Advances in the Study of CD8⁺ Regulatory T Cells

Abstract

Immune tolerance mediated by CD4⁺ and CD8⁺ regulatory T (Treg) cells is important in the control of inflammatory and autoimmune diseases. While CD4⁺FoxP3⁺ Treg cells are well studied, there remained several critical gaps in our current knowledge of the biology of CD8⁺ Treg cells. A major caveat was our inability to distinguish them from conventional CD8⁺ T cells. In this regard, we have recently discovered an innate-like PLZF⁺CD8α⁺TCRαβ⁺ Treg population (CD8αα Treg cells) which is enriched in the liver in naïve mice and present in healthy humans. We have demonstrated that these CD8αα Treg cells serve as a feedback regulatory mechanism and target only activated effector T cells. Such feedback regulation allows the progression of an immune defense response, while avoiding excessive tissue damage. It is likely that the PLZF transcription program endows the CD8αα Treg cells with the innate features that are important for them to effectively control autoimmune responses by targeting activated T cells in both mice and humans. Additional features of the CD8αα Treg cells include the dependence on IL-15/IL-2Rβ signaling, the expression of NK-inhibitory receptors, and the memory phenotype. Importantly, these cells are expanded following an ongoing immune response and serve as a feedback regulatory mechanism to control activated effector T cells and, hence, prevent an excessive immune stimulation. In this review, we will briefly summarize recent important findings related to CD8⁺ Treg cells.

INTRODUCTION

The adaptive immune system, mainly consisting of different subsets of T and B cells, is composed of central and peripheral lymphoid tissues. Central lymphoid tissues include the thymus and bone marrow where most lymphocytes develop as well as get educated not to react to self-antigens but retain the ability to mount a vigorous and an effective immune response against foreign antigens derived from microbes as well as tumors. Therefore, self-reactive T cells that respond strongly to self and are able to harm self-tissues are generally deleted by the central thymic mechanism of negative selection. Consequently, fully developed and mature lymphocytes, generally devoid of pathogenic autoreactive lymphocytes populate peripheral lymphoid tissues. Thus, the adaptive immune system acquires the ability to spare one’s own tissues while attacking and eliminating any invading agents, including cancer cells. However, recent molecular studies in both mice and in humans have shown that the negative selection in the thymus is not complete and that potentially pathogenic autoreactive lymphocytes are present in the peripheral lymphoid tissues^{1, 2}. In some cases, these potentially autoreactive lymphocyte are present in healthy individuals at similar frequencies to the microbial antigen-reactive T cells³.

The presence of self-reactive T cells in the periphery indicates that tolerance mechanisms must be in place to prevent these potentially pathogenic autoreactive lymphocytes from causing tissue damage in healthy individuals. In this regard, previous studies have uncovered multiple such tolerance mechanisms including intrinsic (e.g., PD-1, anergy-induction, and exhaustion) and extrinsic (e.g., regulatory T [Treg] cells) mechanisms⁴.

So far, the most extensively studied Treg cell populations both in mice and humans are the CD4⁺FoxP3⁺ Treg cells^5–12. The extensive characterization of the phenotype as well as the role of CD4⁺FoxP3⁺ Treg cells in experimental models have led to several ongoing clinical studies. One of the challenges in these studies has been the use of polyclonal CD4⁺FoxP3⁺ Treg cells. The use of polyclonal but not monoclonal CD4⁺FoxP3⁺ Treg cells is because of the fact that the antigens that CD4⁺FoxP3⁺ Treg cells recognize remain largely unknown^13–18. Furthermore, it is also becoming clear that FoxP3⁺CD4⁺ Treg cells, especially in humans, are quite heterogeneous and not all of them are suppressive¹⁹. In some instances, FoxP3⁺CD4⁺ Treg cells can revert back to inflammatory T cells capable of causing disease^{20, 21}. Thus, so far, clinical trials using polyclonal CD4⁺FoxP3⁺ Treg cells for the treatment of immune-mediated diseases have met with limited success^{13–18, 22–31}. For this reason, it is crucial that exploration of other Treg cells, specifically a detailed characterization of CD8⁺ Treg cells in both experimental settings as well as humans is an important area of investigation.

Consistent with the role of CD8⁺ Treg cells, recent studies have also shown that most CD8⁺ T cells infiltrating the tissues where autoimmune diseases occur were not pathogenic but actually played a regulatory role^{4, 32}. In this regard, it is important to emphasize that CD8⁺ Treg cells were first described in 1970s³³. However, when compared to CD4⁺FoxP3⁺ Treg cells, the study of CD8⁺ Treg cells significantly lags behind because of the inability to differentiate CD8⁺ Treg cells from non-regulatory conventional CD8⁺ T cells³⁴. Despite the difficulty, several studies have now clearly demonstrated that CD8⁺ Treg cells are an important arm of the immune regulation^{32, 35–41}. In the following, we will briefly review the history, phenotype, genetic control, and mechanisms of regulation of CD8⁺ Treg cells. At the end, we will provide our perspective for potential future studies related to the CD8⁺ Treg cells.

EARLY INVESTIGATIONS

In the early 1970’s, Gershon and colleagues demonstrated that T cells from animals tolerant to antigen A, when adoptively transferred, could specifically suppress antibody responses to antigen A in recipient animals³³. Later studies demonstrated that such T suppressor cells could also down-regulate type I hypersensitivity and cell-mediated delayed type hypersensitivity (DTH) reactions^{42, 43}. In addition to inducing T suppressor cells by means of tolerance-inducing protocols, the in vitro generation of T suppressor cells was also reported⁴⁴. At that time, Lyt allo-antisera became available that were used to differentiate between helper T cells (Lyt1⁺ or CD4⁺) and cytotoxic T cells (Lyt2⁺ or CD8⁺). It was demonstrated that the suppressor activity resided within the CD8⁺ T cell subset⁴⁵. When analyzed in vitro, CD8⁺ T suppressor cells appeared to be induced by activated CD4⁺ T cells^{46, 47} and provided a negative feedback regulation, meaning that the CD4⁺ inducer T cells activated CD8⁺ T suppressor cells to mediate their own down-regulation.

CD8⁺ Treg cells were also investigated in animals that were induced for experimental autoimmune encephalomyelitis (EAE) disease in rodents, a model of human multiple sclerosis (MS). In this system, EAE was induced either by active immunization with a myelin self-antigen, i.e. myelin basic protein (MBP), or by passive transfer of Th1 CD4⁺ T cells reactive to MBP. It was found that normal healthy animals, when vaccinated with attenuated pathogenic CD4⁺ T cells reactive to MBP (a process called T cell vaccination or TCV), became resistant to the EAE induction^48–50. To investigate the mechanisms of TCV, Gaur et al. showed that CD8⁺ T cells were required for T cell receptor peptide-induced clonal unresponsiveness⁵¹. In addition, we demonstrated that T cell receptor peptide-specific CD4⁺ Treg cells did not work when CD8⁺ T cells were depleted⁵². In some EAE models, following the disease induction, the animals contracted paralytic disease, recovered spontaneously, and became resistant to re-induction of EAE. To demonstrate the mechanisms underlying the resistance to EAE re-induction, it was shown that mice depleted of CD8⁺ T cells prior to the EAE induction were no longer resistant to re-induction of EAE⁵³ and that CD8^−/− PL/J H-2^u mice displayed a more chronic form of EAE reflected by a higher frequency of relapses⁵⁴. Therefore, previous data have demonstrated that CD8⁺ Treg cells are an important mechanism in the control of immune responses to self-antigens.

In another system, Nanda et al. showed that Vβ^a mice that lacked 10 TCR Vβ gene segments responded to a peptide of hen egg-white lysozyme (HEL). However, in wild-type Vβ^b mice that carried 20 TCR Vβ gene segments, the response to the same peptide could not be detected. However, the peptide-specific T cell responsiveness was revealed in wild-type (Vβ^b) mice when these mice were treated in vivo with anti-CD8 antibody to deplete CD8⁺ T cells⁵⁵. These findings suggest that CD8⁺ Treg cells also regulate immune responses to foreign antigens.

In conclusion, it is clear that CD8 Treg cells play an important role in regulating immune responses in a variety of settings. In the following sections, we will review our current knowledge regarding CD8 Treg cells in the context of cell surface markers, MHC restriction, genetic control, and mechanisms of regulation. At the end, we will provide our perspective for potential future studies.

CELL SURFACE MARKERS

The identification of unique cell surface markers and or specific transcription factors in CD8⁺ Treg is critical for characterizing their functions, as has been exemplified by the study of FoxP3⁺CD4⁺ Treg cells. Despite that CD4⁺ Treg cells were discovered at least a decade late than CD8⁺ Treg cells^{5, 33}, they have been dominating the study of immune regulation because of their easy-to-identify cell surface markers, i.e. CD4⁺CD25⁺CD127^low56 and the expression of FoxP3. For this reason, our laboratories and others have focused on identifying unique cell surface markers and/or transcription factors that can be used to identify CD8⁺ Treg cells and differentiating them from the conventional CD8⁺ T cells.

Firstly, using cloned CD8⁺ Treg cells from TCR peptide-immunized mice, we showed that CD8⁺ Treg cells expressed CD122⁵⁷. In this regard, CD122 is the IL-2 and IL-15 receptor β chain. An early study showed that CD122 deficient mice displayed spontaneous activation of T cells, leading to enhanced granulopoiesis, suppressed erythropoiesis and death after about 12 weeks^{58, 59}. These data suggest that CD122 is important for preventing the abnormal activation of T cells. To further support the role of CD122 in preventing abnormal T cell activation, adoptive transfer of CD8⁺CD122⁺ cells into CD122 deficient neonates rescued the abnormal T cell activation. These data support that CD8⁺CD122⁺ cells are Treg cells³⁵.

In this regard, it is worth mentioning that CD122 is also a marker of central memory T cells. Therefore, it seems unlikely that the entire CD8⁺CD122⁺ T cells are Treg cells. However it is possible that CD8⁺ Treg may be contained within the CD122⁺CD8+ T cells. Consistently, some previous studies showed that depletion and/or blocking of CD122 by monoclonal antibodies restored immune tolerance in mice with type 1 diabetes^60–63 and Celiac disease⁶⁴, while another study showed that depletion of CD122⁺ cells increased the incidence of hyperthyroidism in an animal model of Graves’ disease⁶⁵. Thus, it is clear that CD122⁺ cells contain a heterogenous population and, therefore, cannot be an exclusive marker for CD8⁺ Treg cells.

Several efforts have been made to identify markers that can further define CD8⁺CD122⁺ Treg cells, which resulted in several subpopulations that appear to narrow down the phenotype for CD8⁺CD122⁺ Treg cells. Using CD8⁺CD122⁺ Treg cell clones, we showed that CD8αα is a unique marker that identifies CD8⁺CD122⁺ T cells as Treg cells³⁷. However, except for the CD8αα Treg cells that are normally found in the intestinal inter-epithelial lymphocytes (IELs) and are considered as a separate cell lineage different from immune cells in other peripheral lymphoid tissues, we previously were not able to purify CD8⁺CD122⁺ Treg cells from peripheral lymphoid tissues of naïve mice because of their lower frequency. Recently, we have investigated different tissues and found that liver is the tissue in naïve mice where CD8⁺CD122⁺ CD8αα Treg cells reside⁶⁶. Importantly, we were also able to identify the CD8⁺CD122⁺ CD8αα Treg in human peripheral blood⁶⁶. In addition, it has been shown that the expression of Ly49 defines a population within the CD8⁺CD122⁺ T cells as Treg cells^{3, 67–71}. However, these CD8⁺CD122⁺Ly49⁺ Treg cells are only detectable in immunized animals^{3, 67–70}. It is important to emphasize that CD8⁺CD122⁺ CD8αα Treg cells also express of inhibitory but not activating Ly49 receptors^{3, 67–70}. Therefore, the presence of CD8⁺CD122⁺ CD8αα Treg cells expressing inhibitory Ly49 receptors in the liver of naïve mice as well as in the human peripheral blood⁶⁶ represent a naturally occurring CD8⁺ Treg cells. It is likely that de novo primed CD8⁺CD122⁺Ly49⁺ CD8⁺ Treg cells may represent induced CD8⁺ Treg cells. However, it is not known if there is an overlap among these populations with respect to the expression of CD8αα versus the CD8αβ chain.

In addition to the above two populations of CD8⁺CD122⁺ Treg cells, in in vitro and in vivo systems in which CD8⁺CD122⁻ cells were used as effector T cells, it was shown that the regulatory activity of CD8⁺CD122⁺ Treg cells resided in the CD49^low but not in the CD49^high population³⁶. Moreover, in an animal model of skin allograft, it was shown that PD-1 defined CD8⁺CD122⁺ T cells as Treg cells, meaning that CD8⁺CD122⁺PD-1⁺ cells are Treg cells while the CD8⁺CD122⁺PD-1⁻ cells are memory T cells^{72, 73}.

Finally, beside the aforementioned extensively investigated cell surface markers, other CD8⁺ Treg cell surface markers have also been described in the literature such as CD8⁺CD28^low74–76, CD8⁺CD45RC^{low77, 78}, and CD8⁺CD25⁺CD127^low79–81. Hence, more investigations are urgently needed to determine the relationship among these various cell surface markers for the CD8⁺ Treg cells.

MHC RESTRICTION

In the late 1970s to early 1980s, an I-J determinant was serologically and biologically defined to be a restriction element for suppressor T cells and suppressor factors⁸². However, genetic analysis of the MHC I region did not support the I-J hypothesis³⁴ and remained a puzzle⁸³. As described above using several in vivo models since 1990s, CD8⁺ Treg cells have been shown to be restricted by both classical as well as non-classical MHC-I molecules^{3, 38, 66}.

Historically, a nonclassical MHC I molecule, i.e. Qa-1 (HLA-E in humans), has been extensively investigated as a restriction element for CD8⁺ Treg cells. In 1970s, it was shown that activated CD4⁺ T helper (Th) cells could prime CD8⁺ T cells. Such primed CD8⁺ T cells could subsequently suppress the ability of the activated CD4⁺ Th cells to provide help for B cells to produce antibodies (feedback regulation)^{46, 84}. Interestingly, Qa-1⁺ but not Qa-1⁻ CD4⁺ Th cells could activate CD8⁺ T suppressors (or CD8⁺ Treg cells)⁴⁷. These data suggest that Qa-1 molecules on the cell surface of activated CD4⁺ Th cells may be necessary for the activation of the CD8⁺ Treg cells. These early findings later led to more extensive investigations on CD8⁺ T cell-mediated, Qa-1-dependent regulation of activated CD4⁺ T cells in other experimental systems.

In one experimental system, mice were administered with a superantigen staphylococcus enterotoxin B (SEB) that can bind specifically to the TCRVβ8 protein and activates TCRVβ8⁺ T cells independent of antigen-specificity. Following the SEB administration, TCRVβ8⁺ CD4⁺ T cells in the mice were specifically activated and expanded to the maximal level on day 4. Subsequently, the number of TCRVβ8⁺ CD4⁺ T cells decreased to about 30%−40% below baseline levels^85–88. Study of the SEB-administered mice showed that the numeric reduction of TCRVβ8⁺ CD4⁺ T cells was dependent on CD8⁺ T cells³⁹. Further investigation demonstrated that CD8⁺ T cells from the SEB-administered mice showed TCRVβ8-specific cytotoxicity towards targets and the cytolytic activity was blocked by antisera to Qa-1, but not by antibody to classical MHC class Ia molecules. These data indicated that the CD8⁺ T cell-mediated killing of activated TCRVβ8⁺ cells depended on Qa-1 but not classical MHC Ia molecules. The concept of Qa-1-dependent immune regulation of activated CD4⁺ T cells was also supported by findings in Qa-1-deficient mice in which exaggerated secondary CD4⁺ T cell responses to foreign and self-peptides were observed³⁸. Accordingly, the foregoing data clearly demonstrate that CD8⁺ Treg cells can be de novo primed by activated CD4⁺ T cells and perform a Qa-1-dependent feedback immune regulation.

Since Qa-1-restricted CD8⁺ T cells may express both Qa-1-binding T cell receptors (TCRs) and NKG2 receptors, the foregoing findings could not conclude which receptors were responsible for the Qa-1-dependent feedback regulation. This question was recently addressed using Qa-1 mutant mouse strains. Studies of these mutant mice showed that, in the absence of TCR-Qa-1 interaction, the immune responses to self-antigens^{67, 89}, cancers⁹⁰, and viruses⁹¹ were significantly augmented.

In our study, we showed that CD8⁺CD122⁺ CD8αα Treg cells were significantly decreased in Qa-1-deficient mice⁶⁶. Importantly, we were also able to generate CD8αα Treg cell clones specific for a TCR peptide from the Vβ8.2 chain and showed that the cloned cells were restricted by Qa-1 molecules^{57, 92, 93}. In addition, we recently identified a Qa-1-binding epitope (MOG196) in myelin oligodendrocyte glycoprotein and showed that MOG196 immunization effectively suppressed ongoing EAE and activated MOG196-specific, Qa-1-restricted CD8 T cells which specifically migrated into the draining lymph nodes⁷⁰.

Therefore, the aforementioned extensive studies have convincingly demonstrated that Qa-1/HLA-E molecules are important restricted elements for CD8⁺ Treg cells. In addition to Qa-1/HLA-E, classical MHC I molecules have also been shown to be restriction elements for CD8 Treg cells^{3, 94}. Recent studies also have shown that CD8⁺CD122⁺Ly49⁺ CD8⁺ Treg cells in the EAE mice as well as in MS patients can also be restricted by the class Ia MHC molecules. Thus, it is clear that both class Ia as well as class Ib MHC molecules can be used by CD8⁺ Treg cells. Since the peptides that generally binds to nonclassical MHC Ib molecules are much more hydrophobic in comparison to those that bind to classical MHC Ia MHC molecules, it is not known whether the target antigens recognized could be different for these different MHC-restricted CD8⁺ Treg subsets.

GENETIC CONTROL

The FoxP3 has been described as the master regulatory transcription factor for CD4⁺ Treg cells. However, FoxP3 is not required for the development and function of CD8⁺ Treg cells^{66, 67, 72, 75, 76}, although some reported CD8⁺ Treg cells could express FoxP3^77–81. While investigating the potential CD8⁺ Treg population in the liver mononuclear cells from naïve B6 mice, we found that most of the CD8αα⁺ Treg cells were NK1.1⁺, CD69⁺, CD44^hi, and CD62^lo. The expression of these innate-like, unconventional cell surface markers in CD8⁺ Treg cells prompted us to investigate whether these features are driven by the expression of the PLZF (Zbtb16) transcription factor that has been shown to control the development of innate-like T cells^{95, 96}. We found that CD8⁺CD122⁺ Treg cells are PLZF⁺ and their development/survival is dependent upon the expression of PLZF⁶⁶. Using mice in which eGFP expression was under the control of PLZF regulatory elements⁹⁶ and the mice in which PLZF⁺ cells were permanently labeled with tdTomato (red fluorescence), we found that CD8⁺CD122⁺ CD8αα Treg cells, but not conventional CD8⁺ T cells, expressed PLZF⁶⁶. Importantly, PLZF deficiency led to a significant decrease in the number of naturally occurring CD8⁺CD122⁺ CD8αα Treg cells⁶⁶. Hence, for the first time we have identified a transcription factor that is crucial for the development of CD8⁺ Treg cells. Also importantly, since PLZF is stable and not inducible upon activation of CD8⁺ T cells⁹⁶, PLZF expression in CD8⁺ Treg allows differentiating regulatory CD8⁺ Treg cells from non-regulatory or conventional CD8⁺ T cells. How and why PLZF expression endows the ability of CD8⁺ Treg cells to control immune response is not clear at present.

In addition to PLZF that controls the development/survival of CD8⁺ Treg cells, Helios was suggested to be important for the functional stability of CD8⁺ Treg cells. Helios was initially proposed as a marker specific for thymus-derived CD4⁺ Treg (tTreg) cells^97–100. However, other studies showed that peripherally induced CD4⁺ Treg (pTreg) cells are frequently Helios^+101–103. Comparison between Helios⁺ and Helios⁻ CD4⁺ Treg cells showed that Helios⁻ CD4⁺ Treg cells expressed the genes that are normally associated with conventional CD4⁺ T cells^104–106 and Helios⁺ CD4⁺ Treg cells were more suppressive and stable^{104, 107, 108}. In addition, Helios appears to be important for the stability of FoxP3 expression. It was shown that, when transferred into immune deficient recipients, Helios⁻ CD4⁺ Treg cells, when compared to Helios⁺ CD4⁺ Treg cells, lost 50% more FoxP3 expression. The above findings were consistent with the other finding that Helios⁺ CD4⁺ Treg cells demethylated more Treg cell-specific regions because such demethylation is an indication of stable FoxP3 expression¹⁰⁹. In addition to CD4⁺ Treg cells, it has been shown that Helios is also required for the stability of CD8⁺ Treg cells¹¹⁰. However, the mechanism by which Helios stabilizes CD8⁺ Treg cells, which do not express FoxP3, are not known.

MECHANISMS OF REGULATION

Different regulatory mechanisms have been described to mediate the regulatory function of CD8⁺ Treg cells. We will briefly summarize some of these mechanisms, including killing of the target cells, secretion of suppressive cytokines, and induction of tolerance in antigen presenting cells via cell-cell contact or negative signaling.

The direct killing of target T cells has been shown to mediate the regulatory function of CD8⁺CD122⁺ Treg cells restricted by both classical and nonclassical MHC I molecules. Firstly, Qa-1-restricted CD8⁺ Treg cells have been shown to down regulate target T cells via cytotoxic mechanisms. An early study showed that CD8⁺ T cells from SEB-administered mice showed TCRVβ8-specific cytotoxicity towards activated CD4⁺ T cells and that the cytolytic activity could be blocked by antisera to Qa-1 molecules but not antibodies to classical MHC Ia molecules³⁹. This ability of Qa-1-restricted CD8⁺ T cells to specifically kill target cells was also observed in Qa-1-restricted CD8⁺ T cell clones specific for a TCRVβ8.2-derived peptide^{37, 92}. The above in vitro data were also supported by in vivo data. In the study of TCR peptide-mediated suppression of EAE, it was shown that the regulatory functions of Qa-1-restricted TCR peptide-specific CD8⁺ Treg cells depended on perforin but not Fas/FasL¹¹¹. Additionally, in the study of the regulation of T follicular helper (Tfh) cells, it was also found that regulatory functions of Qa-1-restricted CD8⁺ Treg cells depended on perforin⁶⁷. Besides perforin, IFN-γ and IL-15 have also been shown to be necessary for Qa-1-resgtricted CD8⁺ Treg cells to execute regulatory functions^{67, 111}. To determine the role of IFN-γ, one study administered both CD8⁺ Treg cells and CD4⁺ T cells into mice that do not have functional T, B, and NK cells. It was shown that the expression of interferon-γ (IFN-γ) receptor but not IFN-γ in Qa-1-restricted CD8⁺ Treg cells was required for the suppression of activated CD4⁺ T cells⁸⁹, suggesting that IFN-γ is necessary for the regulatory function of CD8⁺ Treg cells but does not directly participate in the regulation of target T cells. The role of IL-15 is obvious because Qa-1-restricted CD8⁺ Treg cells express CD122 that is the receptor for IL-15. Hence, similar to IFN-γ, IL-15 may not directly participate in the regulation of target cells, but is important for the survival and expansion of CD8⁺ Treg cells. In addition to Qa-1-restricted CD8⁺ Treg cells, CD8⁺ Treg cells restricted by classical MHC I molecules were also shown to use cytotoxicity to execute the regulation of target cells. A recent study showed that the activation/expansion of myelin-specific autoimmune CD4⁺ T cells triggered the expansion of CD8⁺ Treg cells that were restricted by classical MHC I molecules and were not reactive to myelin proteins³. Using H2-D^b yeast peptide-MHC libraries and MHC tetramers, the study identified four peptides that were recognized by the CD8⁺ Treg cells. However, the mouse genome search did not find any matches with these four peptides. Interestingly, immunization with the four peptides suppressed EAE³. Importantly, consistent with our data, such expanded CD8⁺ Treg cells also required perforin to suppress the proliferation of myelin-specific CD4⁺ T cells in vitro³. However, perforin does not seem to be the only cytotoxic mechanism used by CD8⁺ Treg cells, another study showed that CD8⁺CD122⁺PD-1⁺ Treg cells did not suppress Fas-deficient effector T cells but still performed suppressive activity in the absence of perforin, suggesting that their suppressive activity depended on the Fas/FasL interaction⁷³. Similarly, CD8⁺CD122⁺CD49d^low Treg cells are shown to kill activated T cells via Fas/FasL-mediated cytotoxicity³⁶.

Immunosuppressive cytokines have also been reported to play an important role in some subsets of CD8⁺ Treg cell-mediated suppression. In the study of the role of CD8⁺CD28^low T cells in the control of colitis, it was shown that IL-10 deficient CD8⁺CD28⁻ T cells did not prevent colitogenic T cell-induced colitis and that CD8⁺CD28^low T cells did not prevent colitis if colitogenic CD4⁺ T cells were unable to respond to transforming growth factor (TGF-β)¹¹². In addition, it was shown that suppression of both proliferation and IFN-γ production by CD8⁺CD122⁻ cells by CD8⁺CD122⁺ Treg cells was blocked by anti-IL-10 but not anti-TGF-β monoclonal antibody¹¹³. Moreover, it was shown that adoptive transfer of CD8αα⁺ IELs from IL-10^+/+ mice but not from IL-10^−/− mice resulted in the downregulation of colitis in recipient lymphopenic mice¹¹⁴. Thus, it is likely that in certain experimental conditions, suppressive cytokines, such as IL-10 may also contribute in CD8⁺ Treg cell-mediated immune regulation.

Finally, CD8⁺ Treg cells have been shown to execute their regulatory functions by induction of tolerance in antigen-presenting cells via cell-cell contact. In the study of the role of CD8⁺CD28^low Treg cells in the control of EAE, it was shown that CD8⁺CD28^low Treg cells required antigen-presenting cells to suppress IFN-γ production by myelin oligodendrocyte glycoprotein-specific CD4⁺ T cells. Specifically, the CD8⁺CD28⁻ Treg cells suppressed the activation of CD4⁺ T cells by preventing the upregulation of costimulatory molecules in antigen-presenting cells¹¹⁵. Since CD8⁺ Treg cells also express several cell surface regulatory molecules, such as CD200 and Ly49^{3, 67–71}, it will be interesting to find out whether they could potentially use negative signaling in the tolerance induction in APC.

FUTURE PERSPECTIVE

Ample experimental evidence in both mice and humans has shown that CD8⁺ Treg cells play an important role in the maintenance of peripheral immune tolerance. One important feature of CD8⁺ Treg cells is that they appear to be expanded following the initiation of an immune response and prevent the over activation of activated effector T cells. In this regard, CD8⁺ Treg cells may play a complementary role along with the FoxP3⁺CD4⁺ Treg cells. Another important feature of CD8⁺ Treg cells is their innate-like characteristics, including the expression of PLZF and NK inhibitory receptors. It is not known as to how these innate characteristics are involved in the regulatory function of CD8⁺ Treg cells. Since CD8⁺ Treg cells are restricted by both classical MHC Ia and non-classical MHC Ib molecules, they have at least two subsets. Hence, future studies should focus on characterizing the peptides recognized by both subsets, their unique molecular signatures, as well as potentially distinct regulatory mechanisms. In addition, it is yet to know whether PLZF expression, similar to the FoxP3, is necessary for the immune regulatory function of CD8⁺ T cells and whether other genetic control mechanisms play a crucial role in enabling a regulatory behavior. Nonetheless, recent characterization of the phenotypes that can differentiate them from conventional CD8⁺ T cells will greatly facilitate the characterization of CD8⁺ Treg cells. Looking forward, one can image that novel strategies will emerge that can activate/expand CD8⁺ Treg cells in vivo for the treatment of autoimmune and inflammatory diseases.

Abbreviations:

EAE

Experimental autoimmune encephalomyelitis

FoxP3

Forkhead box P3

IFN-γ

Interferon-γ

MHC

Major histocompatibility complex

MOG₁₉₆

A Qa-1-binding epitope in myelin oligodendrocyte glycoprotein

PLZF

Promyelocytic leukemia zinc finger

PD-1

Programmed cell death-1

SEB

Staphylococcus enterotoxin B

TCR

T cell receptor; Treg cells: Regulatory T cells