IL-17-Producing γδ T Cells

Introduction/Abstract

IL-17 is produced by CD4+ αβ T cells, but also CD8+ αβ T cells, NKT cells, and γδ T cells, plus some non-T cells, including macrophages and neutrophils. The ability of IL-17 to deploy neutrophils to sites of inflammation imparts this cytokine with a key role in diseases of several types. Surprisingly, γδ T cells are responsible for much of the IL-17 produced in several disease models, particularly early on.

1. Disease models showing a role for IL-17 producing γδ T cells

Initial indications

Although the study of IL-17A (IL-17) production by T cells has largely focused on CD4+ αβTCR+ Th17 cells, several reports already indicated early on that γδ T cells were also a major source of this cytokine. In a 2004 study investigating the strong neutrophil response made by mice injected intraperitoneally with a FasL+ tumor cell line [1], the cytokine IL-17 was found to be key in this process, and to be produced by a high proportion of peritoneal γδ T cells, although CD4-CD8- αβ T cells also produced IL-17, and to a lesser extent CD4+ αβ T cells. These unconventional αβ T cells and γδ T cells were collectively termed “Tn” cells, for neutrophil-regulatory T cells, in another paper published in 2005 [2] examining neutrophilia in adhesion molecule-deficient mice. Here, naive γδ T cells in the spleen and lymph nodes were found to be a potent source of IL-17. In the absence of adhesion molecule-directed transmigration of neutrophils into tissue, the normal apoptosis of neutrophils and their subsequent uptake by macrophages and dendritic cells is interrupted. The macrophages and dendritic cells thus fail to downregulate their production of IL-23, which further enhances IL-17 production, leading to neutrophilia. Consistently, the transcription factor RORγt, which is critical for IL-17 secretion, was also found to be very abundant in intestinal γδ T cells [3]. In the last two years, a flurry of studies reporting IL-17 production by γδ T cells in many different disease models suggests that a large part of the immunological role of γδ T cells may in fact depend upon their production of IL-17 and related cytokines.

Infectious Models

In a study involving mice infected with Mycobacterium tuberculosis [4], most of the IL-17 produced was shown to come from γδ T cells. IL-17-producing CD4+ αβ T cells were rare in this model, although CD4-negative αβ T cells also produced substantial amounts of IL-17. Interestingly, the authors also found that naive γδ T cells made IL-17 when cultured with dendritic cells taken from M. tuberculosis-infected mice, but not when cultured with uninfected dendritic cells, and this appeared to be due to IL-23 production by the dendritic cells. In fact, stimulation of the γδ T cells instead via the TCR only poorly induced them to express IL-17. The γδ T cells were found to produce IL-17 not only early on but also throughout the infection, with the level falling off somewhat by 4 weeks post-infection, whereas IL-17 production by CD4 cells was not evident until about week 2.

In mice infected intratracheally with Mycobacterium bovis BCG [5], γδ T cells were also found to be the predominant producers of IL-17, beginning as early as one day after infection. Non-T cells also produced some IL-17 in this model. As with M. tuberculosis infection, IL-23 appeared to be driving IL-17 production by γδ T cells. IL-17 was also shown to be important for the development of protective immunity in this model, because IL-17-/- mice showed impaired granuloma formation following BCG infection, a decreased percentage of BCG-specific Th1 cells, and a decreased DTH response to mycobacterial antigen.

Even more rapid kinetics for an IL-17 response by γδ T cells were noted in an E. coli intraperitoneal infection model in mice [6]. Here, IL-17 was already detectable 1 hour after infection, and peaked at 6 hours, with peritoneal neutrophils reaching their maximum numbers at 24 hours. Again, the IL-17 appeared to be important for host protection because neutralizing IL-17 by administering an anti-IL-17 monoclonal antibody during infection substantially reduced bacterial clearance. The IL-17 response was critically dependent upon the ability of the host to respond to LPS via TLR4, which results in IL-23 production, and IL-23 alone was shown to be sufficient to induce IL-17 secretion by γδ T cells from the infected mice, and to a lesser extent by naive peritoneal γδ T cells. The responding γδ T cells mainly consisted of the Vγ6Vδ1 γδ T cell subset, a TCR-invariant subset that was shown previously to respond preferentially during intraperitoneal E. coli infection [7, 8].

In a related study, γδ T cells were also found were the be main source of IL-17 in mice infected intraperitoneally with Listeria monocytogenes, again with relatively rapid kinetics [9]. IL-17 was detectable within 1 day following infection, and found to be critical for the formation of small contained granulomatous liver lesions, and for efficient bacterial clearance. Moreover, IL-17 -/- bone marrow chimeric mice containing γδ T cells able to produce IL-17 cleared L. monocytogenes from the liver much more efficiently than did bone marrow chimeras having γδ T cells unable to secrete IL-17, illustrating the importance of IL-17 for γδ T cells responding in early infection. Although in this study both Vγ4+ and Vγ6Vδ1+ cells were found to increase in numbers in the liver and to produce IL-17, we have found that the response of γδ T cells in the peritoneum of L. monocytogenes-infected mice is overwhelmingly dominated by the Vγ6Vδ1 subset, of which over 60% express IL-17 (unpublished results).

Experimental sepsis can be induced in mice by cecal ligation and puncture, a process in which the cecum, an outpocketing at the beginning of the large intestine, is tied off and then deliberately perforated [10]. This results in live, replicating bacteria in the bloodstream, and causes death to 80-90% of the mice within 3 days. A high level of IL-17 is observed in the blood in these mice, and is dependent upon the presence of γδ T cells. The IL-17 is clearly pathogenic, since neutralizing anti-IL-17 monoclonal antibody substantially improved survival. This makes sense because it seems likely that the mice die of septic shock rather than bacterial infection in this model. However, anti-IL-17 treatment also substantially reduced bacterial levels in the blood, which was surprising, especially in light of the concomitant decrease in proinflammatory cytokines and chemokines that resulted from the treatment.

IL-17-producing γδ T cells were recently reported in a mouse model of chronic infection involving an attenuated strain of Salmonella enterica Enteriditis [11]. Here, IL-17 was found to improve bacterial clearance by about 10-fold, both at 20 and 80 days post-infection, in mice infected intraperitoneally with a relative high bacterial dose. The cells producing IL-17 proved to be CD4+ Th17 cells, γδ T cells, and CD4- γδTCR- cells, each in about equal numbers by intracellular cytokine staining. Although IL-17 in general did not appear to be critical in bringing about most responses to the disease, reduced numbers of neutrophils were seen in S. Enteriditis-infected IL-17-/- mice. This chronic infection model shows that γδ T cells can continue to be an important source of IL-17 even late during an immune response.

Common findings in most of these mouse infectious disease models include the relatively rapid induction of IL-17-producing γδ T cells, their frequent dependence upon IL-23, and the observation that they often comprise an important component of host protection against the infectious agent. A number of recent papers have indicated that IL-17-producing T cells are also important in human disease [12], and several parallels between human and mouse CD4+ Th17 cells have been noted. One very recent study also showed that a large proportion of the IL-17-producing cells in the peripheral blood of healthy control subjects consists of γδ T cells, and that these become the majority of the IL-17+ population in tuberculosis patients [13]. Whether IL-17-producing γδ T cells in either mice or humans are also triggered by infection with microorganisms other than bacteria has yet to be shown, however.

Immune-mediated disease models

IL-17 is important in inducing pathology in a variety of autoimmune and atopic disease models in mice, including collagen-induced arthritis, experimental autoimmune encephalomyelitis, experimental autoimmune uveoretinitis, asthma models, and colitis [12]. In addition, IL-17 may play a role in some immunodeficiency diseases as well, because lack of IL-17 in hyper IgE syndrome, which often stems from mutations in STAT3 [14], appears to be at least partially responsible for the susceptibility of these patients to many common pathogens. Although in most such diseases, research on the role of IL-17 has focused solely on CD4+ αβTCR+ Th17 cells, it seems likely that γδ T cells produce at least some of the IL-17 and are involved in the disease process.

In collagen-induced arthritis in mice, we were in fact able to demonstrate that IL-17-producing γδ T cells play an important role [15]. Here, γδ T cells producing IL-17 were limited to a very specific subset of Vγ4Vδ4+ cells, present in elevated numbers in the arthritic joints and their draining lymph nodes. In the diseased mice, these cells were found to be approximately equivalent in number to CD4+ αβTCR+ IL-17-producing cells. After being selectively depleted for Vγ4+ TCRs, mice with collagen-induced arthritis showed a lower disease incidence as well as overall reduced disease scores as compared to sham-treated controls, indicating that these IL-17-producing γδ T cells contribute to pathology.

A new model of hypersensitivity pneumonitis and subseqeunt lung fibrosis was recently developed in mice, in which disease is induced by repeated intranasal instillation of the nonpathogenic bacterial species Bacillus subtilis. This induces a surprisingly strong response by Vγ6Vδ1 γδ T cells, such that Vγ6Vδ1+ cells come to represent more than twice the number of αβ T cells infiltrating the lungs [16]. A large proportion of the infiltrating Vγ6Vδ1+ T cells were found to produce IL-17 in this model [17]. The role of the IL-17 is currently being investigated, but because mice unable to produce this γδ T cell subset {B6.Vγ4/6-/- mice [18]} show accelerated pulmonary fibrosis as compared to wildtype controls [16], it likely has a protective effect.

Indoleamine 2,3-dioxygenase (IDO) is an enzyme involved in the catabolism of tryptophan, which also plays a role in suppressing T cells responses. Mice defective for this enzyme (p47^phox-/- mice) develop an immunodeficiency that resembles chronic granulomatous disease (CGD), and like CGD patients, they are particularly susceptible to pulmonary infection by Aspergillus fumigatus. In p47^phox-/- mice infected with A. fumigatus, a strong response of IL-17 producing Vγ1+ T cells was recently reported, along with enhanced neutrophil infiltration compared to wildtype controls, and subsequent acute lung injury [19]. The lung damage could be ameliorated either by neutralization of the IL-17 or γδ T cell depletion, indicating the importance of the IL-17-producing γδ T cells in disease induction. Interestingly, treatment of the p47^phox-/- mice with a combination of IFN-γ and L-kynurenine, a breakdown product of tryptophan normally produced by p47, not only improved survival and prevented lung injury but also markedly reduced the Vγ1+ lung γδ T cells in bronchoalveolar lavage fluid.

Injury Model

Bleomycin is an anti-cancer drug having known pulmonary toxicity, and has been used to study lung fibrosis in rodents. A recent study showed that bleomycin-induced lung injury in mice evokes a strong response by IL-17-producing γδ T cells [20]. Here, one to two weeks after intratracheal bleomycin instillation, 30-40% of lung γδ T cells were found to produce IL-17 by intracellular cytokine stain when stimulated by PMA/ionomycin, compared to less than 4% of CD4+ cells; γδ T cells were clearly the predominant IL-17 source in the lungs in this model. In bronchoalveolar lavage fluid from γδ TCR-/- mice, neutrophil numbers were found to be drastically reduced compared to wildtype controls, suggesting that the IL-17 normally produced by γδ T cells is important in neutrophil recruitment in this model. The γδ TCR-/- mice also showed increased interstitial inflammation and collagen deposition (fibrosis), which could be partially explained by the ability of IL-17 to stimulate lung epithelial cell proliferation.

2. Characteristics of IL-17+ γδ T cells

In the disease models reviewed above, in several instances, particular IL-17-producing γδ T cell subsets (defined by the TCRs they express) showed preferential responses. In three systems - Listeria infection [9], E. coli infection [6], and B. subtilis-induced hypersensitivity pneumonitis in mice [16] - the TCR invariant Vγ6Vδ1 cells showed a preferential or even exclusive response among γδ T cells. In two models, the response was dependent upon the elaboration of IL-23 by other cells, which may reflect an inherent bias of this subset to immediately respond to IL-23 with no need for prior stimulation via the TCR. In support of this interpretation, efficient IL-17 production in response to IL-23 by mouse Vγ6Vδ1+ cells has been shown to require tyrosine kinase 2 [21], known to be involved in cytokine-mediated signaling. In marked contrast, the response of the Vγ4Vδ4 subset in mice with collagen-induced arthritis appears to involve stimulation through the TCR, because the response is limited not only to cells expressing a certain Vγ and Vδ, but also predominantly involves those having particular junctional motifs as well as conserved CDR3 lengths for both TCR chains [15]. Finally, in the chronic granulomatous disease model, a predominant response by IL-17 producing Vγ1+ cells was seen [19]. This may be a consequence of the induced mutation in these mice, because we have yet to find more than a trace of IL-17 production by Vγ1+ cells whether examining naive cells or in any of several disease models (unpublished results). IL-17-producing γδ T cells thus clearly represent several distinct subsets, which may differ not only in the stimuli needed to bring about their response, but in other ways as well. For example, although we have found that Vγ6Vδ1+ cells which produce IL-17 in the Listeria-infected peritoneum also co-produce IFN-γ, no IFN-γ production was seen among the IL-17-producing Vγ4Vδ4 cells in mice with collagen-induced arthritis (unpublished results). This question requires further investigation.

The Vγ6Vδ1 and Vγ4Vδ4 subsets differ from one another also in their development. Whereas the Vγ6Vδ1+ cells develop only in the fetal thymus and represent one of the first T cell types to colonize the newborn mouse, Vγ4+ cells continue to be produced in the adult thymus, so it seems possible that the two subsets undergo distinct thymic programming, which explains their differing IL-17 responses. Shibata et al. [22] recently addressed some of these questions in a study investigating the cytokines produced by γδ T cells expressing certain markers. Here, they showed that distinct subsets of mouse peritoneal γδ T cells produce different cytokines when stimulated by PMA/ionomycin, and these can be distinguished by their expression of CD25 (IL-2Rα) and CD122 (IL-2Rβ/IL-15Rβ): whereas those expressing CD25 nearly all produced IL-17, those expressing CD122 nearly all instead produced IFN-γ. The number of IL-17 producers was substantially lower in peritoneal cells from IL-2-/- and CD25-/- mice, although the number of IFN-γ+ γδ T cells was unchanged, indicating a role for IL-2 in the development of the IL-17 producers. Conversely, in IL-15-/- mice, the number of IFN-γ+ γδ T cells in the peritoneum was substantially reduced, although IL-17+ γδ T cell numbers were unaffected. In the normal fetal thymus, many IL-17+ γδ T cells were found to be present even though none yet expressed CD25, so role for this receptor is probably post-thymic, after functional imprinting in the thymus has already taken place. IL-17+ fetal γδ T cell numbers peaked on day 19 of development, and appeared to consist almost entirely of Vγ6Vδ1+ cells, which are produced during late fetal/early newborn life only; the Vγ5Vδ1 subset which also develops at this time, and is destined to colonize the epidermis, in contrast was mostly IL-17-negative. The peritoneal γδ T cells in adults included Vγ6Vδ1+ cells although it was not evident from this study whether peritoneal γδ T cells expressing other TCR types were also IL-17+, but this seems likely from the high percentage reported (∼50% of all peritoneal γδ T cells in 7-week old hosts).

In a recent study from the Chien laboratory [23], γδ T cells that develop in the thymus were also found to express either IFN-γ or IL-17, but not both, in accordance with whether or not they had encountered a TCR ligand during thymic development. Using the T10/T22 nonclassical MHC class I reactive γδ T cells as a model, those biased to produce IFN-γ were found to develop when a ligand was present (in β2-microglobulin sufficient hosts), whereas their bias was to produce IL-17 if no ligand was available (in β2-microglobulin deficient hosts). Consistent with the findings of Shibata et al. [22], CD122 was low on thymic and splenic T10/T22 reactive γδ T cells that produced IL-17, but high on T10/T22 reactive γδ T cells that produced IFN-γ. Curiously, although naive lymph node-derived CD122-low γδ T cells also had a tendency to produce IL-17 when stimulated with plate-bound anti-Cδ antibody, IFN-γ was not detected among naive lymph-node derived CD122-high γδ T cells. Moreover, following stimulation with peptide and CFA, both T10/T22-specific and non-specific γδ T cells from the draining lymph nodes produced IL-17, so it would appear that thymic “imprinting” can be altered by peripheral conditions. The Chien laboratory study also presented evidence that the ability of most γδ T cells to develop without the need for a TCR ligand could be due to an inherent ability of γδ TCRs to self-dimerize, by-passing the need for a cross-linking ligand. The epidermal Vγ5Vδ1+ cells appear to be an exception here, since they are unable to develop in the absence of what appears to be a ligand for this TCR, the skint1 gene product, a member of the Ig superfamily [24].

Another very recent study suggests that other mouse γδ T cell subsets also may be imprinted to secrete IL-17 during thymic development. These cells express scart1 and scart2, two newly identified mouse scavengers receptors closely related to WC1, a scavenger receptor found on many γδ T cells in sheep and cattle, and like WC1 these are predominantly expressed by γδ T cells [25]. Using a polyclonal antiserum, scart2+ γδ T cells were revealed in the thymus and skin-draining lymph nodes, but not in the mesenteric lymph nodes or spleen. In the inguinal lymph nodes, 5-10% of the γδ T cells were found to express scart2, and although about 75% of Vγ4+ cells were scart2-negative, nearly all scart2+ cells were Vγ4+. The scart2+ γδ T cells also tended to be CD5-low, HSA-low, and CD44-high. In addition, scart2 is expressed on about 20% of dermal (but not epidermal) Vγ5-negative γδ T cells, so an association between passage through the dermis and scart2 expression may exist. When stimulated with PMA/ionomycin, a high percentage of scart2+ naive lymph node-derived γδ T cells produced IL-17. The almost exclusive expression of Vγ4, high expression of activation markers including CD44, and strong tendency to produce IL-17 when stimulated by PMA/ionomycin, are also all features of the Vγ4Vδ4+ cells we detected that expand preferentially in mice with collagen-induced arthritis [15]. Whether this Vγ4Vδ4+ subset also expresses scart2 remains to be seen.

3. Distinctions from CD4+ αβ Th-17 cells

The differentiation of CD4+ αβTCR+ Th17 cells from naive uncommitted CD4 cells generally takes place in vivo over several days, and is currently thought to depend on both TCR engagement and the cytokines TGF-β and IL-6, at least in mice. Moreover, IL-21, also produced by Th17 cells, was in some studies shown to be able to substitute for IL-6. For human naive Th17 cells, many studies have found a role for IL-1 or IL-23 as well, and a role for TGF-β was less clear, but it now appears that high amounts of TGF-β inhibit Th17 development, whereas low levels promote it. There is still controversy here, and in at least one study, either IL-23 or IL-1 was able to cause human Th17 cells develop from naive CD4 cells, in the presence of cross-linking anti-CD2, anti-CD3, and anti-CD28 antibodies [12]. Memory Th17 cells in contrast begin to secrete IL-17 almost immediately in the presence of IL-23, or for human Th17 memory cells, with a combination of IL-1 and IL-23. It seems possible that some of these studies produced confusing results due to the presence of γδ T cells in the cultures, which can be stimulated by IL-23 alone. In fact, Shibata et al. showed that whereas IL-23 alone can induce IL-17 production by naive γδ T cells, a combination of TGFβ and IL-6 cannot [6]. Moreover, in at least one study, it was shown that for γδ T cells, TCR stimulation works less well to induce IL-17 production in vitro than does IL-23 [4]. Whether these findings are generalizable to all γδ T cells, or apply only to particular subsets, and whether human IL-17-producing γδ T cells behave similarly, have yet to be determined.

4. Conclusions

In several of the disease models reviewed here, evidence was presented showing that IL-17 producing γδ T cells not only respond during the disease, but also make a difference to the disease outcome [5, 6, 9, 15, 19, 20]. Moreover, in some instances, IL-17-producing γδ T cells appeared to play a more important role than CD4+ αβTCR+ Th17 cells [6, 9, 19]. Based on these findings, together with the many examples of diseases in which IL-17-producing γδ T cells have now been shown to respond, we suspect that IL-17-producing γδ T cells are generally as important to disease outcome as are CD4+ αβTCR+ Th17 cells. IL-17-producing γδ T cells appear to differ from Th17 cells in that they can be rapidly induced to produce cytokine, frequently respond in relatively large numbers, and often show no antigen specificity in their response. Thus, they are likely to make a very different contribution than Th17 cells to the immune response (Fig. 1). At least one report has now shown that human γδ T cells also can make IL-17, and these appear to be elevated in tuberculosis patients [13]. Therefore, IL-17-producing γδ T cells will likely prove to be a critical part of both the mouse and human immune systems.

Fig. 1.

Proposed differences between IL-17-producing γδ T cell development and Th-17 development. γδ T cells are presumed to be “naive” in mice that have not been deliberately stimulated, though some have been shown to bear memory markers. The scheme for Th17 cells was modified from that proposed in a review by Betelli et al. [26].