Introduction

Mucosa-associated invariant T (MAIT) cells are a subset of non-conventional evolutionary conserved innate-like T cells that are found in both peripheral blood and mucosal tissues.1 Human MAIT cells express a semi-invariant T-cell receptor (TCR) made up of an invariant Vα7.2 coupled with restricted Jα segments (Jα33, Jα12, or Jα20) and limited Vβ repertoires.2,3,4,5 They develop in the thymus upon interaction with MHC class 1-related protein (MR1) expressed on thymocytes.3,4,6 Yeast and certain bacterial species produce metabolites from the riboflavin (vitamin B2) pathway which are presented on MR1 and can activate MAIT cells in a MR1-restricted manner.7 However, MAIT cells are also able to respond to non-TCR signals, especially in the form of inflammatory cytokines, principally during viral infections.8,9,10

MAIT cells located in tissue are functionally and phenotypically different to their peripheral blood counterparts, but the relationship between potentially tissue-resident and circulating MAIT cells is poorly understood. Salou et al. recently showed that MAIT cell subsets express gene signatures associated with tissue-residency.11 Furthermore, in parabiosis experiments it was demonstrated that murine MAIT cells are resident in lungs, liver, and spleen regardless of whether they are located inside or outside of the vasculature.11 Several studies of human MAIT cells in tissue have confirmed that they express tissue-residency markers such as CD69 and CD103.12,13,14 However, mechanistic studies assessing MAIT cell tissue-residency in the human setting are lacking.

The immune system of the endometrium contributes to successful pregnancy outcome.15,16 The endometrial mucosa is also an important site for immune defense against microbes where it provides a barrier composed of an epithelium covered by mucus and hosts numerous effector cells belonging to the innate and adaptive arms of the immune system.17,18 Tissue-specific pathogens that the genital mucosa needs to defend the host against are for example sexually transmitted bacteria such as Neisseria gonorrhoeae (N. gonorrhoeae), Chlamydia trachomatis, and Treponema pallidum. N. gonorrhoeae is the etiological agent of gonorrheae. The pathology and symptoms of infection are mainly caused by the activation of innate immune responses at the site of infection. Untreated infection in women can result in complications such as infertility and ectopic pregnancy. In addition, it might lead to a disseminated gonococcal infection, potentially giving rise to infectious endocarditis and arthritis.19 WHO has identified clinical strains of N. gonorrhoeae that are resistant to most available antibiotics and highlighted the possibility of widespread untreatable gonorrhea to arise in the near future.

Although MAIT cells have been reported to localize to the female genital tract,17,20 many details regarding their function and tissue residency remain elusive. Furthermore, the potential role of endometrial MAIT (eMAIT) cells as a tissue-resident defense against microbes is largely unexplored. Here, we characterize eMAIT cells during several biological phases of the endometrium, in pre-, and postmenopausal, and pregnant women. We further utilize human uterus transplantation as a model to demonstrate that, in contrast to murine models of parabiosis, human MAIT cells are only transiently tissue-resident in the uterus despite expressing markers of tissue-residency. In addition, we show that MAIT cells are able to respond to the clinically relevant pathogen N. gonorrhoeae. Altogether, our results point to MAIT cells as an active component of the endometrial mucosal defense, provide novel insights into the basic immunology of tissue-resident MAIT cells, and new knowledge into their antimicrobial function.

Results

eMAIT cells represent a stable immune cell population not responding to local regeneration

Flow cytometry was used to determine frequencies of MAIT cells, defined as CD3+CD4-CD161+Vα7.2+, in endometrium and matched peripheral blood (Fig. 1a, b). In endometrium from women of fertile age, the eMAIT cell frequency was similar to matched peripheral blood with a trend towards slightly lower levels of MAIT cells in endometrium as compared to blood (Fig. 1a). These results were in line with a previous report on MAIT cell frequencies in the female genital tract and peripheral blood.20 MAIT cells can be classified as either CD8+ or double-negative for CD4 and CD8 (DN), and similar to peripheral blood, the majority of eMAIT cells were CD8+ (Fig. 1c). The frequency of CD8 and DN MAIT cells was similar in postmenopausal endometrium, as well as in decidua (Supplementary Fig. 1a). To verify our strategy whereby eMAIT cells are identified based on CD161 and Vα7.2 together with exclusion of CD4, results were compared with eMAIT cell identification using the MR1-5OP-RU tetramer and this analysis revealed similar frequencies of eMAIT cells using the two approaches (Fig. 1d).

Fig. 1: MAIT cells are present in endometrium but unaffected by endometrial regeneration.
figure 1

a Representative flow cytometry plots for identification of CD3+CD161+Vα7.2+CD4- cells in endometrium. b Frequency of MAIT cells out of total T cells in fertile endometrium (endo) and matched peripheral blood. c Frequency of CD8+ and DN MAIT cells out of total MAIT cells in endometrium and matched peripheral blood (n = 19). d Representative staining for frequencies of MAIT cells identified as CD3+CD161+Va7.2+ cells compared to with tetramer staining (MR1 5-OP-RU) in one representative endometrial sample. e Frequency of eMAIT cells out of total T cells in matched biopsies taken at day 7 and 21 of the menstrual cycle (n = 9). f Frequency of MAIT cells out of total T cells in postmenopausal (PMP) endometrium, matched PBMC (n = 7), and decidua (n = 3). MAIT cells from fertile endometrium (red, Endo) is shown as a reference. g Frequency of MAIT cells out of total T cells in peripheral blood at five-time points during the menstrual cycle (n = 11). h Correlation of peripheral blood and menstrual blood MAIT cell frequencies in monozygotic twin pairs (n = 8 and n = 6). n.s. not significant.

Full size image

Next, we assessed dynamics within the eMAIT cell population in response to endometrial regeneration (menstrual cycle dynamics) and atrophy (post menopause) as well as in response to pregnancy (Fig. 1f). The frequency of eMAIT cells was similar in the early (follicular phase, day 7) compared to late phase (luteal phase, day 21) of the menstrual cycle (Fig. 1e). Similarly, no difference in Ki67-expression was noted in eMAIT cells early and late in the cycle, suggesting lack of a proliferation during the menstrual cycle (data not shown). Furthermore, eMAIT cells were retained at similar levels in post-menopausal endometrium as in fertile endometrium (Fig. 1f). Finally, also in first-trimester decidua, eMAIT cells could be identified and their frequency was not significantly different from that of non-pregnant pre-menopausal endometrium (Fig. 1f). In line with this, peripheral blood MAIT cell levels remained strikingly stable throughout the menstrual cycle (Fig. 1g). Both endometrial and decidual MAIT cells expressed PLZF and the signature cytokine receptor IL18Ra similar to peripheral blood MAIT cells (Supplementary Fig. 1b).

In conclusion, MAIT cells are present in endometrium but we find no evidence for an influence of the biological changes occurring during the menstrual cycle, post-menopause nor in pregnancy, on the frequency of endometrial MAIT cells.

Environmental impact on the human MAIT cell compartment assessed in monozygotic twins

Next, we assessed the frequency of MAIT cells in peripheral and menstrual blood from monozygotic twins. We have previously reported that menstrual blood sampled early during menstruation represents a robust and non-invasive method to study endometrial immune cells.21 Eight female monozygotic twin pairs were recruited. All were gynecologically healthy and between 25 and 32 years of age. Interestingly, the frequency of peripheral blood MAIT cells from monozygotic twins correlated to a high degree (Fig. 1h). In contrast, no such correlation was observed for eMAIT cells from monozygotic twin pairs (Fig. 1i). Next, when assessing the relationship between the size of the circulating and the endometrial MAIT cell populations in 19 unrelated premenopausal women (not twins), no correlation was found (data not shown). Taken together, this would suggest that the size of the peripheral blood MAIT cell population might be influenced by genetic factors whereas the endometrial MAIT cell population is more influenced by environmental factors.

Human eMAIT cells represent a transiently tissue-resident population

Although MAIT cells are present in many human peripheral tissues, are known to express tissue-residency markers, and have been shown to be tissue-resident in murine parabiosis experiments,11 this has not been formally assessed in the human setting. To study this, we first investigated whether human eMAIT cells exhibit signs of tissue-residency. Both eMAIT cells from pre- and post-menopausal endometrium displayed higher expression of CD69 and CD103 as compared to matched peripheral blood MAIT cells as well as low but clearly detectable levels of CD49a (Fig. 2a, b, Supplementary Fig. 2a). A similar expression pattern was observed in decidual MAIT cells (Supplementary Fig. 2b). We also assessed tissue-residency markers on CD8 + T cells compared to MAIT cells from the same individuals. The expression of CD49a was significantly higher in CD8 + T cells, both in fertile (*p = 0.031) and postmenopausal (**p = 0.001) endometrium, while the expression of CD69 and CD103 was similar (ns, p > 0.05) (Supplementary Fig. 2c). To more formally assess whether human eMAIT cells are tissue-resident we obtained endometrial tissue from two uterus-transplanted women. Uterus transplantation is an extremely rare surgical procedure with fewer than 30 successful transplantations ever performed world-wide in women.22 In more detail, peripheral blood was obtained from both donors and recipients and endometrial samples were obtained from the transplanted uteri at 23 and 34 months after transplantation, respectively. Since uterus transplantations are performed across HLA barriers, cellular origin (donor or recipient) could be determined using anti-HLA antibodies specific for donor or recipient HLA.23 As expected, donor and recipient peripheral blood MAIT cells showed a HLA-staining pattern in accordance with their HLA statuses (Fig. 2c). eMAIT cells were present in transplanted uteri (data not shown) and displayed a similar phenotype with respect to tissue-residency markers as eMAIT cells from normal fertile endometrium (Fig. 2d). Strikingly, the vast majority, if not all, of eMAIT cells from the transplanted uteri expressed recipient HLA molecules suggesting these to be of recipient origin (Fig. 2d). Indeed, very few, if any, of the original donor tissue-resident cells could be detected (Fig. 2d).

Fig. 2: Endometrial MAIT cells are transiently tissue-resident.
figure 2

a Representative flow cytometry plots for expression of CD49a, CD69, and CD103 on peripheral blood and eMAIT cells. b Summary data for expression of CD49a, CD69, and CD103 on eMAIT from fertile women and matched peripheral blood (n = 7). c Anti-HLA stainings on MAIT cells from donor and recipient peripheral blood as well as from endometrium after uterus transplantation for two patients. d Representative flow cytometry plots for expression of CD69 and CD103 on peripheral blood and eMAIT cells from one patient after uterus transplantation. n.s. not significant; *p < 0.05.

Full size image

Altogether, this shows that, although eMAIT cells express markers of tissue-residency, they most likely represent a transiently tissue-resident immune cell population in human endometrium. This also suggests that the endometrial niche can be replenished by circulating MAIT cells entering the tissue and starting to express tissue-residency markers.

MAIT cells in endometrium exhibit a more activated phenotype compared to peripheral blood

We next characterized the expression of several markers related to cell activation and function (CD38, CD39, HLA-DR, PD-1, CD56, and CD127) (Fig. 3a, b). Analysis revealed that eMAIT cells from pre-menopausal endometrium express the activation and metabolism markers CD38, HLA-DR, and PD-1 at a significantly higher frequency compared to MAIT cells from matched peripheral blood. The expression of the ectoenzyme CD39 trended towards being higher on eMAIT cells. There was no significant difference in expression of CD56 (Fig. 3b). Interestingly, eMAIT cells express CD127 at a lower frequency compared to peripheral blood MAIT cells (Fig. 3b). Next, we compared the expression of these markers in the CD8+ and DN MAIT cell subpopulations. No differences were observed except from a significantly higher expression of HLA-DR on CD8+ MAIT cells compared to in the DN population (Supplementary Fig. 2d). We furthermore investigated the expression of the aforementioned markers in decidua and post-menopausal endometrium, and compared this to the findings from women of fertile age. For most markers, expression was similar, however, HLA-DR was significantly higher on eMAIT cells from pre-menopausal compared to post-menopausal or pregnant women (data not shown). In order to simultaneously interpret differences in phenotype for all studied markers between eMAIT cells and peripheral blood MAIT cells, we performed UMAP analysis (Fig. 3c). Next, PhenoGraph was used to identify cell clusters within the high-dimensional data visualized with UMAP. PhenoGraph clustering identified 12 clusters of MAIT cells (Fig. 3c). Three clusters were highly enriched for eMAIT cells, whereas another three clusters were predominately found in peripheral blood MAIT cells (Fig. 3d). When assessing expression of single markers in relation to the UMAP and PhenoGraph clusters, the difference between eMAIT cells and peripheral MAITs were primarily attributed to higher expression of CD103, CD69, PD-1, and CD49a within eMAIT cells (Fig. 3e).

Fig. 3: Endometrial MAIT cells display an activated phenotype.
figure 3

a Representative flow cytometry plots for expression of indicated markers on peripheral blood and endometrial MAIT cells. b Summary data for expression indicated markers on MAIT cells from peripheral blood and endometrium (n = 7, 3, 7, 18, 18, and 18). c UMAP projection of concatenated MAIT cells from matched peripheral blood and endometrium (n = 7) colored according to the tissue of origin (left plot). PhenoGraph clustering was performed on MAIT cells barcoded and concatenated from all samples. The middle plot depicts all 13 identified PhenoGraph clusters and the right plot the six most common clusters in endometrium overlaid on the UMAP projection. d Relative abundance of endometrial and blood MAIT cells within every detected PhenoGraph cluster. e UMAP plots showing expression intensities of activation and tissue-residency markers on MAIT cells. n.s. not significant; *p < 0.05; ***p < 0.001.

Full size image

In conclusion, we observed eMAIT cells to exhibit a more activated phenotype compared to matched peripheral blood MAIT cells.

MAIT cells respond to the sexually transmitted bacteria N. gonorrhoeae

Next, we assessed whether MAIT cells were able to respond to the clinically relevant pathogen N. gonorrhoeae and compared responses to the more classical stimuli E.coli. To this end, we stimulated peripheral blood MAIT cells with either E.coli or N.gonorrhoeae and measured four functional readouts (Granzyme B, IFNγ, TNF, and CD107a) as well as upregulation of the activation marker CD69. MAIT cells responded to the laboratory strain of N.gonorrhoeae in a similar fashion as to E.coli, albeit with overall lower responses for the measured functions (Fig. 4a–c). When stimulating with different amounts of bacteria and then measuring the MAIT cell response as CD69+Granzyme B+ or CD69+IFNγ+ cells the response to N.gonorrhoeae was found to be dose-dependent (Supplementary Fig. 3a and b). Moreover, in line with previous reports MAIT cell IFNγ, TNF, and CD107a responses towards E.coli were MR-1 dependent (Fig. 4a, b).24 In similar experiments but with the N.gonorrhoeae laboratory strain, the MAIT cell MR-1 dependency was less obvious suggesting cytokine priming of MAIT cells may be more relevant towards this particular strain (Fig. 4c). When stimulating decidual MAIT cells with N.gonorrhoeae we observed a similar response as from peripheral blood MAIT cells (Fig. 4d). The only difference noted was that the background level of activation trended to be higher in decidual MAIT cells with higher basal expression of granzyme B in unstimulated cells (Fig. 4d).

Fig. 4: MAIT cells are activated by N. gonorrhoeae.
figure 4

a Representative flow cytometry plots for the indicated markers on peripheral blood MAIT cells after E.coli stimulation. b Summary data for expression of granzyme B, IFNy, TNF, and CD107a on peripheral blood MAIT cells after stimulation with E.coli with or without MR1 blockade (n = 7). c Summary data for expression of granzyme B, IFNy, TNF, and CD107a on peripheral blood MAIT cells after stimulation with a N. gonorrhoeae lab strain with or without MR1 blockade as compared to unstimulated (n = 6). d Summary data for expression of granzyme B, IFNy, TNF, and CD107a on decidual MAIT cells after stimulation with a N. gonorrhoeae with or without MR1 blockade as compared to unstimulated (n = 4–5). e Summary data for expression of granzyme B, IFNy, TNF, and CD107a on MAIT cells after stimulation with three separate clinical isolates of N. gonorrhoeae with or without MR1 blockade as compared to unstimulated (n = 6). f Summary data for expression of granzyme B, IFNy, TNF, and CD107a on peripheral blood MAIT cells after stimulation with N. gonorrhoeae lab-strains, antibiotic sensitive, and antibiotic-resistant strains (n = 2 each for the strains with n = 6 healthy donor MAIT cells being tested against each strain). n.s. not significant; *p < 0.05; **p < 0.01; ***p < 0.001.

Full size image

We next stimulated MAIT cells with three different clinical isolates of N. gonorrhoeae (Fig. 4e and Supplementary Fig. 3c). MAIT cells were also responding to clinical isolates of N. gonorrhoeae and no clear differences were observed between the three isolates (Fig. 4e and Supplementary Fig. 3d). Instead, the magnitude of the response seemed to be more dependent on the donor than on the particular bacterial strain (Supplementary Fig. 3d). MAIT cell TNF and CD107a responses toward the clinical isolates, but not granzyme B upregulation and IFNγ production, could be blocked with an anti-MR-1 blocking mAb (Fig. 4d). This indicated that the response was driven by both MR1-dependent and independent mechanisms. When assessing responses from MAIT cell subpopulations, we observed a higher production of granzyme B in CD8+ MAIT cells compared to DN cells when stimulated with clinical isolates of N.gonorrhoeae (Supplementary Fig. 3c). These results are in line with a recent study on MAIT cell subpopulation responses towards E. coli.25 Except for MAIT cells, also monocytes, CD8 T cells, and NK cells, but not CD4 T cells, responded after N.gonorrhoeae was added to the cultures (Supplementary Fig. 3f). However, this response was MR-1 independent.

Finally, the MAIT cell response to three laboratory strains, two antibiotic sensitive clinical isolates, and two antibiotic-resistant clinical isolates of N.gonorrhoeae was compared side by side (Fig. 4f). We observed that the response rate to the antibiotic-sensitive clinical isolates was significantly higher compared to that against both the laboratory strains and the antibiotic-resistant clinical isolated when comparing response rates of Granzyme B, TNF, and CD107a (Fig. 4f).

In summary, our data indicate that MAIT cells are capable of responding towards N.gonorrhoeae, a relevant pathogen for the human uterus.

Discussion

This study provides a detailed characterization of MAIT cells in human endometrium. The access to several unique cohorts such as women undergoing uterus transplantation and monozygotic twins gave us the opportunity to explore dynamics of the endometrial MAIT cell compartment. We observed no differences in MAIT cell frequencies between pre-, post-menopausal, and pregnant endometrial tissue. This indicates that the hormonal changes during the menstrual cycle, post-menopause, or during pregnancy that modulates the endometrium has little effect on the endometrial MAIT cell compartment. This is in contrast to uterine NK (uNK) cells that are well-known to vary in abundance with the uterine hormonal changes26 and play an important role in pregnancy where they further expand.27 Notably, uNK cells express the IL-15-receptor.28 IL-15 is produced in the uterine mucosa in response to progesterone and subsequently promotes uNK cell proliferation.29 Since eMAIT cells seem to not be affected by changes in progesterone (and its downstream effects on IL-15) this suggest that their proliferation are controlled by other factors. This insensitivity to variations in sex hormones indicates that MAIT cell are constantly present in the endometrium and decidua, suggesting they play an important role in defense against relevant pathogens such as N. gonorrhoeae. Moreover, we found no changes in MAIT cell frequency in peripheral blood during the menstrual cycle.

Notably, it has previously been reported that women, as compared to age-matched men, have a higher frequency of peripheral blood MAIT cells.30,31 However, since we saw no modulation of circulating or endometrial MAIT cells in response to hormonal changes, this would suggest that the sex difference in peripheral blood MAIT cells are attributed to other hitherto unknown factors. Indeed, we here report that levels of peripheral blood MAIT cells are very similar between monozygotic twins suggesting that a genetic component contributes to setting the threshold for MAIT cell numbers. By contrast, no such correlation was found for frequencies of endometrial MAIT cells. This indicates that the MAIT cell compartment at mucosal surfaces are affected by environmental cues awaiting discovery. Menstrual blood sampling is inherently more variable compared to sampling of peripheral venous blood. Thus, it is plausible that this contributed to the larger variation observed within menstrual blood MAIT cells in-between twins. However, to ensure an as consistent sampling as possible of menstrual blood, all study participants only collected samples during the first two days of menstruation.

We further show that eMAIT cells display signs of activation. The expression of activation markers indicates that eMAIT cells possibly are more primed and highly functional compared to peripheral blood MAIT cells. This could be due to their localization within the mucosa where pro-inflammatory cytokine levels may be higher. The activated phenotype of eMAIT cells, with high expression of HLA-DR and PD-1, is also in line with what has been published on MAIT cells in other tissues, such as liver32,33 and oral mucosa.12 However, we observed no significant differences in eMAIT cell phenotype in pre- and post-menopausal endometrium except a higher frequency of HLA-DR expressing eMAIT cells in fertile women. The present data are in line with what has been previously published on MAIT cells in the endometrial compartment. Gibbs et al. also found MAIT cells within human endometrium and reported that they display a distinct phenotype compared to that of peripheral blood MAIT cells.20 Furthermore, studies of MAIT cells in decidua of full-term pregnancies (not first-trimester elective termination material, which was used here), have also suggested them to be activated compared to their peripheral blood counterparts.17,34

The majority of eMAIT cells express tissue-residency markers, suggesting that they are resident within the endometrium. However, our results using cells from human uterus transplantations reveal that the donor eMAIT cell population can be replenished by recipient MAIT cells and that very few, if any, donor eMAIT cells were remaining in the endometrial mucosal 2 years after uterus transplantation. This is different compared to what has been reported for conventional human tissue-resident CD4 and CD8 T cells in intestine and lung, where donor cells persist for a prolonged period of time.23,35 These discrepancies might be due to organ-specific factors. However, similar to lung and intestine, uterus also represents a mucosal tissue exposed to the environment. Alternatively, tissue-resident MAIT cells might have a shorter life-span as compared to conventional T cells and/or the population might not be capable of self-replenishment. In this regard, during fetal development, MAIT cells have been suggested to populate tissues early and to express relatively high levels of Ki67 in vivo compared to in adult peripheral blood.36 This would suggest that MAIT cells have an inherent capacity for in situ proliferation. Nevertheless, our data contrasts recent results from the mouse where MAIT cells in tissues were reported to be tissue-resident.11 Possibly, the constant exposure to a more complex environment that humans experience, as compared to mice, could explain these species differences. However, limitations of the current model should also be taken into consideration: to not reject the transplanted organs these women receive immunosuppressive treatment and our time point for assessing tissue residency is fairly long after transplantation. It is possible that this could affect the observed outcome. Despite this, the clinical setting is unique and allows for novel insights into the recirculation patterns of MAIT cells.

Although previous studies have reported that MAIT cells in the female genital tract were responsive to stimulation,20 this has not been investigated in relation to a tissue-specific pathogen. Similar to other bacteria known to activate MAIT cells, N. gonorrhoeae has a riboflavin synthesis pathway and it is likely that the response results, at least in part, from products of the riboflavin operon. Hence, the MAIT cell responses assessed were probably dependent on a combination of both TCR engagement and the action of cytokines.

Future studies should attempt to investigate the local MAIT cell response during ongoing acute N. gonorrhoeae infection. With the global challenge of antibiotic resistance, there is a need to develop alternative treatment strategies for bacterial infections. As for several bacterial species, there are already strains of N. gonorrhoeae that are resistant to most known antibiotics.37 We show here that MAIT cells are able to respond to antibiotic-resistant strains of N. gonorrhoeae. However, the response rate observed was significantly lower than that towards antibiotic-sensitive strains. The same tendency was observed in a recent study by Boulouis et al.38 Nevertheless, a clear positive response was detected and possibly, in the future, immunotherapy involving MAIT cells could represent an alternative treatment option against multi-resistant N. gonorrhoeae infections.

The role of MAIT cell defense in mucosal barriers has been examined in several organs. In lung, MAIT cells have been suggested to protect against infections with Legionella longbeachae,39 Mycobacterium tuberculosis,40 and Haemophilus influenzae.41 MAIT cells in gut have been shown to respond to Helicobacter pylori,13 E.coli,24 and Salmonella typhi.42 Common for these bacterial species is that they all metabolize riboflavin and generate substrates that can be presented by MR1. However, both MR1-dependent activation and activation through cytokines are thought to initiate MAIT cell responses. Except for riboflavin metabolites, other bacterial- or host-specific features such as location of the pathogen (extracellular or intracellular), virulence factors, and type of APCs that respond may also influence the response of MAIT cells.43 A limitation of the current study was the investigation of only a single pathogen in an in vitro setting. Future studies could aim at investigating MAIT cell responses to other clinically relevant pathogens, such as the intracellular bacteria Chlamydia trachomatis, and possibly also in relevant in vivo models. Nevertheless, our findings indicate that MAIT cells are an active component of the endometrial immune system involved in the response against intruding pathogens.

The aim of this study was to investigate the dynamics and function of endometrial MAIT cells. Our analysis shows that eMAIT cells are phenotypically activated and present with a tissue-residency profile. However, by studying human uterus transplantation, we can report that eMAIT cells represents a transiently tissue-resident population in the uterus. In addition, we show that the phenotype as well as size of the eMAIT cell population are largely unaffected by endometrial regeneration occurring during the menstrual cycle and pregnancy as well as endometrial atrophy after menopause. Finally, we report that MAIT cells can respond to the clinically relevant pathogen, N. gonorrhoeae. Together, this provides new insights into human MAIT cell biology, endometrial immunology, and the understanding of how local pathogens are handled.

Methods

Human blood and tissue samples

Samples from several different patient groups were used in the current study. Peripheral blood and endometrial biopsies were obtained from premenopausal and postmenopausal patients undergoing hysterectomy due to non-malignant cause. To study MAIT cells longitudinally endometrial biopsies and peripheral blood were obtained from healthy volunteers sampled at two following menstrual cycles. Peripheral blood was obtained at five-time points during these two menstrual cycles. The endometrial biopsy was taken using Endorette, a vacuum-based method. Decidua was obtained from three women undergoing elective surgical abortion between gestational week 4–8. To study environmental versus heritable influences monozygotic twin pairs were recruited via The Swedish Twin Registry, Karolinska Institutet. In addition, we obtained endometrial tissue from two patients after uterine transplantation as well as peripheral blood from both donor and recipient. Endometrial tissue samples were enzymatically digested for isolation of cells whereas immune cells from menstrual and peripheral blood samples were obtained after Ficoll centrifugation. The Regional Ethics Committee in Stockholm, Sweden approved the study. All patients provided oral and written informed consent.

Antibodies and flow cytometry

See Supplementary Table 1 for antibodies used. Flow cytometry staining was performed as previously described.44 In brief, cryopreserved PBMCs and MNCs were thawed in complete medium (RPMI supplemented with 10% FCS and 2mM L-glutamine), washed twice and resuspended in PBS containing 2% FCS and 2 mM EDTA. All staining for flow cytometry was performed in the dark at room temperature. All samples were stained with Live/dead Aqua (Invitrogen) to discriminate live and dead cells. Samples were then acquired using a BD LSR Fortessa flow cytometer. The MR1 tetramer technology was developed jointly by Dr. James McCluskey, Dr. Jamie Rossjohn, and Dr. David Fairlie, and the material was produced by the NIH Core Facility as permitted to be distributed by the University of Melbourne.45 Data were analysed using FlowJo version 10.5.3 (TreeStar). After adjustment of the compensation matric, samples were concatenated and analyzed using FlowJo plugins Downsample (v2.0.0), UMAP (v2.1), and PhenoGraph (v0.2). UMAP was run using the default settings. PhenoGraph was run using the default number of nearest neighbors (K = 30).

Functional experiments

PBMCs were thawed and resuspended in complete RPMI supplemented with (50 μg/mL) gentamicin (ThermoFisher Scientific). Cells were subsequently stimulated for 18 h with either E.coli (D21 strain, fixed for 4 min in 1X BD Cellfix before used at 10 CFU per cell), N. gonorrhoeae (CCUG 41811 strain, fixed for 4 min in 1X BD Cellfix before used at 0.1, 0.5, 1, 5, 10, and 50 bacteria per cell), N. gonorrhoeae (3 patient isolates), fixed for 4 min in 1X BD Cellfix before used at 10 bacteria per cell or 10 ng/mL IL-12 and 100 ng/mL IL-18. CD107a in FITC was added before adding of the bacteria to assess degranulation. Monensin (BD, Bioscienses) and Brefeldin A (BD, Biosciences) were added for the last 6 h of incubation. To assess MR1-specificity, MR1 (clone 26.5, Biolegend) blocking antibody and the respective isotype (IgG2A clone MOPC-173, Biolegend) was used. Patient isolates were obtained from Clinical Microbiology at Karolinska University Hospital, Huddinge. Titration experiments were performed eight times with 3–4 donors in each experiment, showing similar results.

N. gonorrhoeae growth conditions

Clinical samples were taken from patients with suspected infection according to the local guidelines. The samples were cultured on two agar plates: (1) chocolate agar with polymyxin and vancomycin and (2) chocolate agar with vancomycin, colistin, nystatin, and trimethoprim. The culture plates were incubated at 36 °C for 2 days at 5–6% CO2 atmosphere. The identification was made on colonies by MALDI-ToF MS (Bruker Daltonik, Germany). MALDI-ToF MS scores ≥2.0 indicated identification at the species level. The clinical isolates were tested for antibiotic sensitivity with two of the strains being resistant to ciproxin, and ciproxin and azithromycin, respectively.

Statistical analysis

Data were analyzed using GraphPad Prism (v. 7.0a). First, all data were tested for Gaussian distribution using the D’Agostino & Pearson omnibus normality test. For normally distributed data, comparisons were performed using an unpaired or paired t-test when applicable. For data that failed the normality test, the Mann–Whitney test or Wilcoxon matched-pairs signed-rank test were applied. In all figures, *p < 0.05, **p < 0.01, ***p < 0.001, and n.s denote not significant.