Direct and Indirect endocrine-mediated suppression of human endometrial CD8+T cell cytotoxicity

Regulation of endometrial (EM) CD8+T cells is essential for successful reproduction and protection against pathogens. Suppression of CD8+T cells is necessary for a tolerogenic environment that promotes implantation and pregnancy. However, the mechanisms regulating this process remain unclear. Sex hormones are known to control immune responses directly on immune cells and indirectly through the tissue environment. When the actions of estradiol (E2), progesterone (P) and TGFβ on EM CD8+T cells were evaluated, cytotoxic activity, perforin and granzymes were directly suppressed by E2 and TGFβ but not P. Moreover, incubation of polarized EM epithelial cells with P, but not E2, increased TGFβ secretion. These findings suggest that E2 acts directly on CD8+T cell to suppress cytotoxic activity while P acts indirectly through induction of TGFβ production. Understanding the mechanisms involved in regulating endometrial CD8+T cells is essential for optimizing reproductive success and developing protective strategies against genital infections and gynecological cancers.

The mucosal immune system in the human female reproductive tract (FRT) is precisely regulated to optimize conditions for successful fertilization, implantation, and pregnancy while providing protection against sexually transmitted infections (STIs)1. Unlike other sites in the FRT, the endometrium must protect itself from sexually transmitted pathogens while allowing for the presence of allogeneic spermatozoa and supporting an immunologically distinct fetus. Key to this unique balance within the FRT are the sex hormones, estradiol (E2) and progesterone (P), which are secreted by the ovary and whose levels fluctuate during the menstrual cycle and thereby regulate multiple aspects of innate and adaptive immunity2,3,4,5. During the proliferative phase, E2 is present at low concentrations. However, E2 levels surge at ovulation, decline rapidly, before increasing again during secretory phase, when fertilization and implantation occur. In contrast, P levels remain low throughout the proliferative and ovulatory phases, before rising and peaking in mid-secretory phase. E2 and P act directly on FRT immune cells, epithelial cells and fibroblasts to regulate their function, and modulate the production of cytokines, chemokines and growth factors, which in turn regulate the innate and adaptive immune systems. These direct and indirect effects of sex hormones are essential for successful reproduction, and influence susceptibility to STIs in ways that are site-specific in the upper (endometrium, endocervix) and lower (ectocervix, vagina) FRT.

T cells are central to adaptive immune protection in the FRT and represent a major constituent of leukocytes (30–60%) in the endometrium6,7,8,9. To measure the cytotoxic potential of CD8+T cells, White et al. evaluated the ability of FRT mixed cell suspensions to kill targets using a redirected lysis assay10. In addition to demonstrating cytotoxic activity throughout the FRT, these studies demonstrated that endometrial CD3+CD8+T cell cytotoxic activity is markedly influenced by the stage of the menstrual cycle. Cytotoxic activity was elevated during the proliferative phase of the cycle and suppressed during the secretory phase. Interestingly, cytotoxic activity in the endometrium from postmenopausal women was significantly higher than activity at any point of the cycle in premenopausal women. Cycle dependent suppression occurred without changes in total CD8+T cell numbers. More recently, using a time-lapse imaging system to measure direct killing by purified CD8+T cells from the endometrium, we confirmed changes in cytotoxic activity during the menstrual cycle and increased activity following menopause11. Maximal suppression of CD8+T cell cytotoxic activity in the endometrium coincided with the time when fertilization and implantation is likely to occur, and therefore may represent a mechanism to avoid recognition and rejection of sperm and fetal cells by maternal endometrial CD8+T cells1. Unresolved however, is the extent to which hormonal changes during the menstrual cycle, act directly or indirectly to regulate CD8+T cells cytotoxic activity in the endometrium.

It is well known that sex hormones, which act both locally and systemically, have a spectrum of immunological effects on adaptive immunity and CD8+T cells in particular. E2 is known to affect the maturation, differentiation, and function of T cells. Others have shown that CD8+T cells express ERα and ERβ receptors. For example, in rodents, E2 signaling through ERα was shown to be necessary for full differentiation of CD8+T cells12,13. In the human, retrovirally transduced peripheral blood mononuclear cells (PBMC) containing CD8+T cells incubated with E2 reportedly increased their expression of granzyme B, and IFN-γ14. In contrast, using human CD8+T cells, P has been shown to reduce IFN-γ production upon stimulation ex vivo, although the underlying mechanisms remain to be determined15. Beyond reduced production, other studies suggest that P impairs immune protective functions of antigen-non-specific CD8 T memory cells by inducing IFN-γ gene hyper-methylation16.

Beyond the direct effects of sex hormones, other soluble mediators are capable of modulating CD8+T cell cytotoxicity. For example, TGFβ inhibits intestinal CD8+T cell cytotoxicity through the suppression of perforin, without changes in T-bet, Eomes or granzyme B17. We recently reported that endometrial CD8+T cell cytotoxic function is also susceptible to suppression by TGFβ in both pre- and post-menopausal women, but that the inhibitory effect of TGFβ is reduced after menopause11. In animal studies, we found that TGFβ is produced by epithelial cells and fibroblasts from the FRT and under hormonal control in that E2 treatment increases the basolateral secretion of biologically active TGFβ in the rodent18,19. However, the mechanisms by which TGFβ regulates CD8+T cell function in the endometrium remains to be elucidated.

The effects of sex hormones and TGFβ on cytotoxic activity by CD8+T cells in the human have important implications for women’s health, especially as related to successful pregnancy, immunological infertility and protection against sexually transmitted diseases. The objective of this study was to evaluate the roles of E2, P and TGFβ in regulating endometrial CD8+T cell cytotoxic capacity. Our results demonstrate that E2 and TGFβ act directly on endometrial (EM) CD8+T cells to significantly suppress cytotoxic activity and intracellular perforin and granzyme A levels. In contrast, P has no direct effect on CD8+T cells but stimulates TGFβ secretion by EM epithelial cells. These studies provide insight into mechanisms that control cytotoxic activity in the endometrium to increase reproductive potential.

Estradiol and TGFβ suppress cytotoxic activity in CD8+T cells from the endometrium

We have previously demonstrated that cytotoxic activity of EM CD8+T cells is suppressed during the secretory phase of the menstrual cycle and enhanced after menopause10,11. In the present study, we investigated whether sex hormones are directly involved in the suppression of CD8+T cell cytotoxic activity. T cells were pretreated for 48 h with E2 or P at concentrations known to have physiological effects in vivo20 or with TGFβ for 2 h prior to measuring cell killing11. To measure cytotoxic activity, control and treated CD8+T cells were co-cultured with CFSE-labelled allogeneic target cells, and the number of dead cells quantified by time-lapse imaging as described before11. Based on our earlier studies for optimal killing, we selected an Effector:Target ratio of 1:1 and measured average cell killing during the first 6 h in culture. As seen in Fig. 1a–c, in representative time-course studies, relative to target cells alone, we detected a significant increase in dead target cells in the presence of allogeneic EM CD8+T cells. Moreover, when CD8+T cells were pretreated with E2 (Fig. 1a) or TGFβ (Fig. 1c), but not P (Fig. 1b), cell killing was markedly reduced. Figure 1d–f indicates the effect of hormones and TGFβ treatment on cytotoxicity measured both as total number of dead cells and as fold change in dead target cell numbers, the latter being done to compare experiments with different background mortalities of target cells alone. Consistent with our representative findings, relative to controls, E2 (Fig. 1d), and TGFβ (Figs. 1f), but not P (Fig. 1e), significantly reduced cell killing. Figure 1f confirms our earlier results11 that CD8+T cell cytotoxic activity is significantly reduced following TGFβ treatment of EM CD8+T cells.

EM CD8+T cell expression of perforin and granzyme A is inhibited by E2

To identify the underlying mechanisms involved in suppression of cytotoxic T cell activity, we evaluated the expression of perforin, granzyme A (GZA) and granzyme B (GZB) following preincubation of CD8+T cells with E2 or P. EM tissues were digested to obtain mixed cell suspensions and stained for intracellular cytotoxic molecules prior to analysis by flow cytometry as previously described11. Figure 2a contains representative histogram overlays of CD8+T cells for perforin, GZA and GZB indicating the downward shift in expression of perforin and GZA in response to E2 but not P. Analysis of five separate experiments with different subjects (Fig. 2b), demonstrated that E2 significantly reduced the intracellular content (MFI) of perforin and GZA in CD8+T cells. In contrast, E2 had no significant effect on GZB. Hormone treatment also affected the percentage of cells expressing cytotoxic molecules, as seen in Fig. 2c, with a significant decrease in the percentage of cells positive for perforin and GZA cells following incubation with E2. Overall, these studies demonstrate that E2 decreases the expression of perforin and GZA in EM CD8+T cells, which correlates with reduced cell cytotoxicity. Further, it indicates that these changes are hormone specific given that P had no effect on either perforin, GZA, and GZB intracellular expression or the percentage of positive cells.

EM CD103+and CD103- CD8+T cell expression of perforin and granzyme A is inhibited by estradiol

In previous studies, we demonstrated that human EM CD8+T cells consist of tissue resident cells and non-resident CD8+T cells that are CD103+and CD103−, respectively8,11. When cytotoxicity was compared between CD103+and CD103−CD8+T cells, direct cytotoxic activity and expression of cytotoxic molecules was lower in CD103+compared to CD103−T cells11. Therefore, in this study we explored whether the observed effects of E2 in reducing cytotoxic molecule expression were due to effects on both CD103+and CD103−T cell populations. Mixed cell suspensions from EM tissues were incubated with E2 for 48 h prior to measuring granular content by flow cytometry. Figure 3a shows a representative example of the content of perforin, GZA and GZB before and after treatment with estradiol. As shown in Fig. 3b, using matched samples from multiple patients, E2 significantly suppressed the content of perforin and GZA (MFI) in both tissue resident (CD103+) and non-resident (CD103−) CD8+T cells. In contrast, E2 had no effect on GZB. Interestingly, as seen in Fig. 3c, E2 reduced the percent positive CD103+and CD103−CD8+T cells containing perforin but had no effect on the percentage of cells that were GZA or GZB positive. These findings suggest that E2 regulates both CD103+and CD103−CD8+T cells and that suppression of the levels of cytotoxic molecules is unique in that each of the cytotoxic responses to E2 are separate and distinct. This is consistent with our previous results that menopausal status equally affects resident and non-resident T cells11, suggesting that tissue residency does not modify sensitivity to hormonal control.

Perforin, granzyme A and B Content in EM CD8+T Cells is suppressed by TGFβ

Next, we investigated the underlying mechanisms of TGFβ-mediated cytotoxicity suppression. CD8+T cells in mixed cell suspensions were incubated with TGFβ prior to perforin, GZA, and GZB analysis by flow cytometry at 48 h. As seen in the histograms (Fig. 4a) and analyses of intracellular content (Fig. 4b), TGFβ significantly reduced the concentration of all three intracellular cytotoxic molecules. In our patients we observe a wide range of perforin expression from 0.5 to 32.6% as shown in this study and in our previous publication11. The representative plots presented in Figs. 2 and 4 are from two different patients. Figure 2 is from a patient with low perforin expression, while Fig. 4 is from a patient with high perforin expression. These profiles are examples that reflect the range in expression profiles between patients. We also found that TGFβ reduced the percentage of total CD8+T cells that were perforin positive, but had no effect on the percentage of cells expressing GZA and GZB (Fig. 4c). We next investigated the effect of TGFβ on cytotoxic molecules in CD103+and CD103−CD8+T cell compartments (Fig. 5a). As seen in Fig. 5b, TGFβ significantly reduced the expression of perforin, GZA and GZB within both CD103+and CD103−CD8+T cell populations. Interestingly, TGFβ reduced the number of perforin positive cells but had no effect on cells positive for GZA or GZB (Fig. 5c).

Progesterone increases the basolateral secretion of TGFβ by polarized EM epithelial cells

In addition to their directs effects, sex hormones are known to indirectly regulate immune cells by acting on adjacent cells in the tissue environment. TGFβ is known to be synthesized and secreted by EM epithelial cells and stromal cells from the uterus18,19. Our studies in a rodent model demonstrated that polarized EM epithelial cells preferentially secrete biologically active TGFβ into the basolateral compartment18. Therefore, we analyzed TGFβ production by polarized human EM epithelial cells. As seen in Fig. 6a, TGFβ was secreted constitutively by polarized epithelial cells with preferential release basolaterally, in the direction of underlying tissues. Since we have previously demonstrated that human polarized epithelial cells are hormonally responsive to E221, we hypothesized that E2 might increase TGFβ secretion by epithelial cells as a secondary mechanism for regulating CD8+T cell cytotoxic activity. To test this, polarized EM epithelial cells were grown to confluence and incubated with E2 or P for 48 h prior to collection of basolateral media. Unexpectedly, as shown in a representative experiment (Fig. 6b), we found that P increased the basolateral secretion of TGFβ relative to controls and that E2 had no effect on TGFβ production. To more fully analyze this response, we normalized the data from 6 experiments to express TGFβ production as the fold change relative to control cells, to compare experiments with different secretion backgrounds by control cells. As shown in (Fig. 6c), following normalization, P but not E2 consistently stimulated the secretion of TGFβ by EM epithelial cells. Overall, these studies suggest that during the secretory phase menstrual cycle, under the influence of progesterone, EM epithelial cells secrete increased amounts of TGFβ basolaterally into the tissue compartment to suppress CD8+T cell cytolytic activity.

Our study demonstrates that E2 and TGFβ directly suppress EM CD8+T cell cytotoxic activity by downregulating intracellular perforin, GZA, or GZB expression. Further, we demonstrate that whereas E2 acts directly on CD8+T cells to suppress killing, P has no direct effect. However, P upregulates TGFβ secretion by EM epithelial cells, which could represent an indirect mechanism of control of cytotoxic activity in vivo. Overall, as depicted in Fig. 7, our findings demonstrate direct and indirect hormone-dependent mechanisms in the human endometrium able to suppress CD8+T cell cytolytic capacity in premenopausal women during the secretory phase of the menstrual cycle, without which successful implantation and pregnancy would be unlikely to occur.

CD8+T cells are key effector cells for immune protection, but their function in the endometrium is tightly regulated to avoid rejection of the semi-allogeneic fetus1,24. Our previous studies demonstrate that EM CD8+T cell cytotoxic activity is suppressed in premenopausal women compared to postmenopausal women, particularly during the secretory phase of the menstrual cycle, suggesting that sex hormones are key regulators of CD8+T cell cytotoxicity10,11. However, the direct effect of sex hormones in suppression of cytotoxic function of EM CD8+T cells had never been addressed before. Therefore, we investigated whether sex hormones directly contribute to the suppression of CD8+T cell cytotoxic activity seen in premenopausal women.

We found that E2 directly suppresses cytotoxic activity of EM CD8+T cells, and that the reduction in killing capacity is due to decreased expression of the cytotoxic molecules perforin and GZA in CD8+T cells. In contrast to our studies which showed no effect of E2 on GZB expression, Navarro et al.14 found that E2 increased the expression of GZB in blood. Differences observed are most likely due to their use of buffy-coat blood suspension containing CD8+T cells that were retrovirally transduced and activated, whereas we used purified EM CD8+T cells without any in vitro activation stimulus to measure the effects of E2. Furthermore, to test GZB production, they used a preparation of total T cells, containing both CD4+and CD8+T cells, while we used purified CD8+T cells. Whether activation status alters EM CD8+T cell responsiveness to E2 remains to be determined.

In contrast to E2, we found that P had no direct effect on EM CD8+T cell cytotoxic activity or cytotoxic granule content. This lack of effect is probably due to the previously reported absence of nuclear P receptors in CD8+T cells25 and contrasts with findings showing P changes in cytokine production and reduction in T-cell proliferation following P treatment of peripheral blood CD8+T cells, which the authors attributed to P membrane receptors15. These differences suggest a finite specificity of P effects that are unique to the class of P receptor present or to functional differences between peripheral blood CD8+T cells and EM CD8+T cells.

Confirming our previous results, we found that EM cytotoxic function was rapidly suppressed by TGFβ after short-time treatment11. In the present study, we demonstrate that the suppressive effects of TFGβ are due to a reduction in the content of intracellular cytotoxic molecules (perforin, GZA and GZB). Given that in our experimental setting, cells were pre-treated with TGFβ for 2 h prior to cytotoxicity analysis and 48 h prior to intracellular cytotoxic molecule assessment, the underlying mechanism for rapid regulation of cytotoxic capacity remains unclear, and further investigation is warranted. Others have demonstrated that perforin and granzymes work synergistically to induce apoptosis in target cells recognized by CD8+T cells26. In addition to reductions in GZA and GZB, our findings show a reduction in both the levels of intracellular content of perforin as well as the percent positive cells. This is consistent with reports by others showing that cytotoxic function is suppressed by TGFβ through the reduction of perforin positive CD8+T cell numbers in the gastrointestinal tract17, and may indicate a general mechanism by which TGFβ controls cytotoxic activity of CD8+T cells in tissues. To the best of our knowledge, our findings are the first to demonstrate that E2and TGFβ in culture reduce both the number of percent perforin positive CD8+T cells as well as the intracellular content of cytotoxic molecules of these cells.

We and others have demonstrated that TGFβ is produced by epithelial cells and fibroblasts from the FRT and under hormonal control in a rodent model18,19. Interestingly, in the human, TGFβ mRNA expression in the endometrium is increased during the secretory phase of the menstrual cycle27, when EM CD8+T cells cytotoxic activity is maximally suppressed10,11. Our current study extends these findings by demonstrating that P acts directly on human EM epithelial cells to stimulate the secretion of TGFβ. Interestingly, we found that E2 had no effect on epithelial cell TGFβ production, in contrast to our previous findings with the rodent model18. Consistent with our results, early studies in non-human primates also found that P increased mRNA of TGFβ2 in the endometrium28. We have previously demonstrated broad effects of E2 treatment on epithelial cells, including stimulation of intracellular genes, such as Cytosolic 5′-nucleotidase and P receptor29, and the secretion of SLPI along with other antimicrobial peptides with broad antimicrobial, antiviral and antifungal activity21,30. Our findings in the present study highlight the selectivity of hormonal effects on epithelial cells. Whether after its secretion in response to P, other hormonally dependent steps further process TGFβ to increase biological activity remains to be determined31,32. Taken together, these findings suggest E2 and P precisely regulate adaptive immunity in the endometrium, by acting indirectly on epithelial cells to increase the production of TGFβ.

The menstrual cycle is divided into a proliferative and a secretory phase, during which fertilization and implantation occurs. Others have shown that shortly after ovulation, during the secretory phase of the menstrual cycle, levels of P rise sharply and are sustained for 10–12 days33. Accompanying this increase in P is a secondary mid-secretory rise in E2 which is essential for implantation. In contrast, the proliferative phase of the cycle is dominated by E2. These hormonal changes control the cyclical remodeling of the endometrium in premenopausal women to optimize conditions for implantation. Based on our findings in the present study, we propose that in premenopausal women, E2 directly suppresses cytotoxic activity by EM CD8+T cells and reduces the intracellular content of perforin and GZA. With ovulation and unique to the endometrium, P induces the production of TGFβ by epithelial cells, which in turn acts on EM CD8+T cells to further suppress cytotoxic activity, and reduce the expression of perforin, GZA and GZB, achieving maximum suppression during the secretory phase of the menstrual cycle (Fig. 7).

In our previous studies, we demonstrated that cytotoxic activity is significantly increased in postmenopausal women10,11. Following menopause, sex hormones levels (E2 and P) drop and remain low and constant. Our results in this study suggest that the reduction in sex hormones removes the direct and indirect suppressive mechanisms from EM CD8+T cells, resulting in an increase in cytotoxic activity and intracellular content of cytotoxic molecules, consistent with our previous findings11.

We and others have shown that CD8+T cells in the endometrium consist of both CD103+and CD103−T cells, likely representing tissue resident and non-resident memory T cells respectively. We demonstrated that cytotoxic activity and intracellular content of cytotoxic molecules was significantly higher in CD103−compared to CD103+EM CD8+T cells11. Here we investigated whether the regulatory effects of E2 and TGFβ were preferential to either subset. We found that E2 and TGFβ affected CD103+and CD103−CD8+T cells equally and reduced the content of cytotoxic molecules in both subsets. These findings suggest that tissue-resident and non-resident CD8+T cells are equally susceptible to regulation by sex hormones, and that transition into tissue-residency status does not modify sensitivity to regulation by sex hormones and the tissue environment. Our results here are also in agreement with our previous observation that postmenopausal women tended to have decreased expression of GZB and increased expression of GZA compared to premenopausal women in both CD103−and CD103+CD8+T cells11. This would suggest that removal of sex hormone suppression after menopause induces changes in intracellular content of cytotoxic molecules in both cellular subsets.

Given that cytotoxic activity following menses returns during the follicular phase of the menstrual cycle10,11, suggests either that the absence of P and TGFβ reverses intracellular granular content and cytolytic activity or that a new population of CD8+T cells enters the endometrium. That cytolytic potential can be restored was suggested from studies in which in vitro neutralization of TGFβ partially restored perforin expression in gut CD8+T cells17. Moreover, using an inhibitor that blocks TGFβ signaling in EM CD8+T cells, we rescued cytotoxic activity in some samples11. Whether cytolytic potential is restored as hormone levels change with the menstrual cycle or the result of recruitment of CD8+T cells to the endometrium with the start of a new menstrual cycle remains to be determined. Further elucidation of EM cytotoxic T lymphocyte (CTL) activity and its implications for successful reproduction and protection against sexually transmitted diseases, including HIV, will provide essential information for the development of vaccines and immunotherapies targeted to the FRT.

Regarding study limitations, in addition to the mechanisms addressed in our study, alternative mechanisms of regulation of cytotoxic activity by sex hormones remain to be explored. For example, while we detected significantly lower levels of cytotoxic molecules following E2treatment, whether the observed effects were due to direct actions of E2 on the production of cytotoxic molecules or secondary to decreased activation status of CD8+T cells remains to be determined. Further, in addition to the perforin/granzyme pathway, the FAS/FASL pathway also mediates cell death34. In preliminary studies we have observed no modification of sFASL production by EM CD8+T cells after incubation with hormones under resting conditions. Further research is needed to address whether the FAS/FASL pathway is significantly modified by E2 after T cell stimulation, and if so, the potential contribution of the FAS/FASL pathway to cytotoxic function of EM CD8+T cells.

Another important aspect not addressed in this study is the extent to which sex hormones modify degranulation of T cells. In previous studies, we demonstrated a significant increase in degranulation capacity of CD8+T cells from the FRT after menopause11, suggesting a potential implication of sex hormones in regulation of degranulation capacity. Future studies are needed to investigate the extent to which E2 and P directly modify degranulation of CD8+T cells. Furthermore, in addition to cytotoxic molecules, FRT CD8+T cells can secrete and array of cytokines and chemokines upon stimulation, including IFN-γ and TNFα35. In preliminary studies, we observed no differences in secretion of IFN-γ and TNFα in CD8+T cells treated with E2 or P under resting conditions; however, whether cytokine production after stimulation is modified by sex hormones remains to be addressed in future studies.

Although our study is focused on CD8+T cells, a full understanding of the regulation of cytotoxic activity in the FRT during the menstrual cycle will require the functional evaluation of CD4+T cells and NK cells. We have previously demonstrated decreased proportions of CD4+T cells in the EM from postmenopausal women9. However, whether E2 and P modify the presence and function of EM CD4+T cells throughout the menstrual cycle is unknown. Endometrial NK cells, which have overlapping cytotoxic functions with CD8+T cells, increase in numbers from approximately 3% of leukocytes in proliferative phase to 30% in secretory phase tissues36,37,38 where they surround the spiral arteries39, due to increased chemotaxis and proliferation. Whether sex hormones directly influence cytotoxic function of endometrial NK cells is unclear. Previous studies using peripheral blood NK cells have shown contradictory evidence of a menstrual cycle effect on cytotoxicity40,41. However, others have demonstrated that endometrial NK cells in the early proliferative phase have lower cytotoxicity compared to other stages of the menstrual cycle42. Together these findings suggest that sex hormones modulate cytotoxic activity of multiple cell types as part of a larger program to optimize conditions in the FRT for successful reproduction.

In conclusion, our study demonstrates two unique hormone-dependent regulatory mechanisms that regulate CD8+T cell cytotoxic function in the uterine endometrium, one which involves the direct effect of E2 and the other, the indirect effect of P-mediated epithelial cell production of TGFβ. In response to E2 and TGFβ, CD103+and CD103−CD8+T cell cytolytic potential is suppressed. These results are central to our understanding of hormonal regulation of cytotoxic activity during the menstrual cycle, during pregnancy and following menopause, and to the development of therapeutics for both the prevention and control sexually transmitted infections and gynecological cancers.