy of available evidence is characterized; it ranges from anecdotes and case studies to large randomized clinical trials. TABLE 2 summarizes selected interventions that have been associated with meaningful clinical changes in
Dietary supplements are consumed in a formulation containing a predetermined dose (such as in capsule form) and include vitamins, minerals, essential fatty acids, herbal medicines, amino acids and enzymes. Nutraceuticals are dietary supplements that deliver a concentrated form of a presumed bioactive agent from a food, are presented in a non-food matrix and are used to enhance health in dosages that exceed those that could be obtained from normal foods7. A subset of dietary supplements functions as immunostimulatory nutrients because they have the potential to modulate the activity of the immune system23. Among them, the most studied are vitamins A and E, the mineral zinc and omega-3 polyunsaturated fatty acids (PUFAs). Vitamins. Oxidative stress occurs when there is an imbalance between the generation of free radicals by reactive-oxygen species and the level of endogenous anti-oxidants in cells and tissues. The consequences of oxidative stress might include the development of degenerative disorders, such as cancer, cardiovascular disease and ageing (with its attendant immune senescence)24,25. Higher organisms have evolved various endogenous defence mechanisms to prevent the generation of reactive-oxygen species. For example, catalase and glutathione peroxidase decompose the hydrogen peroxide that is involved in the killing of microorganisms23. Dietary supplements such as vitamin E can provide an essential exogenous source of anti-oxidants.
Table 1 | Some ongoing, large phase III trials of CAM modalities CAM approach
Disease/condition
Status of trial
Acupuncture
Osteoarthritis
Study completed; data being analysed
Glucosamine chondroitin sulphate
Osteoarthritis
Enrolment completed; studies ongoing
Vitamin E
Prostate cancer
Enrolment completed; study ongoing
Shark cartilage
Lung cancer
Enrolment ongoing; study ongoing
St John’s wort (Hypericum perforatum)
Minor depression
Recruitment ongoing
EDTA-chelation therapy
Coronary-artery disease
Recruitment ongoing
Saw palmetto (Serenoa repens)
Benign prostate hypertrophy
Final protocol developed in 2004
CAM, complementary and alternative medicine; EDTA, ethylenediaminetetraacetic acid.
914
| NOVEMBER 2004 | VOLUME 4
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
PERSPECTIVES
Large randomized clinical trials Small randomized clinical trials Uncontrolled trials Observational studies Case studies Anecdotes
Figure 2 | Hierarchy of evidence. Information regarding the efficacy and safety of any clinical approach, including those of complementary and alternative medicine, spans a continuum that ranges from anecdotes and retrospective studies to small randomized, controlled trials (phase II clinical trials) and large randomized, controlled trials (phase III clinical trials).
Vitamin E (α-tocopherol) is fat soluble and functions as an anti-oxidant and freeradical scavenger in cell membranes by blocking the peroxidation of PUFAs26–28. So, vitamin E protects cell-membrane lipids from peroxidation. A deficiency in vitamin E is associated with decreased T-cell and B-cell mitogenesis29, interleukin-2 (IL-2) production, natural killer (NK)-cell activity30 and neutrophil phagocytosis in both rats and
humans. Conversely, supplementation with vitamin E increases T-cell mitogenesis, IL-2 production31 and NK-cell activity32 in rats, as well as T-cell and B-cell mitogenesis33 and antibody production34 in mice. Vitamin E deficiency is exacerbated in elderly people who are malnourished35. Supplementation of elderly individuals with pharmacological doses of vitamin E augments their immune responses. Specifically, vitamin E has been shown to increase delayed-type hypersensitivity (DTH) responses in elderly individuals, with 200 mg each day having the maximal effect36. The 200 mg dose also increased antibody responses after vaccination with hepatitis-B-virus surface antigen, tetanus toxoid or pneumococcal polysaccharides. Recently, Meydani and co-workers reported on a randomized, double-blinded, placebocontrolled trial of 617 people of 65 years of age or more who were given 200 international units (IU) of vitamin E each day for 1 year37. Supplementation with 200 IU of vitamin E each day did not have a significant effect on the incidence of lower respiratory-tract infections in elderly nursing-home residents. However, the authors observed a protective effect of vitamin E supplementation on
upper respiratory-tract infections (particularly the common cold) that merits further investigation. The risks of contracting some infectious diseases are increased in the setting of vitamin deficiency. The evidence for this is strongest for vitamin A (retinol)25. In turn, infectious diseases can then precipitate or exacerbate vitamin deficiencies by promoting decreased food intake38, impairing nutrient absorption39 and causing direct nutrient loss in sweat, stools and urine40. Vitamin A deficiency broadly impairs innate immunity by reducing the function of neutrophils, macrophages and NK cells41,42. In addition, vitamin A deficiency inhibits antibody responses directed by T helper 2 (TH2) cells43 and impedes the normal regeneration of mucosal barriers that are damaged by viral infection25. Large community-based studies have shown that vitamin A deficiency is associated with a markedly increased risk of child mortality44. Worldwide, ∼10% of the vitamin-Adeficient children who die are infected with measles virus, often complicated by bacterial pneumonia. Supplementation with high doses of vitamin A improves recovery from measles by decreasing both the duration of
Table 2 | Some CAM modalities that might mediate their effects through the immune system CAM intervention
CAM domain
Level of evidence
Condition
Dose
Subjects
Observation
Reference
Vitamin A
Biological (dietary supplement)
Meta-analysis of 12 randomized clinical trials
Vitamin-Adeficient children
50,000 IU per day
NA
30% reduction in childhood mortality
45
Vitamin E
Biological (dietary supplement)
Small randomized clinical trial
Healthy elderly people (>65 years)
200 mg per day
88
Increase in clinically relevant immunity
36
Zinc
Biological (dietary supplement)
Large randomized clinical trial
Zincdeficient children
10 mg per day
579
26% reduction in childhood diarrhoea
56
Omega-3 PUFA
Biological (dietary supplement)
Large randomized clinical trial
Patients surviving myocardial infarction
1,000 mg per day omega-3 PUFA
11,323
45% reduction in sudden cardiac death
68
Echinacea
Biological (herbal product)
Small randomized clinical trial
Common cold
5 ml per day Echinacea purpurea extract*
80
Reduction in rhinorrhoea from 9 to 6 days
79
Thunder God vine
Biological (herbal product)
Small randomized clinical trial
Rheumatoid arthritis
360 mg per day for 20 weeks
35
66% of patients had a 20% reduction in disease severity
83
Lactobacillus casei GG
Biological (probiotic)
Small randomized clinical trial
Childhood rotaviral diarrhoea
3 × 109 colonyforming units
100
Reduction in diarrhoea from 6 to 3 days
90
T’ai chi
Mind–body
Small randomized clinical trial
Shingles (varicellazoster virus infection)
15-week intervention, 45 sessions
36
50% increase in varicella-zoster-virusspecific immunity
103
Hypnosis
Mind–body
Small randomized clinical trial
Genital herpes (herpes-simplex virus infection)
6 weeks of self-hypnosis
21
50% reduction in herpes-simplex virus recurrence
104
*Negative results were obtained using a different echinacea preparation (see text for details). CAM, complementary and alternative medicine; IU, international units; NA, not applicable; PUFA, polyunsaturated fatty acids.
NATURE REVIEWS | IMMUNOLOGY
VOLUME 4 | NOVEMBER 2004 | 9 1 5 ©2004 Nature Publishing Group
PERSPECTIVES illness and its overall mortality rate44,45. Moreover, in compelling studies of this inexpensive intervention, vitamin A supplements were shown to significantly reduce overall early-childhood mortality from measles by 30% (REF. 45). It is therefore recommended that vitamin A be given to all children suffering from measles, regardless of whether they are deficient in vitamin A46. It is less clear whether supplementation with vitamin A improves morbidity and mortality caused by other infections, and the benefits of vitamin A in terms of measles do not justify universal supplementation. Furthermore, modest adverse effects, such as exacerbation of the inflammatory process and worsening of clinical symptoms and signs, were noticed in children with bacterial pneumonia who were given high doses of vitamin A47. Vitamin A might be useful for treating patients with specific immunodeficiency states. For example, in one small study of patients with common variable immunodeficiency, decreased serum levels of vitamin A were observed48. Vitamin A supplementation of these patients resulted in a significant increase in CD40-stimulated production of IgG, serum levels of IgA and phytohaemagglutinininduced proliferation of peripheral-blood mononuclear cells48. Several studies also found that low serum concentrations of vitamin A correlate with HIV-associated disease severity and progression49. However, researchers studying the effects of vitamin A supplementation on the maternal–fetal transmission of HIV have reported conflicting findings50,51. Minerals. Zinc is a cofactor for many enzymatic reactions. Patients with acrodermatitis enteropathica — an autosomal recessive disorder attributed to a defect in zinc metabolism — suffer from T-cell dysfunction and diarrhoea, and are more susceptible to viral, bacterial and fungal infections52,53. In humans, zinc deficiency is associated with a decreased number of lymphoid precursors and impaired activity of TH1 cells, as manifested by decreased production of interferon-γ (IFN-γ), IL-2 and tumour-necrosis factor (TNF), but it has no effect on the production of IL-4, IL-6 or IL-10 (REFS 50,54). There is growing evidence that zinc deficiency is a clinically important problem. In some studies of malnourished children, supplementation with zinc decreased the incidence of diarrhoea by more than 50% (REF. 55). In other studies, supplementation with zinc was effective for the treatment of diarrhoea in undernourished children only when it was begun within the first 4 days of
916
illness56. Zinc has also been evaluated for both the prevention and treatment of viral colds, with conflicting results being reported57. Even in developed countries, malnutrition can be prevalent in elderly people, further impairing an already senescent immune system and increasing the risk and severity of infections. Supplementation of malnourished older adults with zinc enhances DTH responses and increases lymphocyte numbers and NK-cell activity58. However, it is unclear whether supplementation with zinc results in a reduced number of infections, because studies evaluating zinc were carried out in combination with supplementation with other trace elements. For example, studies of elderly people who are malnourished have shown a clinical benefit from supplementation with zinc in combination with selenium, with or without vitamins A, C and E. During 2 years of supplementation, individuals who received the two trace elements alone or in combination with the vitamins had significantly fewer infectious events compared with patients who were administered a placebo or vitamins alone59. It should be noted, however, that excessive zinc intake impairs immune responses60, indicating the importance of dose regulation of dietary supplements. A high zinc intake can result in depletion of copper, and copper deficiency by itself impairs immune function. Copper is required for the proper structure and function of cytochrome C oxidases in the mitochondrial electron-transport chain of immune cells61. From the available studies, no consensus has been reached regarding the implications of vitamin and mineral supplementation for the general population. At present, dietary-supplement manufacturers and some authorities are pressing for universal supplementation. Other leading authorities do not consider that the evidence justifies supplementation for all healthy people who have normal diets. Essential fatty acids. The PUFA content of immune cells varies according to the cell type and the lipid fraction examined. Typically, the phospholipid portion of human peripheralblood mononuclear cells comprises 15–25% arachidonic acid (an omega-6 PUFA) and only 0.1–0.8% eicosapentaenoic acid (EPA), docosahexaenoic acid (DHA) and omega-3 PUFAs62,63. Arachidonic acid gives rise to the eicosanoid family of inflammatory mediators (prostaglandins, leukotrienes and related metabolites), which are involved in increasing the intensity and duration of inflammatory and immune responses.
| NOVEMBER 2004 | VOLUME 4
Ingestion of fish oils — a good source of omega-3 PUFAs — decreases the amount of arachidonic acid in the cell membrane that is available for eicosanoid production and thereby decreases inflammation62,63. Additional anti-inflammatory effects of omega-3 PUFAs are elicited by eicosanoidindependent mechanisms that involve downstream intracellular-signalling pathways. In animals and humans, ingestion of fish oils decreases generation of reactive-oxygen species, production of TNF, IL-1 and IL-6 (REF. 64), proliferation of lymphocytes and release of IL-2 and IFN-γ 65. So, supplementation of the diet with high levels of fish oils has a potent anti-inflammatory effect but impairs neutrophil, monocyte and lymphocyte responses. The impairment of immune function could be ameliorated without restoring the inflammatory response by taking 200 mg of vitamin E each day 63,66. Recently, the Agency for Healthcare Research and Quality in the United States evaluated the effects of omega-3 PUFAs on cardiovascular disease, in which inflammation has a central role67. The anti-inflammatory effects of omega-3 fatty acids might contribute to their beneficial cardiac effects. On the basis of 39 studies, the report concludes that the consumption of omega-3 PUFAs, fish or fish oil reduces mortality from all causes and reduces the incidence of various outcomes of cardiovascular disease, such as sudden death, cardiac death and myocardial infarction. For example, among 11,323 patients surviving myocardial infarction, there was a 45% reduction in sudden cardiac death in the group of patients given 1 g of fish-oil-derived omega-3 PUFAs each day 68. Several trials have assessed the impact of dietary supplementation with fish oils on chronic inflammatory disorders, such as rheumatoid arthritis69, Crohn’s disease70, asthma71,72 and IgA nephropathy73. In some of these studies, lower levels of leukotriene B4, IL-1 and C-reactive protein were observed. Reduced symptoms were also observed in these small-scale clinical trials of fish-oil supplementation, with the strongest evidence for rheumatoid arthritis69. It is unclear from these studies what the optimal dose of fish oil is and whether there is any benefit of using different omega-3 PUFAs in combination. Herbal medicines. Herbal products were the mainstay of indigenous medical practices for millennia and remain the only widely available treatments in some countries8. Also, it is estimated that approximately half of all pharmaceuticals are derived from natural products74, including morphine, digitalis, quinine,
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
PERSPECTIVES
Table 3 | Some herbal products that modulate immune responses Common name
Latin name
Active components
Plant sources
Common uses
Immune effects
Cat’s claw
Uncaria tomentosa
Indole alkaloids
Root
Immune stimulant
Increases lymphocyte proliferation
110
Curcumin (turmeric)
Curcuma longa
Curcuminoids
Rhizomes
Anti-inflammatory
Inhibits production of pro-inflammatory cytokines
111
Echinacea
Echinacea purpurea, Echinacea angustifolia and Echinacea pallida
Echinacosides
Root, shoots and leaves
Against common cold and influenza
Extract of E. purpurea results in the production of pro-inflammatory cytokines
77
Ginseng
Panax ginseng, Panax quinquefolium and Eleutherococcus senticosus
Ginsengosides
Root
General tonic
P. ginseng induces the production of pro-inflammatory cytokines
112
Marijuana
Cannabis sativa
Cannabinoids
Leaves and flowers
Narcotic
Suppresses the cell-mediated immune response by induction of apoptosis
113
Thunder God vine
Tripterygium wilfordii Hook F
Terpenoids (di-, tri- and sesquiwilfortrine)
Root
Antiinflammatory
Inhibits production of pro-inflammatory cytokines
vincristine, taxol and artemisinin. Today, herbal products remain popular, even among people who also use pharmaceuticals. TABLE 3 lists several herbal products that have been shown to modulate the immune response in the following ways: stimulation of the immune response (Cat’s claw); inhibition of the immune response (marijuana); induction of pro-inflammatory cytokine production (echinacea and ginseng); and inhibition of pro-inflammatory cytokine production (curcumin and Thunder God vine). Clinical trials of echinacea highlight the complexity of studies of herbal products. Three different species of Echinacea (Echinacea angustifolia, E. purpurea and E. pallida) are used for medicinal purposes and marketed as echinacea75. Preparations are made from roots, above-ground parts (stems, leaves and flowers) or a mixture of both. Furthermore, some preparations of echinacea are aqueous extracts, whereas others are alcoholic extracts. Differences between the species, the soil in which the plants are cultivated, the parts of the plant used and the extraction procedures can result in substantial differences in chemical composition and biological activity76. Immunological studies of echinacea have used aqueous or alcoholic extracts, as well as purified polysaccharides — the compounds thought to be responsible for the non-specific immune stimulatory effects of echinacea. Both extracts of echinacea and purified polysaccharides were shown to stimulate phagocytosis by animal and human neutrophils in vitro75. Human macrophages that were
exposed in vitro to commercial preparations of echinacea released substantially more TNF, IL-1, IL-6 and IL-10 — a pattern that is consistent with that of activated macrophages77. Despite these reproducible in vitro activities, it has been difficult to confirm in placebocontrolled trials that echinacea extracts have salutary effects on the common cold78. Whereas echinacea has not proven to be effective in preventing the common cold, conflicting results have been observed in trials designed to treat the common cold. A reduction in median duration of illness (6 days compared with 9 days for those taking the placebo) was observed in a double-blinded, placebo-controlled trial of 80 adults with early symptoms of the common cold who were given 5 ml of pressed juice from freshflowering E. purpurea twice daily for 10 days79. By contrast, no benefit was observed in a placebo-controlled trial of 148 students with early cold symptoms who were given a mixture of unrefined E. purpurea herb (25%) and root (25%) and E. angustifolia root (50%) taken in 1 g doses six times on the first day of illness and three times on each subsequent day of illness for up to 10 days80. Although herbal products such as echinacea can non-specifically enhance immune activities, other herbal products can inhibit them. For example, Thunder God vine (Tripterygium wilfordii Hook F), which has been used in traditional Chinese medicine for more than two centuries, has both immunosuppressive and anti-inflammatory effects81,82. It contains triptolide and related diterpenoid
NATURE REVIEWS | IMMUNOLOGY
Reference
81
components. Triptolide inhibits the proliferation of T cells by preventing transcription of the gene encoding IL-2 in response to antigenic and mitogenic stimulation. Triptolide also inhibits the production of pro-inflammatory cytokines and mediators, including TNF, IL-1, IL-6, IL-8 and prostaglandin E2. In animal models, triptolide prolongs allograft survival and inhibits graft-versus-host disease81. Thousands of patients in China with autoimmune and inflammatory diseases, particularly rheumatoid arthritis, have been treated with triptolide82. The first small randomized, placebo-controlled trial of triptolide showed a significant dose-dependent response in patients with rheumatoid arthritis83. However, in larger, uncontrolled studies, triptolide was observed to have renal, cardiac, haematopoietic and reproductive toxicities81,82. Probiotics. The health benefits of fermented foods have been recognized since ancient times, although the mediators of fermentation — bacteria and yeast — were only identified in the nineteenth century. Today, food supplements that contain live microorganisms (in the form of capsules, powders, enriched yoghurts and milks) are commonly consumed to treat gastrointestinal and urogenital-tract infections84; these are known as probiotics, derived from the Greek words meaning ‘for life’. Lactobacillus and Bifidobacterium species are the most frequently used in probiotics. Multiple mechanisms are postulated to explain their actions, including the following: reduction of lactose
VOLUME 4 | NOVEMBER 2004 | 9 1 7 ©2004 Nature Publishing Group
PERSPECTIVES
Glossary AYURVEDA
The traditional Indian system of medicine. Ayurvedic (meaning ‘science of life’) medicine is a comprehensive system that places equal emphasis on the body, mind and spirit, and it strives to restore the innate harmony of the individual.
an illness cures it when administered in very small doses. Homeopathic physicians believe that the more dilute the remedy, the greater its potency. Therefore, homeopaths treat illness by using small doses of specially prepared plant extracts and minerals to stimulate the defence mechanisms and healing processes of the body.
BIOFEEDBACK
A process for monitoring a body function (such as breathing, heart rate or blood pressure) and altering the function through relaxation or imagery. CHIROPRACTIC
A system of treatment that is based on the relationship between structure (mainly of the spine) and function, and how that relationship affects the preservation and restoration of health.
HYPNOSIS
An alternative state of consciousness in which the attention of an individual is focused away from the present reality and towards particular images, thoughts, perceptions, feelings, motivations, sensations, behaviours or any combination of these.
The manipulation of the soft tissues of the body to normalize them. MEDITATION
A self-directed method for relaxing the body and calming the mind. The practitioner makes a concentrated effort to focus on a single thought to still the inclination of the mind to mull over the many demands and details of daily life.
ELECTROMAGNETIC FIELDS
Magnetic fields can be used therapeutically to create a static force on the body for the purpose of relieving pain.
NATUROPATHIC TREATMENT
HERBS
A system of treatment that views disease as a manifestation of alterations in the processes by which the body naturally heals itself. It emphasizes restoration of health, as well as treatment of disease. Naturopathic physicians use an array of healing practices, including diet and clinical nutrition, homeopathy, acupuncture and herbal medicine.
Plants or plant products that produce or contain chemicals that act on the body.
QIGONG
HERBAL MEDICINES
Individual herbs or mixtures of herbs that are used for therapeutic value.
HOMEOPATHY
A Western system of medicine that is based on the principle that ‘like cures like’ — that is, the same substance that in large doses produces the symptoms of
content in milk products by β-galactosidase (which is produced by Lactobacillus spp.); release of antimicrobial agents, such as organic acids, free fatty acids, ammonia, hydrogen peroxide and bacteriocins85; competition for the ecological niche that would otherwise be occupied by pathogenic organisms86; and immunomodulation87,88. Importantly, probiotic bacteria can modulate the intestinal IgA response. For example, an increase in the number of IgA-secreting cells and the amount of rotavirus-specific IgA in the serum was observed in patients with rotaviral diarrhoea who were treated with Lactobacillus casei GG89. Probiotics have been evaluated as treatments for several conditions: diarrhoea in children and adults, atopic disease in children, inflammatory bowel disease and urogenitaltract infections. The strongest data are for the prevention and treatment of rotaviral diarrhoea in children, for which administration of Lactobacillus spp.-containing probiotics was associated with significant reductions in
918
Means ‘universal life energy’ in Japanese. It is based on the belief that, by channelling spiritual energy through the practitioner, the spirit is healed, and the spirit, in turn, heals the physical body. SPECIAL DIETS
Diet therapies that are believed to prevent and/or control illness and/or promote health, such as those proposed by Drs Atkins and Ornish. SPIRITUALITY
An inner sense of something greater than oneself. Recognition of a meaning to existence that transcends one’s immediate circumstances.
MASSAGE
DIETARY SUPPLEMENTS
Products that contain one or more ingredients (such as vitamins, minerals and herbs) that are intended to supplement the diet, are intended for human use and are in the form of a tablet, capsule, powder or another preparation that is not a conventional food.
REIKI
TRADITIONAL CHINESE MEDICINE
A system of treatment that emphasizes the proper balance or disturbances of qi (vital energy) in health and disease, respectively. Traditional Chinese medicine consists of a group of techniques and methods, including acupuncture, herbal medicine, oriental massage and qigong (a form of energy therapy). VISUAL IMAGERY
A flow of thoughts that includes sensory qualities such as smell, touch, hearing, taste, motion and images. VITAMINS
A general term for various unrelated organic substances that occur in many foods in small amounts and that are required for the normal metabolic functioning of the body. YOGA
A component of traditional Chinese medicine that combines movement, meditation and the regulation of breathing to enhance the flow of vital energy (qi, pronounced chi) in the body, to improve blood circulation and to enhance immune function.
the duration and severity of the infection90,91. Furthermore, in a double-blinded, placebocontrolled trial, Kalliomaki and co-workers showed that ingestion of L. casei GG prevented atopic disease in children who were at high risk92. However, conflicting results were obtained in adults with traveller’s diarrhoea93,94. The value of probiotics for the treatment of inflammatory bowel disease95 and urogenital-tract infections96 remains to be determined. Mind–body approaches
Mind–body medicine focuses on interactions between the brain, mind, body and behaviour and the powerful ways in which emotional, mental, spiritual and behavioural factors can directly affect health. Mind–body medicine uses interventions such as relaxation, HYPNOSIS, VISUAL IMAGERY, meditation, yoga, BIOFEEDBACK, T’ai chi, QIGONG, SPIRITUALITY and prayer to promote health. Numerous studies indicate that the psychological state of an individual influences
| NOVEMBER 2004 | VOLUME 4
An ancient system of practices originating in India. It is aimed at integrating mind, body and spirit to enhance health and well-being. There are many different forms of yoga. Hatha yoga — the most widely practised form of yoga in the Western world — uses specific postures and breathing exercises.
many facets of health, including susceptibility to illness, outcomes of infection, general wellbeing and senescence97. Researchers in the field of psychoneuroimmunology have documented immune defects in association with several diseases and with stress98. However, most of these studies have been correlative, which thereby prevents statements of causality among identified psychological, immune and health outcomes. Among the most compelling studies have been those by Cohen and colleagues99, who looked at the relationship between psychological stress and resistance to infection. To evaluate this, they prospectively studied the association between psychological stress and the frequency of documented clinical colds among individuals who were intentionally exposed to respiratory viruses. They observed that psychological stress was associated in a dose-dependent manner with an increased risk of acute infectious respiratory illness and that this risk was attributable to increased rates of infection.
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
PERSPECTIVES Psychological stress can affect immunity through the hypothalamic–pituitary–adrenal (HPA) axis and the sympathetic–adrenal medullary (SAM) axis100. Stress-induced activation of the HPA axis results in the release of neuroendocrine hormones, such as adrenocorticotropin, from the anterior pituitary gland. Adrenocorticotropin then circulates through the bloodstream to the adrenal glands, where it induces the release of glucocorticoids, which then bind receptors at the cell surface of lymphoid and myeloid cells101. Lymphoid cells can also respond to signals from the SAM axis that are activated by stress, because they have receptors for catecholamines, adrenaline and noradrenaline. Catecholamines induce changes in cellular trafficking, lymphocyte proliferation and antibody production. They also suppress the synthesis of IL-12 and increase the production of IL-10 (REF. 102). This shifts the tenor of cell-mediated immunity from TH1-type to TH2-type responses, which results in increased production of antibodies. It is hypothesized that this TH2-type orientation could increase the risk of viral, fungal and mycobacterial infections. Because powerful links between the brain and the immune system have evolved, this indicates a purpose and biological advantage for these links. Yet, research on the effects of specific interventions is in its infancy. For example, both the incidence and the severity of shingles (caused by reactivation of varicellazoster virus, VZV) increase markedly with increasing age and in association with a decline in VZV-specific cell-mediated immunity. Instructing elderly individuals to practise T’ai Chi resulted in an increase in their VZVspecific cell-mediated immunity103. This is the first study to show that a behavioural intervention can influence a virus-specific cell-mediated immune response that is important in protection against symptomatic re-infection. Suggestive evidence also comes from hypnosis and conditioning trials. For example, the impact of self-hypnosis, relaxation and guided imagery were evaluated in 21 patients with recurring infection with genital herpes-simplex virus 2 (HSV2). In this small study, after 6 weeks of self-hypnosis using guided imagery, disease recurrence was reduced by almost 50% (REF. 104), which correlated with increased HSV-specific activity of NK cells. A systematic review of the literature105 revealed moderate evidence of efficacy for mind–body interventions such as relaxation, meditation, imagery, biofeedback and hypnosis for the treatment of coronary-artery disease, headaches, insomnia, incontinence and chronic lower-back pain. However, a
meta-analysis of 85 trials revealed only modest evidence that these interventions can reliably alter an immune response106. Conclusions
The data regarding CAM approaches have mainly been obtained using in vitro assays, which have uncertain relevance to in vivo effects, and small in vivo studies in which the immunological effects observed could not be associated with meaningful clinical changes. The exceptions to this general conclusion are the data obtained from studies of individuals with a marked vitamin A or zinc deficiency, remediation of which increases immune responses and improves outcomes. The value of probiotics for the treatment of inflammatory bowel disease and urogenital-tract infections merits additional study, because current treatments are not always satisfactory. Probiotics alone or in combination with other agents might result in a useful therapy. Mind–body therapies that purport to address psychoneuroimmunological circuits are another fertile area of research. Further progress will arise from studies that evaluate the impact of brain–body interactions on concurrent measures of immune parameters, disease progression, resistance to infection, and health and well-being. Moreover, studies of any of the CAM modalities will benefit from the use of newer and more powerful technologies. So far, studies have mainly relied on relatively simple and insensitive methodologies, such as phenotyping of lymphocytes and quantitation of antibody, proliferative or cytokine responses. Studies of brain correlates of behaviour and immunity using functional magnetic-resonance imaging and positron-emission tomography, and microarray analyses of the expression of genes that mediate immune and neuroendocrine responses, could prove sufficiently sensitive to identify more proximate and powerful contributors to health and to provide objective evidence (in the context of well-controlled trials) of the salutary effects of particular CAM or conventional interventions.
1.
2.
3.
4.
5.
6.
7. 8.
9.
10.
11.
12. 13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
Martin H. Goldrosen is at the Office of Scientific Review, National Center for Complementary and Alternative Medicine, National Institutes of Health, Suite 401, 6707 Democracy Boulevard, Bethesda, Maryland 20817, USA. Stephen E. Straus is at the Office of the Director, National Center for Complementary and Alternative Medicine, National Institutes of Health, Building 31, Room 2B11, 31 Center Drive, Bethesda, Maryland 20892-2182, USA. Correspondence to S.E.S. e-mail:
[email protected] doi:10.1038/nri1486
NATURE REVIEWS | IMMUNOLOGY
23. 24.
25. 26. 27.
28.
National Center for Complementary and Alternative Medicine. Expanding Horizons of Healthcare: Five-Year Strategic Plan 2001–2005 (National Institutes of Health, Bethesda, 2000). Ernst, E. Prevalence of use of complementary/alternative medicine: a systematic review. Bull. World Health Organ. 78, 252–257 (2000). Barnes, P. M., Powell-Griner, E., McFann, K. & Nahin, R. L. Complementary and alternative medicine use among adults: United States, 2002. Adv. Data 343, 1–19 (2004). Eisenberg, D. M. et al. Trends in alternative medicine use in the United States, 1990–1997: results of a follow-up national survey. JAMA 280, 1569–1575 (1998). Ni, H., Simile, C. & Hardy, A. M. Utilization of complementary and alternative medicine by United States adults: results from the 1999 National Health Interview Survey. Med. Care 40, 353–358 (2002). Harris Interactive. Widespread ignorance of regulation and labeling of vitamins, minerals and food supplements, according to a National Harris Interactive Survey. Harris Interactive Health Care News (Dec 2002). Harris Interactive [online], http://www.harrisinteractive.com/news/allnewsbydate.as p?NewsID=560 Zeisel, S. H. Regulation of ‘nutraceuticals’. Science 285, 1853–1855 (1999). World Health Organization. Regulatory Situation of Herbal Medicines: A Worldwide Review (World Health Organization, Geneva, 1998). European Parliament and the Council of the European Union. The approximation of the laws of the member states relating to food supplements directive (2002/46/EC). Official J. Eur. Communities L183/51 (10 Jun 2002). Astin, J. A. Why patients use alternative medicine: results of a national study. JAMA 279, 1548–1553 (1998). Nortier, J. L. et al. Urothelial carcinoma associated with the use of a Chinese herb (Aristolochia fangchi). N. Engl. J. Med. 342, 1686–1692 (2000). Teschke, R. Kava, kavapyrones and toxic liver injury. Z. Gastroenterol. 41, 395–404 (2003) (in German). Chen, C., Biller, J., Willing, S. J. & Lopez, A. M. Ischemic stroke after using over the counter products containing ephedra. J. Neurol. Sci. 217, 55–60 (2004). Markowitz, J. S. et al. Effect of St John’s wort on drug metabolism by induction of cytochrome P450 3A4 enzyme. JAMA 290, 1500–1504 (2003). Piscitelli, S. C., Burstein, A. H., Chaitt, D., Alfaro, R. M. & Falloon, J. Indinavir concentrations and St John’s wort. Lancet 355, 547–548 (2000). Mathijssen, R. H., Verweij, J., de Bruijn, P., Loos, W. J. & Sparreboom, A. Effects of St. John’s wort on irinotecan metabolism. J. Natl Cancer Inst. 94, 1247–1249 (2002). Mansky, P. J. & Straus, S. E. St. John’s Wort: more implications for cancer patients. J. Natl Cancer Inst. 94, 1187–1188 (2002). Bauer, S. et al. Alterations in cyclosporin A pharmacokinetics and metabolism during treatment with St John’s wort in renal transplant patients. Br. J. Clin. Pharmacol. 55, 203–211 (2003). Ernst, E. St John’s wort supplements endanger the success of organ transplantation. Arch. Surg. 137, 316–319 (2002). Hall, S. D. et al. The interaction between St John’s wort and an oral contraceptive. Clin. Pharmacol. Ther. 74, 525–535 (2003). Pfrunder, A. et al. Interaction of St John’s wort with low-dose oral contraceptive therapy: a randomized controlled trial. Br. J. Clin. Pharmacol. 56, 683–690 (2003). Nahin, R. L. & Straus, S. E. Research into complementary and alternative medicine: problems and potential. BMJ 322, 161–164 (2001). Hughes, D. A. Dietary carotenoids and human immune function. Nutrition 17, 823–827 (2001). De la Fuente, M. Effects of antioxidants on immune system ageing. Eur. J. Clin. Nutr. 56 (Suppl. 3), S5–S8 (2002). Stephensen, C. B. Vitamin A, infection, and immune function. Annu. Rev. Nutr. 21, 167–192 (2001). Moriguchi, S. & Muraga, M. Vitamin E and immunity. Vitam. Horm. 59, 305–336 (2000). Beharka, A., Redican, S., Leka, L. & Meydani, S. N. Vitamin E status and immune function. Methods Enzymol. 282, 247–263 (1997). Serafini, M. Dietary vitamin E and T cell-mediated function in the elderly: effectiveness and mechanism of action. Int. J. Dev. Neurosci. 18, 401–410 (2000).
VOLUME 4 | NOVEMBER 2004 | 9 1 9 ©2004 Nature Publishing Group
PERSPECTIVES 29. Langweiler, M., Schultz, R. D. & Sheffy, B. E. Effect of vitamin E deficiency on the proliferative response of canine lymphocytes. Am. J. Vet. Res. 42, 1681–1685 (1981). 30. Meeker, H. C., Eskew, M. L., Scheuchenzuber, W., Scholz, R. W. & Zarkower, A. Antioxidant effects on cell-mediated immunity. J. Leukoc. Biol. 38, 451–458 (1985). 31. Bendich, A., Gabriel, E. & Machlin, L. J. Dietary vitamin E requirement for optimum immune responses in the rat. J. Nutr. 116, 675–681 (1986). 32. Moriguchi, S., Kobayashi, N. & Kishino, Y. High dietary intakes of vitamin E and cellular immune functions in rats. J. Nutr. 120, 1096–1102 (1990). 33. Yasunaga, T., Kato, H., Ohgaki, K., Inamoto, T. & Hikasa, Y. Effect of vitamin E as an immunopotentiation agent for mice at optimal dosage and its toxicity at high dosage. J. Nutr. 112, 1075–1084 (1982). 34. Tengerdy, R. P., Henzerling, R. H., Brown, G. L. & Mathias, M. M. Enhancement of the humoral immune response by vitamin E. Int. Arch. Allergy Appl. Immunol. 44, 221–232 (1973). 35. High, K. P. Nutritional strategies to boost immunity and prevent infection in elderly individuals. Clin. Infect. Dis. 33, 1892–1900 (2001). 36. Meydani, S. N. et al. Vitamin E supplementation and in vivo immune response in healthy elderly subjects. A randomized controlled trial. JAMA 277, 1380–1386 (1997). 37. Meydani, S. N. et al. Vitamin E and respiratory tract infections in elderly nursing home residents: a randomized controlled trial. JAMA 292, 828–836 (2004). 38. Martorell, R., Yarbrough,C., Yarbrough, S. & Klein, R. E. The impact of ordinary illnesses on the dietary intakes of malnourished children. Am. J. Clin. Nutr. 33, 345–350 (1980). 39. Solomons, N. W. Pathways to the impairment of human nutritional status by gastrointestinal pathogens. Parasitology 107, S19–S35 (1993). 40. Stephensen, C. B. et al. Vitamin A is excreted in the urine during acute infection. Am. J. Clin. Nutr. 60, 388–392 (1994). 41. Wiedermann, U. et al. Vitamin A deficiency predisposes to Staphylococcus aureus infection. Infect. Immun. 64, 209–214 (1996). 42. Dawson, H. D., Li, N. Q., DeCicco, K. L., Nibert, J. A. & Ross, A. C. Chronic marginal vitamin A status reduces natural killer cell number and function in aging Lewis rats. J. Nutr. 129, 1510–1517 (1999). 43. Carman, J. A., Pond, L., Nashold, F., Wassom, D. L. & Hayes, C. E. Immunity to Trichinella spiralis infection in vitamin A-deficient mice. J. Exp. Med. 175, 111–120 (1992). 44. Glasziou, P. P. & Mackerras, D. E. Vitamin A supplementation in infectious diseases: a meta-analysis. BMJ 306, 366–370 (1993). 45. Fawzi, W. W., Chalmers, T. C., Herrera, M. G. & Mosteller, F. Vitamin A supplementation and child mortality. A meta-analysis. JAMA 269, 898–903 (1993). 46. American Academy of Pediatrics Committee on Infectious Diseases. Vitamin A treatment of measles. Pediatrics 91, 1014–1015 (1993). 47. Stephensen, C. B. et al. Adverse effects of high-dose vitamin A supplements in children hospitalized with pneumonia. Pediatrics [online], 101, E3 (1998). 48. Aukrust, P. et al. Decreased vitamin A levels in common variable immunodeficiency: vitamin A supplementation in vivo enhances immunoglobulin production and downregulates inflammatory responses. Eur. J. Clin. Invest. 30, 252–259 (2000). 49. Semba, R. D. Vitamin A and human immunodeficiency virus infection. Proc. Nutr. Soc. 56, 459–469 (1997). 50. Coutsoudis, A., Pillay, K., Spooner, E., Kuhn, L. & Coovadia, H. M. Randomized trial testing the effect of vitamin A supplementation on pregnancy outcomes and early mother-to-child HIV-1 transmission in Durban, South Africa. South African Vitamin A Study Group. AIDS 13, 1517–1524 (1999). 51. Fawzi, W. W. et al. Effect of providing vitamin supplements to human immunodeficiency virus-infected, lactating mothers on the child’s morbidity and CD4+ cell counts. Clin. Infect. Dis. 36, 1053–1062 (2003). 52. Barnes, P. M. & Moynahan, E. J. Zinc deficiency in acrodermatitis enteropathica: multiple dietary intolerance treated with synthetic diet. Proc. R. Soc. Med. 66, 327–329 (1973). 53. Zinc therapy of depressed cellular immunity in acrodermatitis enteropathica. Nutr. Rev. 39, 168 (1981).
920
54. Beck, F. W., Prasad, A. S., Kaplan, J., Fitzgerald, J. T. & Brewer, G. J. Changes in cytokine production and T cell subpopulations in experimentally induced zinc-deficient humans. Am. J. Physiol. 272, E1002–E1007 (1997). 55. Roy, S. K. et al. Randomised controlled trial of zinc supplementation in malnourished Bangladeshi children with acute diarrhoea. Arch. Dis. Child. 77, 196–200 (1997). 56. Sazawal, S. et al. Zinc supplementation in young children with acute diarrhea in India. N. Engl. J. Med. 333, 839–844 (1995). 57. Turner, R. B. & Cetnarowski, W. E. Effect of treatment with zinc gluconate or zinc acetate on experimental and natural colds. Clin. Infect. Dis. 31, 1202–1208 (2000). 58. Prasad, A. S. et al. Zinc deficiency in elderly patients. Nutrition 9, 218–224 (1993). 59. Girodon, F. et al. Effect of micronutrient supplementation on infection in institutionalized elderly subjects: a controlled trial. Ann. Nutr. Metab. 41, 98–107 (1997). 60. Chandra, R. K. Excessive intake of zinc impairs immune responses. JAMA 252, 1443–1446 (1984). 61. Failla, M. L. & Hopkins, R. G. Is low copper status immunosuppressive? Nutr. Rev. 56, S59–S64 (1998). 62. Gibney, M. J. & Hunter, B. The effects of short- and longterm supplementation with fish oil on the incorporation of n-3 polyunsaturated fatty acids into cells of the immune system in healthy volunteers. Eur. J. Clin. Nutr. 47, 255–259 (1993). 63. Yaqoob, P., Pala, H. S., Cortina-Borja, M., Newsholme, E. A. & Calder, P. C. Encapsulated fish oil enriched in α-tocopherol alters plasma phospholipid and mononuclear cell fatty acid compositions but not mononuclear cell functions. Eur. J. Clin. Invest. 30, 260–274 (2000). 64. Blok, W. L. et al. Pro- and anti-inflammatory cytokines in healthy volunteers fed various doses of fish oil for 1 year. Eur. J. Clin. Invest. 27, 1003–1008 (1997). 65. Meydani, S. N. et al. Oral (n-3) fatty acid supplementation suppresses cytokine production and lymphocyte proliferation: comparison between young and older women. J. Nutr. 121, 547–555 (1991). 66. Kramer, T. R. et al. Increased vitamin E intake restores fish-oil-induced suppressed blastogenesis of mitogenstimulated T lymphocytes. Am. J. Clin. Nutr. 54, 896–902 (1991). 67. Wang, C. et al. Effects of Omega-3 Fatty Acids on Cardiovascular Disease Summary, Evidence report/technology assessment 94 (Agency for Healthcare Research and Quality, Rockville, 2004). 68. Marchioli, R. et al. Early protection against sudden death by n-3 polyunsaturated fatty acids after myocardial infarction: time-course analysis of the results of the Gruppo Italiano per lo Studio della Sopravvivenza nell’Infarto Miocardico (GISSI)-Prevenzione. Circulation 105, 1897–1903 (2002). 69. Volker, D., Fitzgerald, P., Major, G. & Garg, M. Efficacy of fish oil concentrate in the treatment of rheumatoid arthritis. J. Rheumatol. 27, 2343–2346 (2000). 70. Belluzzi, A. et al. Effect of an enteric-coated fish-oil preparation on relapses in Crohn’s disease. N. Engl. J. Med. 334, 1557–1560 (1996). 71. Broughton, K. S., Johnson, C. S., Pace, B. K., Liebman, M. & Kleppinger, K. M. Reduced asthma symptoms with n-3 fatty acid ingestion are related to 5-series leukotriene production. Am. J. Clin. Nutr. 65, 1011–1017 (1997). 72. Villani, F., Comazzi, R., De Maria, P. & Galimberti, M. Effect of dietary supplementation with polyunsaturated fatty acids on bronchial hyperreactivity in subjects with seasonal asthma. Respiration 65, 265–269 (1998). 73. Donadio, J. V. Jr, Bergstralh, E. J., Offord, K. P., Spencer, D. C. & Holley, K. E. A controlled trial of fish oil in IgA nephropathy. Mayo Nephrology Collaborative Group. N. Engl. J. Med. 331, 1194–1199 (1994). 74. Newman, D. J., Cragg, G. M. & Snader, K. M. Natural products as sources of new drugs over the period 1981–2002. J. Nat. Prod. 66, 1022–1037 (2003). 75. Borchers, A. T., Keen, C. L., Stern, J. S. & Gershwin, M. E. Inflammation and Native American medicine: the role of botanicals. Am. J. Clin. Nutr. 72, 339–347 (2000). 76. Perry, N. B., Burgess, E. J. & Glennie, V. L. Echinacea standardization: analytical methods for phenolic compounds and typical levels in medicinal species. J. Agric. Food Chem. 49, 1702–1706 (2001). 77. Burger, R. A., Torres, A. R., Warren, R. P., Caldwell, V. D. & Hughes, B. G. Echinacea-induced cytokine production by human macrophages. Int. J. Immunopharmacol. 19, 371–379 (1997).
| NOVEMBER 2004 | VOLUME 4
78. Berman, J. D. & Straus, S. E. in Principles and Practices of Infectious Diseases (eds Mandell, G., Bennett, J. E. & Dolin, R.) (Elsevier, Philadelphia, in the press). 79. Schulten, B., Bulitta, M., Ballering-Bruhl, B., Koster, U. & Schafer, M. Efficacy of Echinacea purpurea in patients with a common cold. A placebo-controlled, randomised, double-blind clinical trial. Arzneimittelforschung 51, 563–568 (2001). 80. Barrett, B. P. et al. Treatment of the common cold with unrefined Echinacea. A randomized, double-blind, placebo-controlled trial. Ann. Intern. Med. 137, 939–946 (2002). 81. Tao, X. & Lipsky, P. E. The Chinese anti-inflammatory and immunosuppressive herbal remedy Tripterygium wilfordii Hook F. Rheum. Dis. Clin. North Am. 26, 29–50 (2000). 82. Chen, B. J. Triptolide, a novel immunosuppressive and anti-inflammatory agent purified from a Chinese herb Tripterygium wilfordii Hook F. Leuk. Lymphoma 42, 253–265 (2001). 83. Tao, X., Younger, J., Fan, F. Z., Wang, B. & Lipsky, P. E. Benefit of an extract of Tripterygium wilfordii Hook F. in patients with rheumatoid arthritis: a double-blind, placebo-controlled study. Arthritis Rheum. 46, 1735–1743 (2002). 84. Alvarez-Olmos, M. I. & Oberhelman, R. A. Probiotic agents and infectious diseases: a modern perspective on a traditional therapy. Clin. Infect. Dis. 32, 1567–1576 (2001). 85. Silva, M., Jacobus, N. V., Deneke, C. & Gorbach, S. L. Antimicrobial substance from a human Lactobacillus strain. Antimicrob. Agents Chemother. 31, 1231–1233 (1987). 86. Mack, D. R., Michail, S., Wei, S., McDougall, L. & Hollingsworth, M. A. Probiotics inhibit enteropathogenic E. coli adherence in vitro by inducing intestinal mucin gene expression. Am. J. Physiol. 276, G941–G950 (1999). 87. Christensen, H. R., Frokiaer, H. & Pestka, J. J. Lactobacilli differentially modulate expression of cytokines and maturation surface markers in murine dendritic cells. J. Immunol. 168, 171–178 (2002). 88. Gill, H. S., Rutherfurd, K. J. & Cross, M. L. Dietary probiotic supplementation enhances natural killer cell activity in the elderly: an investigation of age-related immunological changes. J. Clin. Immunol. 21, 264–271 (2001). 89. Kaila, M. et al. Enhancement of the circulating antibody secreting cell response in human diarrhea by a human Lactobacillus strain. Pediatr. Res. 32, 141–144 (1992). 90. Guarino, A., Canani, R. B., Spagnuolo, M. I., Albano, F. & Di Benedetto, L. Oral bacterial therapy reduces the duration of symptoms and of viral excretion in children with mild diarrhea. J. Pediatr. Gastroenterol. Nutr. 25, 516–519 (1997). 91. Van Niel, C. W., Feudtner, C., Garrison, M. M. & Christakis, D. A. Lactobacillus therapy for acute infectious diarrhea in children: a meta-analysis. Pediatrics 109, 678–684 (2002). 92. Kalliomaki, M. et al. Probiotics in primary prevention of atopic disease: a randomised placebo-controlled trial. Lancet 357, 1076–1079 (2001). 93. Buydens, P. & Debeuckelaere, S. Efficacy of SF 68 in the treatment of acute diarrhea. A placebo-controlled trial. Scand. J. Gastroenterol. 31, 887–891 (1996). 94. Oksanen, P. J. et al. Prevention of travellers’ diarrhoea by Lactobacillus GG. Ann. Med. 22, 53–56 (1990). 95. Fedorak, R. N. & Madsen, K. L. Probiotics and the management of inflammatory bowel disease. Inflamm. Bowel Dis. 10, 286–299 (2004). 96. Reid, G. et al. Oral use of Lactobacillus rhamnosus GR-1 and L. fermentum RC-14 significantly alters vaginal flora: randomized, placebo-controlled trial in 64 healthy women. FEMS Immunol. Med. Microbiol. 20, 131–134 (2003). 97. Evans, P., Hucklebridge, F. & Chow, A. Mind, Immunity and Health: The Science of Psychoneuroimmunology (Free Association Books, London, 2000). 98. Kiecolt-Glaser, J. K., McGuire, L., Robles, T. F. & Glaser, R. Psychoneuroimmunology and psychosomatic medicine: back to the future. Psychosom. Med. 64, 15–28 (2002). 99. Cohen, S., Tyrrell, D. A. & Smith, A. P. Psychological stress and susceptibility to the common cold. N. Engl. J. Med. 325, 606–612 (1991). 100. Yang, E. V. & Glaser, R. Stress-associated immunomodulation and its implications for responses to vaccination. Expert Rev. Vaccines 1, 453–459 (2002). 101. Eskandari, F. & Sternberg, E. M. Neural–immune interactions in health and disease. Ann. NY Acad. Sci. 966, 20–27 (2002).
www.nature.com/reviews/immunol ©2004 Nature Publishing Group
PERSPECTIVES 102. Elenkov, I. J., Papanicolaou, D. A., Wilder, R. L. & Chrousos, G. P. Modulatory effects of glucocorticoids and catecholamines on human interleukin-12 and interleukin-10 production: clinical implications. Proc. Assoc. Am. Physicians 108, 374–381 (1996). 103. Irwin, M. R., Pike, J. L., Cole, J. C. & Oxman, M. N. Effects of a behavioral intervention, Tai Chi Chih, on varicella-zoster virus specific immunity and health functioning in older adults. Psychosom. Med. 65, 824–830 (2003). 104. Gruzelier, J. H. A review of the impact of hypnosis, relaxation, guided imagery and individual differences on aspects of immunity and health. Stress 5, 147–163 (2002). 105. Astin, J. A., Shapiro, S. L., Eisenberg, D. M. & Forys, K. L. Mind–body medicine: state of the science, implications for practice. J. Am. Board Fam. Pract. 16, 131–147 (2003). 106. Miller, G. E. & Cohen, S. Psychological interventions and the immune system: a meta-analytic review and critique. Health Psychol. 20, 47–63 (2001). 107. Anderson, J. W., Konz, E. C. & Jenkins, D. J. Health advantages and disadvantages of weight-reducing diets: a computer analysis and critical review. J. Am. Coll. Nutr. 19, 578–590 (2000).
108. Foster, G. D. et al. A randomized trial of a low-carbohydrate diet for obesity. N. Engl. J. Med. 348, 2082–2090 (2003). 109. Ornish, D. Dr. Dean Ornish’s Program for Reversing Heart Disease (Ballantine Books, New York, 2004). 110. Sheng, Y., Bryngelsson, C. & Pero, R. W. Enhanced DNA repair, immune function and reduced toxicity of C-MED100, a novel aqueous extract from Uncaria tomentosa. J. Ethnopharmacol. 69, 115–126 (2000). 111. Chan, M. M. Inhibition of tumor necrosis factor by curcumin, a phytochemical. Biochem. Pharmacol. 49, 1551–1556 (1995). 112. Nakaya, T. A., Kita, M., Kuriyama, H., Iwakura, Y. & Imanishi, J. Panax ginseng induces production of proinflammatory cytokines via Toll-like receptor. J. Interferon Cytokine Res. 24, 93–100 (2004). 113. McKallip, R. J., Lombard, C., Martin, B. R., Nagarkatti, M. & Nagarkatti, P. S. ∆9-tetrahydrocannabinol-induced apoptosis in the thymus and spleen as a mechanism of immunosuppression in vitro and in vivo. J. Pharmacol. Exp. Ther. 302, 451–465 (2002). 114. United States Congress. Dietary Supplement Health and Education Act (Public Law 103-417; codified at US congress 287C-11) (25 Oct 1994). Food and Drug Administration [online], http://www.fda.gov/opacom/laws/dshea.html
Acknowledgements We acknowledge the assistance of C. Pontzer for proof-reading the manuscript.
Competing interests statement The authors declare no competing financial interests.
Online links DATABASES The following terms in this article are linked online to: Entrez Gene: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=gene C-reactive protein | IFN-γ | IL-1 | IL-2 | IL-4 | IL-6 | IL-8 | IL-10 | IL-12 | TNF FURTHER INFORMATION National Center for Complementary and Alternative Medicine: http://nccam.nih.gov Chronic Fatigue and Immune Dysfunction Syndrome Association of America: http://www.cfids.org The Autism Autoimmunity Project: http://www.autismautoimmunityproject.org Access to this interactive links box is free online.
ONLINE CORRESPONDENCE
Nature Reviews Immunology publishes items of correspondence online. Such contributions are published at the discretion of the Editors and can be subject to peer review. Correspondence should be no longer than 500 words with up to 15 references and should represent a scholarly attempt to comment on a specific Review or Perspective article that has been published in the journal. To view correspondence, please go to our homepage and select the link to Correspondence or use the URL indicated below. The following correspondence has recently been published:
Processed switch-region transcripts and retrotranscripts as possible generators of heteroduplex ‘R-loop’ targets for AID-deamination Steele, E. J. http://www.nature.com/nri/archive/correspondence.html This correspondence relates to the article
CLASS-SWITCH RECOMBINATION: INTERPLAY OF TRANSCRIPTION, DNA DEAMINATION AND DNA REPAIR Chaudhuri, J. and Alt, F. W. Nature Reviews Immunology 4, 541–552 (2004).
César Milstein and the discovery of monoclonal antibodies Karpas, A. http://www.nature.com/nri/archive/correspondence.html
A witness is not a judge Alkan, S. S. http://www.nature.com/nri/archive/correspondence.html This correspondence relates to the article
MONOCLONAL ANTIBODIES: THE STORY OF A DISCOVERY THAT REVOLUTIONIZED SCIENCE AND MEDICINE Alkan, S. S. Nature Reviews Immunology 4, 153–156 (2004).
NATURE REVIEWS | IMMUNOLOGY
VOLUME 4 | NOVEMBER 2004 | 9 2 1 ©2004 Nature Publishing Group
