The dysfunction of CD8+ T cells triggered by endometriotic stromal cells promotes the immune survival of endometriosis

Endometriosis is defined as an oestrogen-dependent and inflammatory gynaecological disease of which the pathogenesis remains unclear. This study aimed to investigate the cellular heterogeneity and reveal the effect of CD8+ T cells on the progress of endometriosis. Three ovarian endometriosis patients were collected, and single-cell RNA sequencing (scRNA-seq) progressed and delineated the cellular landscape of endometriosis containing five cell clusters. The endometrial cells (EMCs) were the major component, of which the mesenchymal cells were preponderant and characterized with increased inflammation and oestrogen synthesis in endometriosis. The proportion of T cells, mainly CD8+ T cells rather than CD4+, was reduced in endometriotic lesions, and the cytokines and cytotoxicity of ectopic T cells were depressed. CD8+ T cells depressed the proliferation of ESCs through inhibiting CDK1/CCNB1 pathway to arrest the cell cycle and triggered inflammation through activating STAT1 pathway. Correspondingly, the coculture with ESCs resulted in the dysfunction of CD8+ T cells through upregulating STAT1/PDCD1 pathway and glycolysis-promoted metabolism reprogramming. The endometriotic lesions were larger in nude mouse models with T-cell deficiency than the normal mouse models. The inhibition of T cells via CD90.2 or CD8A antibody increased the endometriotic lesions in mouse models, and the supplement of T cells to nude mouse models diminished the lesion sizes. In conclusion, this study revealed the global cellular variation of endometriosis among which the cellular count and physiology of EMCs and T cells were significantly changed. The depressed cytotoxicity and aberrant metabolism of CD8+ T cells were induced by ESCs with the activation of STAT1/PDCD1 pathway resulting in immune survival to promote endometriosis.

Endometriosis, defined as the presence of endometrium-like tissues outside of the uterine cavity, is considered a benign gynecologic disease that impacts 6%–10% of reproductive-age females [1]. Endometriosis is a disease with many dysfunctional aspects such as hormone effect abnormality and inflammation occurrence [2]. Notably, endometriosis is usually accompanied with chronic pelvic pain, dysmenorrhea and even infertility, which results in negative consequences on the life quality and economic status of patients [3]. Therefore, the treatment of endometriosis has already been upgraded to a public health issue. Recent advances in studying the pathogenesis of endometriosis involved many aspects and viewpoints. The increased oestrogen and aberrant oestrogen receptors played an important role in the development of endometriosis [4, 5]. It was found that abnormal expression of several enzymes contributed to the increased oestrogen level in endometriotic lesions to promote the persistence and growth of ectopic cells [6]. Han et al revealed that oestrogen receptor beta (ERβ) inhibited TNF-α induced apoptosis and interacted with cytoplasmic inflammasome to increase IL-1β, which induced the immune escape and cellular proliferation of ectopic endometrial cells (EMCs) [7]. Agostinis et al considered the complement components as the immunological basis to result in the chronic inflammation characteristic of endometriosis [8]. The differential expression of cytokines was indispensable for the progression of endometriosis via promoting survival, growth and immunoinflammatory response [9]. Therefore, the inflammatory response cooperated with oestrogen to modulate the development of endometriosis. Furthermore, it was reported that the immune cells, including neutrophils, macrophages, NK cells and dendritic cells, appeared to have changes in cell count and physiology that might play a specific role in the angiogenesis, growth and invasion of endometriosis cells [10]. It was suggested that the immune function of the myeloid system was systemically aberrant, and the greater inflammatory phenotype and decreased phagocytic capacity macrophages were found in endometriosis which resulted in defective clearance of EMCs [11]. IL-10+ Th17 cells were significantly increased in the peritoneal fluid of endometriosis and stimulated the proliferation and implantation of ectopic lesions to accelerate the progress of endometriosis [12]. To sum up, the progress of endometriosis is associated with many pathogenic factors including the aberrant oestrogen level, inflammatory response and immunologic function. In this study, a droplet-based scRNA-seq platform (10× Genomics) was employed to profile the architecture of single-cell transcriptome of endometriosis lesions and paired eutopic endometrial tissues from three endometriosis patients. We identified five cell clusters and found the cellular physiological variations between endometrial tissues and endometriotic lesions, of which the ectopic EMCs and T cells were remarkably different. Therefore, the function of T cells in endometriosis was explored from cellular and animal levels. The interaction of T cells and ESCs was studied to disclose the mechanism of immune survival in endometriosis.

Patient and sample collection All three patients suffering from ovarian endometriotic cyst surgery were recruited from the Department of Obstetrics and Gynaecology in the First Affiliated Hospital of Xiamen University. Both eutopic endometria and paired endometriotic tissues were collected after surgery and sent to the laboratory immediately to isolate the single cells for scRNA-seq or usage as the primary cells. The blood samples were collected before the surgery to separate T cells. All patients had regular menstrual cycles and were without hormone treatment for more than 3 months before the surgery. The surgery was chosen at the proliferative phase of patients. And the endometriotic lesions were collected after surgery and examined by pathology. The usage of samples was gated the permission from the ethics committee of the hospital and all patients signed the informed consent. Isolation and culture of primary ESCs The isolation of primary ESCs from endometriotic lesions was referred to the previous method [13]. The endometriotic tissues were cut up into small pieces and then dissociated by Collagenase IV (#A004186-0001; Sangon Biotech; Shanghai, China) for an hour, and Deoxyribonuclease I (DNase I; #B002004-0005; Sangon Biotech) for half an hour. Then the cell suspension was filtered through cell strainers to filter out the undissociated tissue masses and separate epithelial and stromal cells. The isolated primary cells were cultured within DMEM/F12 medium (#SH30023.01; Hyclone; Shanghai, China) containing 10% fetal bovine serum (FBS; #04-001-1A; Biological Industries; Cromwell, USA) in the 37°C incubator. scRNA-seq processing The single-cell suspension was isolated following the above method of isolating primary ESCs with some modifications. Additionally, the cell suspensions after the treatment of collagenase and DNase were density gradient centrifuged by Percoll Separation Fluid (#P8370; Solarbio; Beijing, China) to remove the tissue fibre and red blood cells. The single cells were washed with PBS containing 0.04% BSA and confirmed cell concentration via the automated cell counter. Approximately several cells (1 × 105 cells/mL) were subjected immediately to the Chromium Controller machine (10× Genomics) for generation of Gel Beads-in-Emulsion (GEM) in which mRNA of single cells was released and further processed. After the reverse transcription of mRNA from the barcoded cell, cDNA was recovered, purified and amplified to generate sufficient quantities for library preparation. The sequencing of cDNA libraries was run on the Novaseq for Illumina PE150 sequencing (California, USA). scRNA-seq data analysis and visualization The raw data of scRNA-seq were demultiplexed and mapped via Cell Ranger software (Version 2.1.1). Then, each dataset was log-normalized, scaled and corrected for dataset-specific batch effects. All further analyses were performed using the Seurat package (Version 2.3.3). Principle component analysis (PCA) was performed to the analysis of dimensions reduction. T-Distribution Stochastic Neighbour Embedding (t-SNE) and Uniform Manifold Approximation and Projection (UMAP) dimension representation were used to visualize the cell distribution. The cell type definition was accorded to the conserved markers: endometrial cells/EMCs (DCN, LUM, KRT19) included mesenchymal cells (ACTA2, ENG, VIM, DLK2), epithelial cells (CD24, EPCAM, KRT19, TJP1, CDH1) and proliferative cells (TOP2A, MKI67, PCNA, CDK1); endothelial cells/ECs (VWF, SELE, PECAM); vascular smooth muscle cells/VSMCs (MYH11, PDGFRB, CNN1); lymphocytes (TRAC, CD2, CD3D/E/G) containing T cells/TCs (CD3D/EG, TRAC, CD8A/B), NK cells (GNLY, GZMB, NKG7), B cells (CD79A, IGKC, MZB1); myeloid cells (IL-1β, CD14, CD68) included macrophages (CD14, CD68), neutrophils (FCGR3B, CXCR2, CMTM2). The analysis of pseudotime trajectory was performed via Monocle 2. Enrichment analysis Gene Ontology (GO) enrichment and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways analysis of differentially expressed genes were performed to understand the changes of functions and pathways via referring to STRING database (https://cn.string-db.org/). The corrected p-value of enrichment results less than 0.05 were regarded significantly. Immunofluorescence To examine the cell heterogeneity in different endometrial tissues, the conversed markers were marked via immunofluorescence. The tissue samples were processed by fixation, dehydration and embedment as previously reported [14]. And then the slides of samples were incubated with the primary antibodies at 4°C overnight. The primary antibodies included CD3 antibody (#ab16669; Abcam; 1:100 dilution), CD68 antibody (#ab201340; Abcam; 1:100 dilution) and CD8A antibody (#ab237709; Abcam; 1:100 dilution). After being washed with PBS for 3 times, the slides were incubated with corresponding fluorescent secondary antibodies. The cell nucleus was stained through DAPI resolution. The results were observed and photographed by fluorescence microscope (Olympus Corporation, Tokyo, Japan). RNA extraction and qRT-PCR The treated cells were lysed through RNAiso Plus reagent (#9108; Takala Biotechnology; Kyoto, Japan). The cell lysate was processed by Chloroform, precipitated by Isopropyl-Ketone and washed with 75% ethanol to purify the total RNA. And then, the exacted RNA was applied as the template to synthesize cDNA via using the PrimeScript RT reagent kit (#9108; Takala Biotechnology; Kyoto, Japan). qRT-PCR was reacted using the TB Green Premix Ex Taq II (#RR820A, Takala) in the LightCycler 480 machine (Roche Molecular Biochemicals, Mannheim, Germany). The results were normalized based on GAPDH mRNA level. All the primer sequences used in this study are listed in Table 1. | Gene name | Sequences (5′–3′) | |---|---| | GAPDH | Forward: GGAAGGTGAAGGTCGGAGTCA Reverse: GAGTCCTTCCACGATACCAA | | C3 | Forward: GGGGAGTCCCATGTACTCTATC Reverse: GGAAGTCGTGGACAGTAACAG | | C7 | Forward: AATGGCTGTACCAAGACTCAGA Reverse: GCTGATGCACTGACCTGAAAA | | B2M | Forward: GAGGCTATCCAGCGTACTCCA Reverse: CGGCAGGCATACTCATCTTTT | | IL-6 | Forward: ACTCACCTCTTCAGAACGAATTG Reverse: CCATCTTTGGAAGGTTCAGGTTG | | IL-8 | Forward: ACTGAGAGTGATTGAGAGTGGAC Reverse: AACCCTCTGCACCCAGTTTTC | | CCL2 | Forward: CAGCCAGATGCAATCAATGCC Reverse: TGGAATCCTGAACCCACTTCT | | CXCL2 | Forward: CCAACTGACCAGAAGGGA Reverse: TGCTCAAACACATTAGGCA | | CXCL10 | Forward: GTGGCATTCAAGGAGTACCTC Reverse: TGATGGCCTTCGATTCTGGATT | | FCGR3A | Forward: CCTCCTGTCTAGTCGGTTTGG Reverse: TCGAGCACCCTGTACCATTGA | | CCR8 | Forward: GTGTGACAACAGTGACCGACT Reverse: CTTCTTGCAGACCACAAGGAC | | CXCR6 | Forward: GACTATGGGTTCAGCAGTTTCA Reverse: GGCTCTGCAACTTATGGTAGAAG | | KLRC1 | Forward: AGCTCCATTTTAGCAACTGAACA Reverse: CAACTATCGTTACCACAGAGGC | | ISG15 | Forward: CGCAGATCACCCAGAAGATCG Reverse: TTCGTCGCATTTGTCCACCA | | PDCD1 | Forward: CCAGGATGGTTCTTAGACTCCC Reverse: TTTAGCACGAAGCTCTCCGAT | | CYP17A1 | Forward: GCTGCTTACCCTAGCTTATTTGT Reverse: ACCGAATAGATGGGGCCATATTT | | CYP19A1 | Forward: TGGAAATGCTGAACCCGATAC Reverse: AATTCCCATGCAGTAGCCAGG | Western blot Western Blot was performed in this research to detect the protein expression. The treated cells were lysed by RIPA lysis buffer (#ab156034; Abcam) containing protease inhibitors and phosphatase inhibitors. Then, the protein samples were prepared by treating with SDS-PAGE loading buffer and then separated via SDS-PAGE and transferred onto PVDF membranes (#03010040001; Roche, Basel, Switzerland). Different primary antibody dilutions and corresponding HRP-labelled secondary antibodies were used to incubate the membranes. The antibodies in the experiment were STAT1 antibody (#ab109320, Abcam; 1:10000 dilution), phospho (Y701)-STAT1 antibody (#ab30645, Abcam; 1:1000 dilution), phospho (S727)-STAT1 antibody (#ab109461, Abcam; 1:5000 dilution), CDK1 antibody (#ab133327, Abcam; 1:20000 dilution); CCNB1 antibody (Abcam, #181593; 1:1000 dilution), TOP2A antibody (#ab52934, Abcam; 1:5000 dilution), PCNA antibody (#13110, CST; 1:1000 dilution), CD16 antibody (#ab246222, Abcam; 1:1000 dilution), CCR8 antibody (#ab32399, Abcam; 1:1000 dilution), PDCD1 antibody (#86163, CST; 1:1000 dilution), GPI antibody (#94068, CST; 1:1000 dilution), PKM antibody (#ab137791, Abcam; 1:2000 dilution), PGK1 antibody (#ab199438, Abcam; 1:1000 dilution) and GAPDH antibody (#ab181602, Abcam; 1:10 000 dilution). The signal of blot was visualized by chemiluminescence (ECL) and scanned by ChemiDoc MP Imaging System (Bio-Rad; California, USA). The acquisition and culture of human T cells T cells were collected from the peripheral blood of endometriosis patients through the lymphocyte separation solution (#P8610; Solarbio; Beijing, China). RPMI medium with 10% FBS, 30 U/mL human IL-2 (#HY-P7037; MedChemExpress) and 20 μL/mL CD3/CD28 T cell activator (#11131D; ThermoFisher; Waltham, MA, USA). Flow cytometry was used to distinguish and separate the subtype of T cells according to the combination of CD4 antibody (#ab133616, Abcam; 1:100 dilution) and CD8A antibody (#ab237709, Abcam; 1:500 dilution). Cell Coculture analysis To demonstrate the in vitro interaction between ESCs and T cells, the transwell system was applied in this research. 1 × 106 ESCs were plated in the lower compartment, and 1 × 105 T cells were cultured in upper chambers. The cells were cocultured in the transwell system for 48 h in the 37°C incubator, and then the total RNA was extracted, respectively, for RNA sequencing or qRT-PCR. Cell proliferation assay To investigate the proliferation of cocultured ESCs and T cells, cell counting kit-8 (CCK-8; #HY-K0301; MedChemExpress, Shanghai, China) was employed to quantify the growth rate. After the coculture of ESCs and T cells, the transwell chambers were separated and added the CCK-8 reagent for 2 h at 37°C. The absorbance of supernatants was measured with an ELISA reader spectrophotometer (Dynatec Laboratories, Chantilly, VA). EdU labelling assay EdU labelling assay were performed to evaluate the cell proliferation of ESCs cocultured with T cells. EdU labelling experience was conducted as described in the instructions of EdU assay kit (#C0071S; Beyotime Biotechnology, Shanghai, China). After the coculture with T cells, ESCs were incubated with EdU solution for 3 h. Next, the cells were fixed by 4% paraformaldehyde and permeabilized by PBS containing 0.3% Triton X-100. And then the click additive solution (included in the kit) was added to incubate the cells for 30 min in the dark conditions. Finally, Hoechst 33342 was used to stain the nucleus. The results were observed and captured using fluorescence microscope (Olympus Corporation, Tokyo, Japan). RNA sequencing The total RNA was extracted through TRIzol reagent (#15596026; Invitrogen; California, USA). The extracted RNA was reacted with Oligo(dT) beads to enriched mRNA which next was fragmented into short fragments via fragmentation buffer. The RNA fragments were reversely transcribed into cDNA by NEBNext Ultra RNA Library Prep Kit (#7530; New England Biolabs; Massachusetts, USA). The purified cDNA was end-repaired, added A base and ligated to sequencing adapters. And then, polymerase chain reaction (PCR) was amplified to establish cDNA library which was then sequenced by Illumina Novaseq 6000. Endometriotic mouse model assay To analyse the function of total T cells and CD8+ T cells on the development of endometriosis in vivo, the endometriotic mouse models were established. Both female BALB/c mice and nude BALB/c mice (6–8 weeks old and 15–20 g) were bought and fed in the animal research laboratory of Xiamen University, and all the treatments on mice complied with animal ethics. The method of establishing endometriosis mouse models was followed the previous report [14]. In brief, the uterine of donor mice were surgically collected, and the uterine horns and myometrium were peeled off to obtain the endometrium-rich fragments. The endometrial tissues were cut into pieces and equally injected into the abdominal cavity of recipient mice to establish endometriosis mouse models. The BALB/c mouse models were grouped into EMs group, EMs + CD90.2 group and EMs + CD8A group which were abdominal injected with 200 μL saline, 200 μL saline with 100 μg CD90.2 antibody (#105302, Biolegend; San Diego, USA) and 200 μL saline with 100 μg CD8A antibody (#100746, Biolegend), respectively, every week for three times to deplete the pan T cells and CD8+ T cells [15, 16]. The nude mouse models were grouped into nuEMs group and nuEMs + TCs group. The mice of nuEMs group were injected with 200 μL saline and the nuEMs + TCs group were injected with 200 μL saline containing 1 × 106 CD8+ T cells extracted from the lymph nodes of female BALB/c mice via CD8 T Cells Enrichment Kit (#8804-6822-74; ThermoFisher). All the mouse models were normally raised for a month and then were sacrificed to collect the ectopic lesions, respectively. Statistical analysis All statistical analyses were performed via OriginLab (version 9.7) or R software. Comparisons between two groups were performed with a two-tailed or unpaired student's t-test. All data are presented as means ± standard deviations (SDs).

Single-cell transcriptomic atlas revealed the cellular variation of endometriosis The progress of endometriosis is complicated that involved the transformation of cellular functions and the remoulding of the tissue microenvironment. Therefore, scRNA-seq was employed to detect the cellular variation between eutopic endometrial tissues and endometriotic lesions from endometriosis patients. After the sequencing, quality control filtering, and statistical analysis, the single-cell transcriptomic atlas of 15 300 cells was obtained (Figure 1a). To evaluate the cellular distinction between eutopic endometria and endometriotic lesions, PCA analysis was performed. There were obvious differences of distribution between global eutopic and ectopic cells (Figure 1b) indicating the significantly changed gene expression of ectopic cells. And the further visualization of total cell distribution was performed using the UMAP project (Figure 1c). The conserved marker genes were annotated and quantified (Figure 1d). All the detected cells were assigned into five different cell types including EMCs, endothelial cells (ECs), vascular smooth muscle cells (VSMCs), lymphocytes and myeloid cells (Figure 1e). The cellular proportions of different clusters in eutopic endometria and endometriotic lesions were counted (Figure 1f). The proportion of EMCs (Eut: 70%, EMs: 53%), the major cells in endometrial tissues, was reduced in endometriotic lesions. Accordingly, it was found that ECs (Eut: 2%, EMs: 9%), VSMCs (Eut: 2%, EMs: 9%) and myeloid cells (Eut: 5%, EMs: 12%) were increased, and lymphocytes (Eut: 21%, EMs: 17%) were reduced in endometriosis group. Thus, our results of scRNA-seq identified five cell clusters, and the cellular composition was changed in endometriotic lesions compared to eutopic endometria. There were promoted inflammation and oestrogen synthesis in the ectopic EMCs characterized by epithelial-mesenchymal transition EMCs were the major cellular component in endometrial tissues. The distribution of eutopic and ectopic EMCs were further present via t-SNE analysis in Figure 2a, and the EMCs were divided into four subtypes (EMCs-1, EMCs-2, EMCs-3 and EMCs-4). According to the expression of cell markers (Figure 2b), EMCs-1 significantly increased expressed the mesenchymal markers, and EMCs-4 markedly expressed epithelial markers and lower mesenchymal cells. Thus, EMCs-1 was defined as mesenchymal cells and EMCs-4 was epithelial cells. EMCs-2 equally expressed mesenchymal and epithelial markers and was considered as the cell within a transition state between mesenchymal and epithelial cells. TOP2A, MKI67, PCNA and CDK1 were markedly expressed in EMCs-3 which was defined as proliferative EMCs. The cell proportion analysis found that EMCs-1 (Eut: 27%, EMs: 81%) were greatly enhanced in endometriotic lesions. EMCs-2 (Eut: 58%, EMs: 12%), EMCs-3 (Eut: 8%, EMs: 4%) and EMCs-4 (Eut:7%, EMs:2%) were decreased (Figure 2c). Moreover, the mesenchymal marker ACTA2 and VIM were increased and the epithelial markers KRT19 and TJP1 were decreased in ectopic EMCs (Figure 2d). Therefore, these results suggested that the ectopic EMCs present notable mesenchymal-epithelial transition (EMT) to dramatically increase mesenchymal cells. Comparing the gene expression profiles between the eutopic and ectopic EMCs, 1537 genes were increased and 1135 genes were decreased in ectopic EMCs (Figure 2e). The GO analysis disclosed that angiogenesis, extracellular matrix organization, regulation of cell migration, inflammatory response, chemotaxis and response to cytokines/hormones were enriched in the upregulated genes. The GO terms including sister chromatid segregation, nuclear division, cell division, mitotic cell cycle and cell proliferation were enriched upon the downregulated genes (Figure 2f). KEGG enrichment analysis found that the NK-κB pathway, TNF pathway, IL-17 pathway, complement and coagulation cascades and MAPK pathway were enriched following the increased genes, and the decreased genes mainly enriched the cell cycle (Figure 2g). As shown in (Figure 2h,i), based on the changes of physiological function, the ectopic EMCs increasingly expressed cytokines including CCL2, CXCL1, CXCL2, CXCL3, IL-6 and CXCL8/IL8. And the complements C1S, C1R, C2, C3, C7 and activator CFB, CFD and CFH were markedly enhanced in ectopic EMCs. The increased cytokines and complements might promote the immune cells infiltration and the formation of inflammatory microenvironment. Furthermore, the oestrogen synthesis-associated enzymes STAR, CYP11A1, CYP17A1, CYP19A, HSD3B2 and HSD17B1 were remarkably upregulated in the ectopic EMCs. Therefore, the ectopic EMCs present the mesenchymal transformed and promoted inflammatory response and oestrogen synthesis to establish the pathological microenvironment. The count and immune function of T cells were regressive in endometriotic lesions The immune cells were separated and analysed, and five types of immune cells were distinguished including T cells, macrophages, NK cells, neutrophils and B cells according to the conversed markers (Figure 3a,b). The cell count of macrophages (Eut: 17%, EMs: 37%), neutrophils (Eut: 1%, EMs: 5%) and B cells (Eut: 3%, EMs: 5%) were increased in endometriotic lesions. The proportion of T cells (Eut: 65%, EMs: 38%) were decreased, and NK cells (Eut: 15%, EMs: 15%) remained unchanged (Figure 3c). Therefore, macrophages and T cells were the immune cell clusters presenting the most changed in endometriosis. The macrophage marker CD68 was increased and the marker of T cell CD3 was detected and found decreased in endometriotic tissues through immunofluorescence staining (Figure 3d). Thus, the greatly increased macrophages and decreased T cells might perform important functions in the development of endometriosis. The above results revealed that T cells were the major immune cells in the endometrial tissues and decreased in endometriosis, which suggested the dysfunction of T cells promoted the development of endometriosis. Firstly, the subtypes of T cells were identified according to the expression of CD4 and CD8. As shown in Figure 3e, the endometrial T cells mainly consisted of CD8+ T cells rather than CD4+ T cells. And the expression of CD8 was decreased in ectopic lesions detected through immunofluorescence (Figure 3f). The differently expressed genes between eutopic and ectopic T cells were further analysed. As shown in Figure 3g, there were 378 genes increased and 713 genes decreased in the ectopic T cells according to the results of scRNA seq. The increased genes enriched peptide biosynthetic process, inflammatory response, programmed cell death and cell adhesion. The GO terms including natural killer cell-mediated cytotoxicity, natural killer cell activation, response to interferon-γ, T cell activation and chemotaxis were enriched following the decreased genes (Figure 3h). And the expression of cytokines including interferon-γ/IFNG, IL2, IL4, CCL3, CCL3L3, CCL4, CCL4L2, CSF1, XCL1, XCL2 and cytotoxicity markers containing KLRC4, KLRD1, NCR1, NKG7, PRF1, GZMA, GZMH, GZMK and GZMM were greatly decreased in the ectopic T cells (Figure 3i). Therefore, all these results suggested that the endometrial T cells mainly consisted of CD8+ T cells, and the immune cytotoxicity of ectopic T cells was diminished. CD8+ T cells triggered the immune stimulation and cell cycle arrest of ESCs To explore the function of the decreased CD8+ T cells in the pathology of endometriosis, CD8+ T cells were collected from the peripheral blood of endometriotic patients via CD8 positive isolation kit. To evaluate the interaction of ESCs and CD8+ T cells, the transwell system was used to coculture the primary ESCs and CD8+ T cells (Figure 4a). As exhibited in Figure 4b, the proliferation of ESCs and CD8+ T cells were both reduced within the coculture system. Therefore, ESCs and CD8+ T cells present mutual inhibition in proliferation. The cocultured ESCs and CD8+ T cells were sequenced to detect the differently expressed genes. There were 1229 genes increased and 1199 genes decreased in the cocultured ESCs (Figure 4c). As shown in Figure 4d, the increased gene profiles enriched T cell-mediated cytotoxicity, T cell costimulation, cytokine activity, inflammatory response, angiogenesis, lymphocyte activation, chemotaxis, cell mobility and so on. And the decreased genes enriched mitotic spindle organization, sister chromatid segregation, cell division and cell cycle. Thus, ESCs induced the drastically changed gene expression associated with immune stress (Figure 4e). And the mRNA expression of C3, B2M, IL6, IL8/CXCL8, CCL2 and CXCL10 were enhanced in the cocultured ESCs (Figure 4f). Furthermore, the protein expression of STAT1 and phosphorylation were significantly upregulated in ESCs stimulated by CD8+ T cells (Figure 4g). Consequently, the coculture with CD8+ T cells resulted in the immune stimulation of ESCs accompanied with abundant cytokines secretion and inflammatory response with the activation of STAT1 pathway. Furthermore, the genes involved with cell cycle and division were decreased in cocultured ESCs (Figure 4h). To detect the change of cell cycle of cocultured ESCs, flow cytometry was performed. As shown in Figure 4i, the cocultured ESCs exhibited a reduced proportion of G2/M cells and enhancement of G0/G1 cells, which implied the cell cycle of ESCs was arrested before G2/M phase. And EdU-labelling assay proved that CD8+ T cells inhibited the cell replication of ESCs (Figure 4j). Thus, CD8+ T cells were cytotoxic to ESCs and inhibited the cellular proliferation. The cell cycle regulatory proteins CDK1 and CCNB1 were found reduced in the cocultured ESCs. And proliferation markers TOP2A and PCNA in ESCs were decreased with the stimulation of CD8+ T cells (Figure 4k). The above results suggested that CD8+ T cells exerted an immune cytotoxic role to trigger the drastic immune stimulation and cell cycle arrest of ESCs. ESCs induced the exhaustion of CD8+ T cells through activating STAT1/PDCD1 pathway and remodulating metabolism The gene profiles of CD8+ T cells stimulated by the ESCs were found differently including 274 genes increased and 473 genes decreased (Figure 5a). According to the increased genes, interferon signalling, glycolytic process, negative regulation of viral genome replication, pyruvate metabolic process, defence response to virus, response to hypoxia and programmed cell death were enriched, and the function of mitochondrion including oxidative phosphorylation, electron transport chain and mitochondrial envelop were significantly enriched following the decreased genes (Figure 5b). To estimate the immune activity of cocultured CD8+ T cells, the gene profiles of immune function were detected. As shown in Figure 5c, the genes of interferon signalling and immune response were increased, but the molecules associated with T cell activity including FCGR3A/CD16, CCR8, CCR4, CCL1, KLRC1 and KLRD1 were depressed in the cocultured CD8+ T cells according to the data of RNA seq. The mRNA of FCGR3A/CD16, CCR8, CXCR6 and KLRC1 were significantly decreased, ISG15 and PDCD1 were dramatically increased (Figure 5d). Moreover, the protein expression of FCGR3A/CD16, CCR8 were decreased, and PDCD1 and STAT1 were upregulated in the cocultured CD8+ T cells (Figure 5e). Therefore, CD8+ T cells present the exhausted status with the activation of STAT1/PDCD1 signalling under the influence of ESCs. On the other hand, it found that the genes regulating the glycolytic process including GPI, PGK1, PDK1, ALDOC and ALDOA were enhanced, and the molecules of oxidative phosphorylation were greatly depressed, which implied the metabolism reprogramming of cocultured T cells (Figure 5f). And the protein expression of GPI, PKM and PGK1 were significantly increased in the cocultured CD8+ T cells (Figure 5g). Thus, ESCs triggered the metabolism reprogramming of CD8+ T cells with increased glycolysis and aberrant mitochondrial function, which might result in the low immune activity. CD8+ T cells inhibited the growth of endometriotic lesions in vivo To investigate the function of T cells/CD8+ T cells in the development of endometriosis in vivo, we established the endometriosis mouse models. As exhibited in Figure 6a, BALB/c mice and BALB/c nude mice were used to establish the endometriosis mouse models, and BALB/c mouse models were abdominal injected with CD90.2 (Thy 1.2) antibody or CD8A antibody to inhibit T cells and CD8+ T cells, and BALB/c nude mouse models were injected with CD8+ T cells. As shown in the Figure 6b,c, the lesions of nude mouse models were larger than all other groups, which suggested the deficient immune were conducive to the development of endometriosis. However, the supplement of CD8+ T cells to nude mouse models prominently reduced the weight of endometriotic lesions. Furthermore, the lesions in normal BALB/c mice were most small, and both the inhibition of total T cells via CD90.2 antibody and CD8+ T cells via CD8A antibody notably increased the growth of lesions. In conclusion, the immune function of T cells/CD8+ T cells exerted the important role of suppressing the development of endometriosis.

In this study, we performed scRNA-seq to detect the cellular heterogeneity between eutopic endometria and paired endometriotic lesions from endometriosis patients, and five cell clusters were distinguished. The ectopic EMCs were predominantly mesenchymal cells, promoted inflammatory response by secreting cytokines, and increased oestrogen synthesis. The cytokines secretion and cytotoxicity of ectopic T cells which mainly consist of CD8+ cells were dramatically depressed. Furthermore, the coculture with T cells induced immune stress of ESCs with abundant cytokine secretion via activating STAT1 pathway, and resulted in the cell cycle arrest. And ESCs induced the immune activity of T cell depressed through promoting STAT1/PDCD1 pathway and resulted in glycolysis-promoted metabolism reprogramming. Consequently, we delineated the scRNA transcriptional architecture that demonstrated the cellular heterogeneity of endometriosis and revealed that the dysfunction of T cells that interacted with the ESCs induced immune tolerance to promote endometriosis. To reveal the cellular heterogeneity of endometriosis, many studies have been devoted to investigating. It was reported that endometrial mesenchymal cells were the most abundant cells, T cells/natural killer T cells were the second most cells, and then epithelial cells, myeloid cells, smooth muscle cells, ECs and B lymphocytes [17]. Another research defined five overarching cell types including epithelial, stromal, endothelial, lymphocyte and myeloid, and found that the stromal cells and pro-angiogenesis macrophages remarkably increased, lymphocytes greatly decreased in ovary endometriosis, driving angiogenesis and immunotolerance [18]. Zhu et al declared that the most cell proportion was myofibroblasts and the proportion of ECs and macrophages were significantly increased in endometriotic lesions, which resulted in fibrosis and pro-inflammation [19]. Another previous study identified nine cell types in endometriosis, in which the ectopic T cells, accompanied by decreased effector CD8+ T cells, were less activated and inflammatory, and the ectopic macrophages were increased in endometriosis cysts [20]. Therefore, all these studies illustrated the cellular heterogeneity of endometriotic lesions, and demonstrated the cell count and physiological variation of EMCs and immune cells, although the cellular proportion and physiological function in different research were inconsistent for the different methods of cell preparation and cell definition. Overall, this research distinguished the EMCs including were the major cell type, the endometrial T cells were the second enriched cells, and macrophages were the third most cell type, suggesting these three major cell types might present the most important for the occurrence of endometriosis. Endometriosis is defined as the presence of endometrium-like tissues including endometrial epithelium and stroma outside of the uterine cavity [21]. Thus, the cellular activity of EMCs, the major cells in endometria, determined the incidence of endometriosis. The previous studies reveal that the markers of mesenchymal cells were significantly increased in the ectopic lesions [22, 23]. Our data demonstrated that the mesenchymal cells were significantly enhanced in endometriotic lesions, which indicated the promoted epithelial-mesenchymal transition (EMT) in endometriosis. The increased proportion of mesenchymal cells indicated the enhanced ability of migration and invasion [24]. On the other hand, the expression of cytokines and complements was greatly enhanced in the ectopic EMCs, especially IL6, CXCL/IL8, C3 and C7. IL6 induces the hyperinflammation [25]. and IL8 is responsible of stimulating proliferation, migration, angiogenesis and immune suppression [26]. The dysregulation of complements was found and associated with the local chronic inflammation of endometriosis [27]. Therefore, the transformation of ectopic EMCs increasedly expressing cytokines and complements promoted the cellular activity and inflammatory response. Furthermore, endometriosis is considered an oestrogen-dependent disease [28], and the oestrogen synthesis-associated enzymes including aromatase [29], HSD3B1 [30] and HSD17B1 [31] were noticeably increased. This research demonstrated that it was the ectopic EMCs that highly expressed oestrogen synthetases. In conclusion, the mesenchymal transformed EMCs of ectopic lesions potently promoted the formation of a pathogenesis microenvironment that regulated the proliferation, migration, apoptosis and immune response [32], and greatly affected the therapeutic effect and clinical outcome [33, 34]. The endometrial T cells, the second abundant cells in endometria, were remarkedly decreased in ectopic lesions, which suggested the dysfunction of T cells was associated with the pathology of endometriosis. It was reported that the frequencies and number of PD-1-positive T cells were greatly increased in peripheral blood from endometriosis patients [35]. The level of CTLA-4 T cells, a marker of anergic lymphocytes, was significantly higher in the peripheral blood of patients with advanced endometriosis [36]. Thus, there is functional degeneration of T cells in endometriosis. This study found that CD8+ T cell is the prominent cell in endometrial T cells. CD8+ T cells exhibited cytotoxicity against tumour cells and secreted [37]. It found that CD8+ T cells inhibited the proliferation of ESCs and the development of endometriosis in vivo, which indicated the therapeutic effect of T cells in endometriosis. However, it was reported that CD8+ T cells from the menstrual effluent of endometriosis patients were less cytotoxic for the significantly reduced perforin [38]. Furthermore, the expression of genes associated with T cell activation, including abundant cytokines and immune cytotoxicity genes, were distinguished and decreased according to the results of scRNA seq, such as IFNG, IL2, PRF1, NKG7, KLRD1 and GZMA. Thus, the T cells in endometriosis present the status of immune functional attenuation. Herein, this research explored the interaction between ESCs and T cells and revealed the anergic immune of T cells triggered by ESCs. Firstly, it was found that ESCs could trigger the response to virus and interferon signalling of T cells, which suggested ESCs were regarded as the extracellular stimulation for T cells. The coculture with ESCs resulted in the reduced proliferation of T cells. Therefore, ESCs stressed T cells and inhibited the cell growth. It was reported that FCGR3A/CD16 [39], CXCR6 [40] and KLRC1 [41] represent the immune activity and cytotoxicity of T cells, and the cocultured T cells decreased the expression of these genes. Moreover, the inhibitory receptor PDCD1, the marker of T cell exhaustion [42], was increased in the cocultured T cells, which implied the ESCs induced the exhaustion of T cells. Thus, ESCs might diminish the immune potential of T cells in the endometriotic microenvironment. STAT1 was associated with the reduced survival of CD4+ T cells [43]. It was reported that STAT1 suppressed the immune responsiveness of CD8+ T cells through STAT4-mTOR1 signalling axis [44]. The previous study found JAK–STAT axis controlled the expression of PD-L1 to regulate lymphocyte infiltration [45]. And the activation of STAT1 were involved with the regulation of type I interferon pathway on regulating the expression of PDCD1 and PD-L1 [46]. Moreover, it was reported that STAT1 was a key mediator of eIF4F inducing PD-L1 expression [47]. Another study found that IFNα/β promoted the expression of PDCD1 depending on STAT1 which directly binds to the region of PDCD1 gene [48]. Thus, STAT1 could affect the PDCD1 pathway at the aspect of transcription. Our results suggested that both STAT1 and PDCD1 were upregulated with the stimulation of ESCs, which indicated ESCs impaired the immune activity of CD8+ T cells. Besides, the coculture with ESCs triggered the physiological abnormality of T cells including the promoted glycolysis, pyruvate metabolic process and aberrant mitochondrial function. The survival and immune function of T cells are dependent on energy from glycolysis and oxidative phosphorylation, and the metabolic transition from oxidative phosphorylation to glycolysis in CD8+ T cells impaired their survival and immune function [49, 50]. Therefore, the ESCs induced the aberrant metabolism to impact the immune function of T cells. In conclusion, the immune cytotoxicity of T cells was depressed with the interaction of ESCs via metabolism reprogramming and PDCD1 pathway to support immune tolerance. To sum up, this study demonstrated the cellular variation of endometriosis and distinguished the number and physiological changes of the ectopic EMCs and T cells. The immune cytotoxicity of ectopic CD8+ T cells was depressed by ESCs through upregulating the PDCD1 signalling and glycolysis metabolism to promote the immune surveillance. Therefore, the dysfunction of T cells triggered by ESCs induced an abnormal immune microenvironment to promote the development of endometriosis. AUTHOR CONTRIBUTIONS Zhi-Xiong Huang and Dian-Chao Lin contributed equally to this research, completed the experiments, analysed the data and wrote the manuscripts. Hua-Ying Zhang and Jia-Hao Chen undertook primary cell culture and analysed partial experimental data. Meng-Jie Yang and Xin-Yu Ding reviewed and analysed the sequencing data and drew the graphics. Song-Juan Dai, Yi-Huang Hong and Gui-Shuang Liang were in charge of clinical samples and case information collection. Qing-Xi Chen and Qiong-Hua Chen conceived the study, provided financial support and revised the manuscript.

Thanks to Dr. Qian-Sheng Huang, Dr. Wei-Dong Zhou and Dr. Rong-Feng Wu for their valuable comments and excellent technical assistance in this research. The author(s) disclosed receipt of the following financial support for the research: the Natural Science Foundation of China (No. 82301852 and No. 82271678), and the National Science Foundation of Fujian Province (No. 2022J02058). CONFLICT OF INTEREST STATEMENT The author(s) announced there were no potential conflicts of interest in this article. DATA AVAILABILITY STATEMENT The data that support the findings of this study are available from the corresponding author upon reasonable request.