NATURE IMMUNOLOGY VOLUME 6 NUMBER 4 APRIL 2005 353 Infectious challenges to the host are met by a wealth of humoral and cellular responses. Some agents are difficult to control and the host response to them often results in tissue damage. This tissue damage might be more intense were it not for many regulatory mechanisms that contain the ���zeal��� of both innate and adaptive effector responses. The once-disfavored idea that suppressor cells with antigen specific- ity form part of the regulatory mechanisms has now been revital- ized1. Indeed, it is has been conceded that several types of regulatory cells exist, some of which are induced in response to infectious chal- lenge and some that are considered natural regulators2,3. Inducible regulatory T cells (Treg cells) such as TR1 or T helper type 3 (TH3) cells can develop from conventional CD4+ T cells that are exposed to specific stimulatory conditions such as the blockade of costimulatory signals, deactivating cytokines or drugs. These cell types have been discussed in several reviews2,4,5. Natural Treg cells, however, arise dur- ing the normal process of maturation in the thymus and survive in the periphery as Treg cells. This segregation between natural Treg cells and induced Treg cells could prove to be arbitrary, with the relation- ship between the populations requiring clarification. Nevertheless, natural Treg cells obey defined rules and express a specific set of markers3,6. For example, only natural Treg cells constitutively express CD25, the T cell inhibitory receptor CTLA-4 and the glucocorti- coid-inducible tumor necrosis factor receptor (GITR). The unique transcription factor Foxp3 is required for the generation of natural Treg cells, and this represents their most specific marker identified so far (reviewed by Fontenot and Rudensky6 in this issue). Natural Treg cells can respond to a large variety of self antigens, although grow- ing evidence suggests that these cells may also respond to antigens expressed by microbes. Although inducible Treg cells may control various infectious processes4, our review focuses only on infections for which an association with natural Treg cells has been suggested (Table 1). Understanding the unique properties of natural Treg cells and their mode of action may result in new therapeutic avenues use- ful for the control of infectious diseases. In most cases in which natural Treg cells participate in responses to infection, these are chronic infections. As discussed below, the influ- ence of natural Treg cells may favorably affect the outcome or can be harmful to the host. However, the outcome is also affected by other factors. These include the stage of infection, dose of the pathogen and genotype and immunological status of the host as well as the presence of concomitant disease or other infections. We also discuss whether enhanced pathogen survival is one consequence of natural Treg cell function. Influence of regulatory: effector cell balance Some of the earliest studies of natural Treg cells emphasized that such cells help control the extent of immune-mediated pathology. In fact, a chief function of natural Treg cells may be to respond to signals associated with tissue destruction and then to minimize col- lateral tissue damage they cause7. A well documented example of this situation is the involvement of natural Treg cells in gastrointestinal homeostasis. Commensal gut bacteria can, in cases of immune dys- regulation, trigger harmful inflammatory diseases. Extensive work in mouse models of colitis has demonstrated that natural Treg cells act as chief regulators of such lesions. Adoptive transfer of naive T cell populations lacking natural Treg cells into T cell���deficient mice causes massive gut inflammation. Transfer of CD4+CD25+CD45RBlo T cells together with those naive T cells suppresses disease devel- opment, an effect mediated by interleukin-10 (IL-10), transforming growth factor-�� (TGF-��) and CTLA-4. Similarly, in mice deficient in the recombination-activating gene(s), Helicobacter hepaticus causes 1Division of Molecular Immunology, Cincinnati Children���s Hospital Research Foundation, Cincinnati, Ohio 45229, USA. 2Department of Microbiology, Walters Life Sciences Building, University of Tennessee Knoxville, Tennessee, 37996, USA. 3Present address: Laboratory of Parasitic Diseases, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland 20892, USA. Correspondence should be addressed to Y.B. (ybelkaid@niaid.nih.gov). Published online 22 March 2005 doi:10.1038/ni1181 Natural regulatory T cells in infectious disease Yasmine Belkaid1,3 & Barry T Rouse2 This review discusses the control exerted by natural CD4+ CD25+ regulatory T cells (natural Treg cells) during infectious processes. Natural Treg cells may limit the magnitude of effector responses, which may result in failure to adequately control infection. However, natural Treg cells also help limit collateral tissue damage caused by vigorous antimicrobial immune responses. We describe here various situations in which the balance between natural Treg cells and effector immune functions influences the outcome of infection and discuss how manipulating this equilibrium might be exploited therapeutically. R E V I E W REGULATORY T CELLS �� 2005 Nature Publishing Group http://www.nature.com/natureimmunology

354 VOLUME 6 NUMBER 4 APRIL 2005 NATURE IMMUNOLOGY mild intestinal inflammation, but this is enhanced considerably by the adoptive transfer of CD4+CD25���CD45RBhi T cells or T cells from IL- 10-deficient mice8,9. However, transfer of CD4+CD45RBloCD25��� or CD25+ cells together with those cells prevents inflammation in an IL-10- and TGF-��-dependent way10. This model was the first, to our knowledge, to demonstrate that the targets of natural Treg cells could include components of the innate immune system as well as patho- genic T cells9. In T cell���deficient mice infected with Helicobacter pylori, transfer of CD4+CD25��� T cells provides better control of the infection, but enhanced gastric inflammation also occurs11. Natural Treg cells may also control H. pylori infection in humans12. When isolated from a chronically infected person, Treg cells are able to suppress H. pylori���specific T cell responses but not responses to unrelated antigens12. The preservation of host homeostasis by natural Treg cells is not restricted to inflammation caused by gastrointestinal bacteria. There is similar involvement in other infection models in which the infected sites require more control. These include the lung, skin and liver as well as the eye. For example, Pneumocystis carinii infects mice defi- cient in recombination-activating gene 2 without inducing detect- able pathology. The transfer of CD4+CD25��� T cells results in better control of infection but also triggers florid pneumonitis that becomes lethal. This outcome can be prevented by transfer of natural Treg cells13. During infection of mice with Candida albicans, a reduction in natural Treg cell numbers induces better control of the infection, but unfortunately enhanced inflammatory gastrointestinal pathology also occurs14. In a nonhealing model of Leishmania major infection, cutaneous infection results in progressive lesions caused by a robust TH2 response15. In this model, the amplitude of the response and subsequent pathology is held in check by natural Treg cells16,17. This model exemplifies the dual function of natural Treg cells during a given infection. The early disease exacerbation in the absence of natural Treg cells is followed by better control of infection in the later stages of the syndrome than in control mice18. The outcome of schistosomal infection in mice depends on TH2 polarization. The inhibitory effects of natural Treg cells on the TH1 response have been shown to promote TH2 polarization and to pro- tect the host from lethal inflammatory pathology19. One of the main targets of natural Treg cells in this model is the production of IL-12 by activated dendritic cells, an inhibitory effect that is mediated by IL-10. In the late stages of the schistosomal infection, most of the pathology is a granulomatous fibrosis. IL-10-producing natural Treg cells puri- fied from parasite egg���induced granulomas are an important factor for host survival20. Natural Treg cells also seem to be as important in the disease caused by hepatitis C virus (HCV). A chief complication of this chronic infection is massive liver damage that often requires organ transplant. Liver biopsies obtained at the time of the transplant show an inverse correlation between the number of natural Treg cells in the periphery and the histological inflammatory score21. Control of inflammatory reaction by natural Treg cells might be especially important in delicate tissues such as the eye. This organ���s function requires that the path of light to the retina not be impeded by defracting inflammatory cells. In a model in which a blinding keratitis was caused by herpes virus infection induced by a CD4+ T cell���orchestrated reaction, lesions were much more severe in mice whose natural Treg cells had been depleted22. Indeed in the absence of natural Treg cells, nonpathological doses of virus can readily induce keratitis22. One other consequence of the modulation of excessive immune responses by natural Treg cells is enhanced pathogen survival and, in some cases, long-term persistence. Thus, pathogen persistence may represent a ���com- promise��� reached by the host with pathogens when new homeostatic conditions become established (Table 2). The mouse model of L. major infection in which natural Treg cells are a necessary component of the pathogen���s survival provides a good example of this23. Self-healing C57BL/6 mice infected with a low dose of parasites develop small self- healing lesions, and immunity to re-infection requires persistent infection24. Natural Treg cells accumulate at sites of infection and limit the efficacy of TH1 immune responses (by both IL-10-dependent and IL-10-indepen- dent pathways). As a consequence, the natu- ral Treg cells promote pathogen persistence and potential transmission to other hosts. Removal of natural Treg cells leads to ���sterile cure���, a state that is not compatible with the preservation of long-term immunity23. Another example of this ���entente��� between host and pathogen has been provided by ocular infection of mice with herpes simplex virus (HSV). A low dose of virus infection protects mice from CD4+ T cell���mediated pathology by promoting natural Treg cells, a situation that is compatible with the estab- lishment of immunity to re-infection22. Table 1 Microbial infections for which a regulatory function for natural Treg cells has been suggested Microbe Species Antigen specificity Reference Helicobacter hepaticus Mouse ND 8,10 Human Helicobacter pylori Mouse ND 12,49 Human Listeria monocytogenes Mouse ND 38 Pneumocistis carinii Mouse ND 13 Leishmania major Mouse Yes 16���18,23,37 Schistosoma masoni Mouse Yes 19,20 Candida albicans Mouse ND 14,80 Herpes simplex virus Mouse ND 22,28 Friend virus Mouse ND 30,75 Human immunodeficiency virus Human Yes 32���34,59,60 Hepatitis C virus Human Yes 21,35,36 Cytomegalovirus Human ND 32 Murine AIDS Mouse ND 31 Feline immunodeficiency virus Cat ND 56,81 ND, not done. REVIEW �� 2005 Nature Publishing Group http://www.nature.com/natureimmunology