Recent bioinformatic analyses looking at several CD8 T cell subtypes found changes in TN, TCM, and TEM cell scores between the eutopic endometrium of patients and controls (Table 3). However, 5 out of 9 papers reported no difference in the concentration of the total CD8 T cell population between patients with endometriosis and healthy controls.

Using gene set enrichment analysis (GSEA) on data from eutopic endometrial biopsies taken during the proliferative phase with CIBERSORT, Zhong et al. found no difference in CD8 T cell fraction between patients with severe endometriosis and controls (2). Nevertheless, in a similar analysis, investigators found increased CD8 T_EM cell gene enrichment scores in women with revised American Society for Reproductive Medicine (rASRM) stage I/II as well as stage III/IV disease (p = 0.000 and p = 0.027, respectively) and decreased CD8 T_N and T_CM scores (p = 0.001 and p = 0.000; p = 0.035 and p = 0.782) when compared to controls (27). Another GSEA by a different group shows an increased CD8 T cell score in endometriosis versus controls when analysed using CIBERSORT (0.1659 ± 0.0968 vs 0.1086 ± 0.0939, respectively; p = 0.0033), MCP-counter (p = 0.00018) and ImmuneCellAI (p = 7.2e-8) (28). Comparing different subsets with ImmuneCellAI, decreased CD8 T_N cells (p = 7.6e-5), T_CM (p = 0.00076), and T_EM cells (p = 1.1e-6) were present in the endometrium of women with endometriosis when compared with controls. In a second analysis (GSE120103), performed with MCP-counter, increased CD8 T cell scores were revealed in endometrium of infertile versus fertile women with endometriosis (p = 0.0019, AUC (area under the receiver operating characteristic curve (ROC)) = 0.722). Immunohistochemical validation of their transcriptomic data confirmed increased proportions of CD8+ cells among all leukocytes in the eutopic endometrium of endometriosis patients versus controls in both the proliferative and early secretory phases (0.2292 ± 0.0591 vs 0.1790 ± 0.0562, p = 0.0132, AUC = 0.727) (28).

Similarly, increased quantity of CD8+ cells per mm² was detected in severe endometriosis versus controls, of note samples were collected in the proliferative phase (0.019 ± 0.004 vs 0.007 ± 0.002 cell/mm², respectively; p < 0.01) (29). Focusing on the eutopic endometrial tissue in the secretory phase, single-cell RNA sequencing analysis did not reveal any difference in the expression of the CD8A gene between endometriosis and controls (30). Again, no difference was found in the microscopic analysis of the secretory phase tissue from infertile women with endometriosis and fertile controls (31). In samples equally distributed across the uterine cycle, no difference was found in either glandular or intraepithelial CD8+ cell count in either of the uterine cycle phases between endometriosis and controls (32), or between endometriosis and controls with laparoscopically proven tubal factor infertility resulting from prior pelvic inflammatory disease (PID) (33).

Finally, in the findings from Bunis et al., the endometrial CD8 T_EM signature had higher gene enrichment scores across all menstrual phases in endometriosis patients when compared to the control (34). Additionally, they were significantly increased in the mid-secretory phase when compared to the preceding phases in both disease and control (NES = 1 for unstratified samples). The activation potential of eutopic endometrial CD8 T cells was decreased in the endometriosis patients when compared to controls and genes with functions known to be relevant to phagocytosis and complement activation were stimulated in the eutopic endometrium of patients. No papers were found assessing the proliferative and cytotoxic capacities of eutopic endometrial CD8 T cells in endometriosis.

All 5 studies researching ectopic endometrium reported an increased proportion of CD8 T cells in endometriotic lesions when compared to matched eutopic endometrium and/or healthy eutopic endometrium (Table 4). Additionally, 2 studies also investigated their activation and cytotoxic properties.

Zhong et al. performed transcriptome meta-analysis with CIBERSORT on 24 ectopic, 82 eutopic and 68 healthy endometrial samples in the proliferative phase (GSE120103, GSE51981, GSE25628, GSE37837, GSE7846, GSE6364, GSE7305) (2). They described an increased CD8 T cell fraction among all infiltrating immune cells in the endometriosis lesions when compared to the eutopic endometrium of endometriosis patients (p = 0.085) and endometrium of controls (p = 0.033). Similarly, analyzing scRNA-seq data of proliferative phase timed samples, a significantly higher percentage of CD8 T_N cells was reported in endometriotic lesions when compared to matched eutopic endometrium and normal endometrium (43.48% vs 0.91% and 0.33%, respectively), and a lower proportion of T_E subtypes (35). In addition, the authors describe decreased cytotoxic T cell populations in lesions, however, data was not shown. Interestingly, a smaller immunohistochemistry study on proliferative phase samples confirmed increased CD8+ cell counts in ectopic endometrium when compared to the control endometrium (0.019 ± 0.002 vs 0.007 ± 0.002 per mm²; p < 0.01), but not in ectopic versus matched eutopic endometrium (29).

Investigating samples in the secretory phase of the menstrual cycle, genes associated with several immune cell surface markers, including CD8A, were upregulated in ectopic tissues compared with eutopic endometrium of patients as well as control endometrium (p < 0.05) (30). In the same study, they also found increased expression of immune genes for T cell co-stimulation (e. g. CD27 and CD28) as well as less specific HLA genes associated with T cell activation, such as HLA-DPA1, HLA-DPB1, HLA-DQA1, HLA-DRA, and HLA-DMA (p < 0.05). On the other hand, a different IHC study found no difference in the CD8+ cell count between the glandular epithelium of ectopic lesions and matched eutopic endometrium in the proliferative and late secretory phase but confirmed an increased count in the early secretory phase (p = 0.0002) (32).

Of the 9 studies researching CD8 T cell ratios in the peritoneal fluid, 3 found no difference between endometriosis patients and controls, 1 paper noted no difference in CD8 T cell fraction among lymphocytes but increased proportions among all peritoneal cells in patients with endometriosis and 2 found increased concentrations in the disease (Table 5). Similarly, as shown, no consensus was reached on the activation status of the peritoneal CD8 T cells: 3 studies found no difference, 2 reported decreased and 1 increased activation in endometriosis patients.

The proportion of CD8 T cells among total mononuclear cells, between patients with rASRM stage I/II endometriosis and fertile controls are unchanged in the proliferative and secretory phases as well as when uncontrolled for the cycle phase (36). Similarly, some authors found comparable rates of CD8 T cells among mononuclear cells in women with and without the disease in the proliferative phase (37) and others reported no difference in the percentage of CD8+ cells among all lymphocytes in the periovulatory phase between endometriosis patients and healthy controls (38). While Deng et al. identified no difference in the proportion of CD8+ cells among lymphocytes in the proliferative phase, they found increased numbers of CD8+ cells among all peritoneal fluid cells in samples from patients with endometriosis (p = 0.0032) (39).

Increased proportions of specific CD8 T cell subsets were reported in studies studying baboons and MAIT cells. In baboons, a significant increase in the proportion of peritoneal CD8 T cells was detected in the group with long-term spontaneous endometriosis, when compared to other groups with no disease, induced disease, and recent spontaneous endometriosis (74.3 ± 6% vs 67.1 ± 8.3% vs 65.8 ± 8% vs 54.8 ± 21.9%, respectively; p = 0.01) (40). In humans among MAIT cells defined as CD3+CD161+Vα7.2+ enhanced CD8+ MAIT cell subsets were found in the disease compared to control (6.52 ± 1.05 vs 2.77 ± 0.42, respectively; p = 0.0454) (41). In the same study, peritoneal fluid CD8+ MAIT cells from endometriosis patients displayed higher levels of the activation marker CD38, when compared to non-endometriosis participants (2.98 ± 0.53 vs 0.36 ± 0.09, respectively; p = 0.0071). The difference was more pronounced in the group with stages I/II than with stages III/IV when compared to controls.

On the other hand, Gallinelli et al. showed decreased proportions of CD3+CD8+HLA-DR+ cells in the endometriosis versus the control group (42.8 ± 6.8 vs 57.1 ± 7.6, respectively; p < 0.02) (42). Additionally, they found a significantly higher CD8/CD4 ratio in the disease (2.8 ± 0.5 vs 2.1 ± 0.4, respectively; p < 0.05). Nevertheless, 2 publications report no difference in activation potential between endometriosis and non-endometriosis groups (38, 43). In a study from Ho et al., they found no differences between stage I/II and III/IV endometriosis and controls in any of the identified activated phenotypes, including HLA-DR+, CD25+, CD28+, and CD69+ CD8 T cells (43).

Schmitz et al. investigated the variance in immune cells between menstrual effluent of patients with endometriosis and healthy controls (44). They found no statistically significant differences in either the proportion of CD8 T cells or the CD4/CD8 ratio. However, a decreased proportion of perforin-high CD8 T cells was present in the menstrual effluent of patients versus controls (4.0 ± 4.8 vs 11.2 ± 10.4, respectively; p = 0.029). In the study of Shih et al. four distinct CD8 T cell clusters were differentially enriched between patients with surgically confirmed endometriosis and community controls, although the differential markers for those four clusters have not been explored (45).

The eutopic endometrium is an important mucosal barrier and the location of embryo implantation. Not only is immune dysregulation in eutopic endometrium associated with subfertility it may also play a role in lesion establishment (77, 78). Therefore, along with other immune cell populations, CD8 T cells are an important consideration in endometriosis. Although the data regarding eutopic endometrial CD8 T cells in endometriosis altogether are controversial, our systematic review gives an overview not only of the whole CD8 T cell population, but of alterations in other CD8 T cell subtypes in the disease. CD8 TN and TCM cell gene enrichment scores were found to be decreased while TEM cells seem to be increased (27, 28, 34). There is increasing evidence demonstrating an association between endometriosis and autoimmunity and CD8 T_EM cells are likely an important player in the pathogenesis of autoimmune diseases due to their effectiveness and durability (72). Nevertheless, evidence is not uniform in other autoimmune disease where for example, CD8 T_EM cells are increased in the cerebrospinal fluid of patients with multiple sclerosis. However the opposite was found in the gut of patients with inflammatory bowel disease (79, 80). More research is certainly needed to ascertain their role in the pathogenesis of endometriosis (81).

T_RM are thought to be the most prevalent CD8 T subpopulation in mucosal tissues but our understanding of this subtype in the human endometrium is poor (12). T_RM in the gut are widely researched and there is strong evidence they are implicated in the pathogenesis of inflammatory bowel disease (IBD) and other intestinal immunopathologies such as coeliac disease (13, 82). For example, authors investigating T_RM in IBD show considerably lower levels of CD103+ T_RM in patients with flares versus patients with endoscopic remission and healthy controls, which suggests their contribution to tissue homeostasis and immune regulation (83). Higher levels of CD69+CD103+ T_RM were associated with active disease in patients with vitiligo, and importantly, these cells can attract cytotoxic effector CD8 T cells from the blood, which is critical for disease persistence (84). In other gynaecological pathologies, authors revealed a modified eutopic endometrial CD8 T_RM phenotype in recurrent miscarriage with considerably lower expression of tissue residency marker CD69 (85).

In adenomyosis, CD8 T cell subsets are increased not only in the eutopic, but also in the ectopic endometrium of patients compared to healthy controls (86). Similar seems to be true for endometriosis, although there are discrepancies in the results between the lesions and matched eutopic endometrium. However, we emphasize that studies reporting no difference in CD8 T cells between eutopic and ectopic endometrium did not investigate specific phenotypes (2, 29). Although the potential roles of CD8 T cells in lesions are not fully defined, some parallels may be drawn from cancer research, where CD8 T cells are regarded as the most potent effectors in the anti-cancer immune response and form the basis of some current cancer immunotherapies (87). For example, increased intraepithelial CD8 T cell count in the tumour microenvironment is associated with prolonged survival in colorectal, ovarian and endometrial cancer (88–90) and the intraepithelial subpopulation of CD8 T cells was found to be increased in ectopic lesions in the early secretory phase when compared to matched eutopic and healthy endometrium (32). However, no studies were conducted comparing intraepithelial CD8 T cells between lesions of patients with different endometriosis stages and between those who benefited from the surgery or relapsed, which would give us valuable insights into the prognostic role of this population.

It has been suggested that endometriotic lesions ought to be investigated along with peritoneal fluid, since this dynamic environment strongly influences their pathophysiology (91). Nevertheless, to date, studies have examined lesions and peritoneal fluid in isolation. Similar to eutopic endometrium, data on peritoneal CD8 T cells are contradictory and no definitive conclusions can be made about the role of the CD8 T cell population in this regard. On the other hand, macrophages are a well-researched population in this setting as they are the most abundant peritoneal immune cell population, representing around 50% of peritoneal cavity leukocytes (92). They are required for lesion growth, development, vascularization, and innervation, and are even associated with pain symptoms. In endometriosis, an increased ratio of altered peritoneal macrophages has been indicated, and since they are involved in antigen presentation to CD8 T cells, this points to potentially related dysfunctions within the CD8 T cell compartment. Similarly, changes were found in regulatory T cells (Tregs) in endometriosis, with findings showing an increased number of activated Tregs in the peritoneal fluid of patients with endometriosis (93). Like macrophages, Tregs are involved in CD8 T cell activation and play an important role in the generation of high-avidity initial responses and effective memory (94).

One aim of this review was to gain a better understanding of the activation characteristics of the CD8 T cells in endometriosis. Unfortunately, no functional analyses have been performed to date on either eutopic or ectopic endometrial CD8 T cells, despite clear evidence of endometriosis-related dysfunctions in key regulators of CD8 T function, such as programmed cell death protein 1 (PD-1), programmed death-ligand 1 (PD-L1) and CD4+ T regulatory cells (29, 53, 95). Functional analyses of CD8 T cells in peritoneal fluid demonstrate inconsistent findings, where data mostly report no difference or decreased activation potential of peritoneal CD8 T cells with only one study reporting an increase in activated CD8+ MAIT cells. There is some evidence of upregulated activation marker CD69 expression on peritoneal fluid T cells in endometriosis patients, although it is important to emphasize that the investigated cohort had been taking hormonal treatment hence this study was not included in our systematic review (19).

Studies on immune cells in the female genital tract are challenging given the invasive nature of collecting mucosal and peritoneal samples (96). As previously mentioned, most research depends on biopsies and hysterectomies from women with underlying pathologies and samples from healthy women are limited. Menstrual effluent contains cells of endometrial origin and could be a valuable non-invasive source of endometrial immune cells, but has so far been overlooked in endometriosis research (44). Authors found decreased expression of perforin in CD8 T cells of patients when compared with controls and argued that perforin-mediated cytotoxicity may play a crucial role in the establishment of the lesions, drawing from research on CD8 T cells in susceptibility and elimination of cancer cells. Importantly, this could provide some answers to why only a few women develop endometriosis while retrograde menstruation occurs in most females of reproductive age (97). However, further investigation of menstrual effluent is required to understand the processes responsible for endometrial cell attachment and development outside of the endometrium, especially given the key differences in immune milieu between eutopic and ectopic endometrium and peritoneal fluid which we have confirmed in our review.

The ease of sampling makes peripheral blood one of the most studied tissues in endometriosis leading to valuable insights into the systemic nature of endometriosis and its association with autoimmune diseases, including abnormalities in the T cell population (72, 93, 98). Peripheral blood immune cells are influenced by sex hormones, with higher CD8 T cell numbers in males, but greater activation, proliferation and cytotoxic capacity in females (99). Although longitudinal studies investigating CD8 T in the peripheral blood at different menstrual phases are limited, modifications in Treg cells (100) and Th1/Th2 ratios (101) indicate that it is important to consider the menstrual cycle when investigating CD8 T population in the peripheral blood. However, this was not the case in some of the included studies. While most of our data does not suggest discrepancies in the peripheral CD8 T cells, it is important to highlight that most papers did not describe CD8 T cell subpopulations. They were shown to be significantly changed in patients in one of the recent analyses with increased central memory and decreased effector memory subtypes in the disease (46). Interestingly, several authors confirmed a decreased frequency of peripheral effector memory CD8 T cells in patients with multiple sclerosis at the onset of the disease and throughout its clinical course, which underlines its importance in the initiation of the disease rather than being its consequence (102). Additionally, there is evidence of changes in innate immune cells in the peripheral blood of patients with endometriosis, such as monocytes, which are important for antigen presentation and CD8 T cell activation (103). Reduced numbers of classical and intermediate monocytes were found in the peripheral blood of endometriosis patients, while the levels of plasmacytoid dendritic cells and non-classical monocytes were increased. Nevertheless, to better understand the systemic role of CD8 T cells in endometriosis, detailed phenotypic studies are needed, which will address specific subtypes and their activation characteristics found to be changed in the peripheral blood of other immune diseases (84).

Although the pathophysiology of endometriosis-associated subfertility remains uncertain, several findings support altered endometrial immune receptivity as one of the possible causes (77, 104). The exact nature of CD8 T in eutopic endometrium is poorly understood, but the suggested hormonal influence on LA formation and CD8 T cell cytotoxicity underline the importance of optimal regulation during implantation and pregnancy (5, 68). As noted in our review, authors found significantly increased CD8 T cell scores in infertile women with endometriosis when compared to fertile women as well as in infertile versus fertile women with or without endometriosis (28). Despite the limited sample size of this GSEA analysis, it reveals possible changes in CD8 T cells in patients with endometriosis-associated infertility and urges for more investigation into the phenotypic characterisation of the population in this related disorder. It is likely that different CD8 T phenotypes could contribute to subfertility and pregnancy complications as their findings correspond to the results of several authors, who confirmed altered endometrial CD8 T characteristics in endometrial biopsies obtained during the window of implantation from patients with recurrent pregnancy loss, pre-eclampsia, and preterm labour (85, 105–107).