{"paper_id":"cca8bb4f-34ec-442b-8dc5-e7326cda793f","body_text":"Shih et al. BMC Medicine          (2022) 20:315  \nhttps://doi.org/10.1186/s12916-022-02500-3\nRESEARCH ARTICLE\nSingle-cell analysis of menstrual endometrial \ntissues defines phenotypes associated \nwith endometriosis\nAndrew J. Shih1  , Robert P . Adelson1, Himanshu Vashistha1, Houman Khalili1, Ashima Nayyar1, Radha Puran1, \nRixsi Herrera1, Prodyot K. Chatterjee1, Annette T. Lee1,2, Alexander M. Truskinovsky2,3, Kristine Elmaliki1, \nMargaret DeFranco1, Christine N. Metz1,2*   and Peter K. Gregersen1,2*   \nAbstract \nBackground: Endometriosis is a common, complex disorder which is underrecognized and subject to prolonged \ndelays in diagnosis. It is accompanied by significant changes in the eutopic endometrial lining.\nMethods: We have undertaken the first single-cell RNA-sequencing (scRNA-Seq) comparison of endometrial tissues \nin freshly collected menstrual effluent (ME) from 33 subjects, including confirmed endometriosis patients (cases) and \ncontrols as well as symptomatic subjects (who have chronic symptoms suggestive of endometriosis but have not \nbeen diagnosed).\nResults: We identify a unique subcluster of proliferating uterine natural killer (uNK) cells in ME-tissues from con-\ntrols that is almost absent from endometriosis cases, along with a striking reduction of total uNK cells in the ME of \ncases (p <  10−16 ). In addition, an IGFBP1+ decidualized subset of endometrial stromal cells are abundant in the shed \nendometrium of controls when compared to cases (p <  10−16 ) confirming findings of compromised decidualization \nof cultured stromal cells from cases. By contrast, endometrial stromal cells from cases are enriched in cells expressing \npro-inflammatory and senescent phenotypes. An enrichment of B cells in the cases (p = 5.8 ×  10−6 ) raises the possibil-\nity that some may have chronic endometritis, a disorder which predisposes to endometriosis.\nConclusions: We propose that characterization of endometrial tissues in ME will provide an effective screening tool \nfor identifying endometriosis in patients with chronic symptoms suggestive of this disorder. This constitutes a major \nadvance, since delayed diagnosis for many years is a major clinical problem in the evaluation of these patients. Com-\nprehensive analysis of ME is expected to lead to new diagnostic and therapeutic approaches to endometriosis and \nother associated reproductive disorders such as female infertility.\nKeywords: Menstrual blood, Menstrual effluent, Inflammation, Senescence, Fibrosis, Biomarkers, Decidualization, \nEndometriosis, Single-cell RNA sequencing\n© The Author(s) 2022. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which \npermits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the \noriginal author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or \nother third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line \nto the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory \nregulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this \nlicence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. The Creative Commons Public Domain Dedication waiver (http:// creat iveco \nmmons. org/ publi cdoma in/ zero/1. 0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.\nBackground\nEndometriosis is a common and heterogeneous disor -\nder that is characterized by the growth of endometrial-\nlike tissues outside of the uterus, most commonly in the \nperitoneal cavity and associated with inflammation [1]. \nWhile the pathogenesis of endometriosis is not under -\nstood, retrograde menstruation of endometrial cells and \nOpen Access\n*Correspondence:  cmetz@northwell.edu; pgregers@northwell.edu\n2 Donald and Barbara Zucker School of Medicine, 500 Hofstra Blvd, \nHempstead, NY, USA\nFull list of author information is available at the end of the article\n\nPage 2 of 16Shih et al. BMC Medicine          (2022) 20:315 \ntissues via the fallopian tubes is one accepted theory for \nthe development of endometriosis lesions in the peri -\ntoneal cavity [2, 3]. However, retrograde menstruation \noccurs in nearly all women [4], yet endometriosis occurs \nin approximately one in ten females in their reproduc -\ntive years [3]. Thus, other factors must contribute to the \ndevelopment of endometriosis. While there is a signifi -\ncant genetic component to endometriosis [5], very little \nis known about how these putative risk alleles function. \nOn the other hand, the eutopic endometrium of patients \nwith endometriosis is significantly different when com -\npared to the endometrium of those without endome -\ntriosis, with inflammatory changes noted in the setting \nof endometriosis [6–9]. We have undertaken a detailed \nanalysis of endometrial tissues and cells present in men -\nstrual effluent (ME), since ME is the critical biological \nsample transferred to the pelvic cavity, where most endo-\nmetriosis lesions grow.\nMost previous investigations of ME have involved \nthe phenotypic analysis by immunofluorescence, flow \ncytometry, and/or in  vitro culture of single-cell suspen -\nsions collected using menstrual cups [10–14]. Our previ -\nous flow cytometry studies showed that uterine natural \nkiller (uNK) cells were relatively depleted in ME from \nendometriosis cases vs. controls [11]. However, this study \nwas limited by the analysis of relatively few cell types in \nME, with no assessment of specific subsets of cells. In \naddition, we demonstrated a defect in decidualization \ncapacity of endometrial stromal cells grown from the ME \nof patients with endometriosis when compared to ME-\nstromal cells grown from healthy controls [11, 15]. While \nthese earlier results potentially provided a basis for a \nscreening test for endometriosis, these analyses relied on \nlaborious and expensive cell culture and in  vitro assays, \nmaking them impractical for clinical application.\nHerein, we investigated fresh ME as an unexplored \nand important biological specimen for the development \nof non-invasive diagnostics based on the direct analysis \nof endometrial tissue fragments. We show that ME con -\ntains large numbers of shed fragments from endometrial \ntissues. Using enzymatic digestion of ME and associated \ntissues followed by single-cell RNA sequencing (scRNA-\nSeq) analysis, we compared the major cellular differences \nand gene expression profiles found in ME collected from \nhealthy controls (without symptoms of endometriosis) \nand patients diagnosed with endometriosis (confirmed \nby laparoscopic surgery with positive confirmation by \npathology), as well as patients with symptoms of endo -\nmetriosis (e.g., recurrent dysmenorrhea, persistent \nabdominal bloating, dyspareunia, dysuria, and/or dysche-\nzia) who are not yet diagnosed. In order to gain insight \ninto the pathogenesis of endometriosis, we particularly \nfocused on the phenotypes of stromal and uNK cells in \nME through scRNA-Seq because these are abundant and \nhave been previously shown to be abnormal in  patients.\nMethods\nHuman subjects and menstrual effluent collections\nMenstrual effluent (ME) was collected as previously \ndescribed [11, 15]. Briefly, women of reproductive age \n(N = 33, age 20–45 years, average age 33.6 years) living in \nNorth America who were not pregnant or breastfeeding, \nwho were menstruating, and who were willing to provide \nME samples were recruited mainly via social media and \nconsented to the ROSE study (IRB#13-376A; https:// feins \ntein. north well. edu/ insti tutes- resea rchers/ insti tute- molec \nular- medic ine/ robert- s- boas- center- for- genom ics- and- \nhuman-  genet  ics/ rose- resea  rch- outsm  arts-  endom etrio \nsis). Women with histologically confirmed endometrio -\nsis (determined following excision laparoscopic surgery \nand documented in a pathology report without revised \nAmerican Society for Reproductive Medicine (rASRM) \nstaging/classification) were enrolled as “endometriosis” \nsubjects (N = 11). Women who reported chronic symp -\ntoms consistent with endometriosis (e.g., recurrent dys -\nmenorrhea, persistent abdominal bloating, dyspareunia, \ndysuria, and/or dyschezia), but not yet diagnosed with \nendometriosis (or not) were enrolled as ‘symptomatic’ \nsubjects (N = 13). Control subjects living in North Amer -\nica who self-reported no gynecologic history suggestive \nof a diagnosis of endometriosis (and the absence of poly -\ncystic ovarian syndrome, and pelvic inflammatory dis -\nease) were recruited mainly via social media and enrolled \nas “controls” (N = 9).\nEndometriosis, symptomatic, and control subjects col -\nlected their ME using an “at home” ME collection kit for \n4–8 h on the day of their heaviest menstrual flow (typi -\ncally day 1 or 2 of the cycle) with a menstrual cup (pro -\nvided by DIVA International), except for one subject \nwho collected ME using a novel menstrual collection \nsponge (as previously described [15]). After collection, \nME was shipped priority overnight at  4°C to the labora -\ntory for processing. ME collected from menstrual cups \nwas mixed 1:1 with DMEM for processing. For the sat -\nurated menstrual collection sponge, ME tissue was col -\nlected after rinsing the sponges with PBS to collect cells \nand tissue. Demographic and gynecologic/health data \n(including hormone usage, menstrual cycle information, \nand pain/pain medications) for controls, endometriosis \nsubjects, and symptomatic subjects (and the total cohort) \nare shown in Table 1.\nImmunostaining of ME‑derived tissue fragments\nME-derived tissue fragments were obtained from con -\ntrols, symptomatic subjects, and endometriosis patients \n(n = 2 each); tissue fragments were collected by pouring \n\nPage 3 of 16\nShih et al. BMC Medicine          (2022) 20:315 \n \nME over a 70μ filter, fixed, and transferred to the clini -\ncal pathology lab for paraffin embedding and hema -\ntoxylin and eosin (H&E) staining. CD10 was chosen for \nimmunohistochemical analysis because it is a sensitive \nmarker of eutopic endometrial stroma [16] and because \nadjunctive use of CD10 immunostaining with H&E \nstaining enhances the histologic detection of endome -\ntriosis [17]. CD56 was chosen because uNK cells stain \nbrightly with CD56. H&E slides and immunostained \nslides were examined microscopically and imaged by a \npathologist. Representative images are shown in Fig. 1 .\nProcessing menstrual effluent for scRNA‑Seq analyses\nWhole (unfractionated) ME (2.5–10  ml) was digested \nwith Collagenase I (1  mg/ml, Worthington Biochemi -\ncal Corporation, Lakewood, NJ) and DNase I (0.25mg/\nml, Worthington Biochemical Corporation) at 37  °C for \n10–30 min using the  gentleMACSTM Tissue Octo Disso -\nciator (Miltenyi Biotec, Cambridge, MA) using C tubes \nTable 1 Subject group characteristics—control (CTRL), diagnosed (Dx), and symptomatic (Sx)\nNo comparisons of CTRL vs. Sx were significant for the above characteristics, except pain (yes/no): *p = 0.008; +one patient used 200mg of progesterone (Prometrium \nvaginal suppositories) 2 days prior to menses\nCTRL Dx Sx Total P value \n(CTRL vs \nDx)\nAge (years) (mean ± SD) 33.4 ± 5.4 35.2 ± 4.4 32.3 ± 8.2 33.6 ± 6.3 0.42\nBMI (kg/m2) (mean ± SD) 24.2 ± 7.1 28.4 ± 5.4 26.8 ± 7.6 26.5 ± 6.8 0.15\nAge at menarche (years) (mean ± SD) 12.0 ± 0.9 11.6 ± 1.6 12.6 ± 1.7 12.0 ± 1.5 0.51\nRace/ethnicity 0.34\n Caucasian 7/9 (78%) 0/11 (91%) 13/13 (100%) 30/33 (91%)\n Black 1/9 (11%) 0/11 (0%) 0/13 (0%) 1/33 (3%)\n Mixed 1/9 (11%) 0/11 (0%) 0/13 (0%) 1/33 (3%)\n Other 0/9 (0%) 1/11 (9%) 0/13 (0%) 1/33 (3%)\n Hispanic 0/9 (0%) 0/11 (0%) 1/13 (8%)\nTypical cycle length (days) 0.52\n 21–25 days 0/9 (0%) 1/11 (9%) 3/13 (23%) 4/33 (12%)\n 26–31 days 8/9 (89%) 7/11 (64%) 8/13 (62%) 23/33 (70%)\n 32–39 days 1/9 (11%) 2/11 (18%) 2/13 (15%) 5/33 (15%)\n > 40 days 0/9 (0%) 1/11 (9%) 0/13 (0%) 1/33 (3%)\nTypical bleed time (days) 0.36\n < 3 days 0/9 (0%) 2/11 (18%) 0/13 (0%) 2/33 (6%)\n 3–5 days 5/9 (56%) 6/11 (55%) 8/13 (62%) 19/33 (58%)\n 6–8 days 4/9 (44%) 3/11 (23%) 5/13 (38%) 12/33 (36%)\nTypical flow 0.96\n Light 1/9 (11%) 1/11 (9%) 0/13 (0%) 2/33 (6%)\n Moderate 2/9 (22%) 2/11 (18%) 3/13 (23%) 7/33 (21%)\n Moderately heavy 3/9 (33%) 5/11 (45%) 9/13 (69%) 17/33 (52%)\n Heavy 3/9 (33%) 3/11 (23%) 1/13 (8%) 7/33 (21%)\nHormone use 0.35\n Yes 0/9 (0%) 1/11 (9%)+ 1/13 (8%) 2/33 (6%)\nPain in this cycle 0.06\n Yes 5/9 (56%) 10/11 (91%) 13/13 (100%)* 28/33 (85%)\n None 4/9 (44%) 1/11 (9%) 0/13 (0%) 5/33 (15%)\n Mild 2/9 (22%) 3/11 (23%) 3/13 (23%) 8/33 (24%)\n Moderate 3/9 (33%) 6/11 (55%) 4/13 (31%) 13/33 (40%)\n Severe 0/9 (0%) 1/11 (9%) 6/13 (46%) 7/33 (21%)\nPain medication in this cycle (Midol, Advil, Tylenol, Naproxen, Hydromorphone 2 mg/Baclofen/diazepam/Ketamine 8/10/15 mg)\n Yes 1/9 (11%) 6/11 (55%) 10/13 (77%) 17/33 (52%)\n\nPage 4 of 16Shih et al. BMC Medicine          (2022) 20:315 \nand Program 37CMulti_E_01 (31 min). After diges -\ntion, the sample was sieved over a 70μ filter and washed \nwith DMEM 10% fetal bovine serum (FBS) to neutralize \ndigestion enzymes; the flow through was sieved over a \n40μ filter and washed with DMEM 10%FBS. After col -\nlecting the single cells (from the flow through) follow -\ning centrifugation (350×g for 5 min), neutrophils were \nremoved using the  EasySepTM HLA Chimerism Whole \nBlood CD66b Positive Selection Kit (STEMCELL, Cam -\nbridge, MA), according to the manufacturer’s protocol. \nThe neutrophil pellet was frozen at – 80 °C and used as \na source of subject DNA for genotyping (see below). The \nresultant cells were depleted of red blood cells using the \nEasySep™ RBC Depletion Reagent (STEMCELL), accord-\ning to the manufacturer’s protocol, and then washed and \nsubjected to density gradient centrifugation using Ficoll-\nPaque PLUS (Sigma-Aldrich, St. Louis, MO) to collect \nmononuclear cells, according to manufacturer’s direc -\ntions. To collect ME-tissue, whole ME (2.5–10  ml) was \nsieved over a 70μ filter and washed with DMEM; the ME-\ntissues trapped on the filter were collected and digested \nwith Collagenase I (1mg/ml, Worthington Biochemi -\ncal Corporation, Lakewood, NJ) and DNase I (0.25mg/\nml, Worthington Biochemical Corporation) at 37  °C for \n10 min and processed as described above for whole ME, \nexcept without a density gradient centrifugation step. The \nresultant whole ME cells were enumerated, and viability \nwas assessed using ViaStain ™ AOPI Staining Solution \nand the Nexcelom Cellometer (Lawrence, MA). Prepara -\ntions with > 80% viability were processed for scRNA-Seq. \nCells were immediately fixed in methanol for scRNA-Seq, \nas described by Chen for peripheral blood mononuclear \ncells [18]. Briefly, cells were washed and resuspended in \na 200  μl  Ca++ and  Mg++-free PBS, followed by drop -\nwise addition of chilled 100% methanol (800 μl, final 80% \nw/v). Fixed cells were stored at –  20  °C for 20  min and \nthen stored at – 80 °C until used for scRNA-Seq. A pilot \nexperiment was performed with a single ME sample, \nwhich was processed and either prepared immediately \nfor scRNA-Seq (without methanol fixation and freezing) \nor was fixed in methanol and frozen, as described above \nto optimize our scRNA-Seq methods. The data showed \nalmost identical scRNA-Seq results using both methods \n(see Additional file  1) reassuring us that the methods of \nChen et al. [18] can be applied to ME samples. Thus, all \nME samples were methanol fixed and frozen. An advan -\ntage of this approach is that it allows cost-effective pool -\ning of samples collected at different times and reduces \nthe potential batch effects of running samples separately \nfor scRNA-Seq.\nProcessing of samples for single‑cell sequencing\nMethanol-fixed cells were removed from –  80  °C and \nplaced on ice for 5 min before centrifugation (1000×g \nfor 5 min). Methanol-PBS supernatant was completely \nremoved and cells were rehydrated in 0.04% bovine \nFig. 1 ME contains endometrial tissues. Histological analysis of endometrial tissues isolated from the menstrual effluent (ME) from 4 separate \nsubjects: A control subject, B, C two subjects with pathologically confirmed endometriosis, and D subject with chronic symptoms of endometriosis \n(not yet diagnosed). Upper panels for A–D: H&E staining is shown in two panels at two magnifications for each individual: A 40X (left) and 200X \n(right), B 100X and 200X, C 100X and 200X, and D 40X and 200X; arrowheads point to glandular epithelium. Sections show typical late secretory/\nmenstrual endometrium with expanded stroma containing scattered inflammatory cells and secretory and inactive type glands. Lower panels for \nA–D: immunostaining with anti-CD10 and anti-CD56 antibodies to detect stromal cells (left) and uterine NK (uNK) cells (right), respectively, at 100X. \nScale bars are shown in each image\n\nPage 5 of 16\nShih et al. BMC Medicine          (2022) 20:315 \n \nserum albumin (BSA) + 1  mM dithiothreitol (DTT) + \n0.2 U/μl RNase Inhibitor in 3X SSC (saline sodium citrate \nbuffer solution) Buffer (Sigma). An aliquot of fixed cells \nwas stained with Trypan Blue and visualized under the \nmicroscope. The cells were counted and pooled from dif -\nferent donors at equal ratios, filtered using 40 µ strainer \n(Falcon), recounted and brought up to a final conc. of \n2000 cells/μl, and proceeded immediately for GEM gen -\neration and barcoding on a 10X Chromium using Next \nGEM 3’ v3.1 reagents (10X Genomics). Libraries were \nconstructed following 10X Genomics’ recommendations \nand quality was assessed on a High Sensitivity DNA chip \non a BioAnalyzer 2100 (Agilent) before loading (1.8 pM) \nand sequencing on an Illumina Nextseq 500 using a High \nOutput kit v2.5 (150 cycles).\nFive subjects were pooled together into a single 10X \nlane with at least one of each phenotype per run with a \ntotal of 8 runs for ME-tissue and 3 runs of whole ME. The \nME-tissue runs had 44,135 total cells, of which 5147 had \nambiguous calls in Demuxlet, 2632 were doublets, and \n36,356 were singlets; only the singlets were analyzed. The \nwhole-ME runs had 30,090 total cells, of which 5556 had \nambiguous calls, 2776 were doublets, and 21,758 singlets \n(and hence analyzed). A total of 43,054 cells were ana -\nlyzed in this study following filtering and QC (thresholds \nof > 10% mitochondrial reads < 500 nUMI (number of \nunique molecular identifiers) or > 50000 nUMI or > 6000 \nunique features per cell).\nSingle‑cell RNA sequencing and analyses and statistics\nSamples were converted from raw bcl files to gene by cell \nmatrices using CellRanger 6.0 aligned to 10x Genomics’ \nGRCh38-3.0.0 reference. Individuals were demultiplexed \nvia Demuxlet [19] using genotypes taken from SNPs on \nthe Illumina GSAv3 genotyping array, run on DNA pre -\npared from neutrophils isolated from ME. The thresholds \nin Demuxlet were adjusted to the expected doublet rate, \nand those marked as doublets were removed. Down -\nstream analysis and visualization were done using Seurat \n4.0 [20]. Briefly, there were at least 25,000 reads per cell \non average per 10X run, and the mean number of genes \ncaptured was 1388 (± 896) (mean ± standard devia -\ntion [SD]). There was no significant difference between \nthe various clinical groups (controls, cases, symptomat -\nics) in these values. Genes were filtered out if they were \nexpressed in less than 3 cells while cells were filtered \nout if they had > 10% mitochondrial reads, 500 < nUMI \n< 50000 and > 6000 unique features. For the analysis of \nME-tissue samples, only subjects with information on at \nleast 500 cells per subject were retained. After filtering, \nthe cell yields were comparable in each group (mean ± \nSD: 1256 ±732 and 1319 ±767 in ME-tissue and whole \nME, respectively). Gene expression normalization and \ncell clustering was done using the SCTransform pipeline \n[21] with percent mitochondrial reads regressed out and \nperson specific batch effects corrected using Harmony \n[22]. Identification of cell clusters was done using known \nmarker genes (Additional file  2) [23–30] with differen -\ntial gene expression calculated using a Wilcoxon rank \nsum test. Enrichment of cell clusters of specific pheno -\ntypes was done using MASC (mixed-effects modeling of \nassociations of single cells) (https:// github. com/ immun \nogeno mics/ masc), which essentially uses a percentage of \ncells per cluster while also taking into account technical \ncovariates; 10X library batch, preparation (whole ME or \nME-tissue), nUMI per cell, percent mitochondrial reads, \nand phase are accounted for. All datasets are deposited \nin the National Center for Biotechnology Information/\nGene Expression Omnibus (GEO) accession number \nGSE203191.\nResults\nEndometrial tissue fragments are present in fresh \nmenstrual effluent\nWe carried out histological assessment of fresh men -\nstrual effluent (ME)-associated tissues isolated from ME. \nRepresentative H&E sections of ME-derived tissue frag -\nments from four subjects (1 control, 2 laparoscopically/\nhistologically confirmed endometriosis subjects, and 1 \nsymptomatic subject) show the presence of endometrial \ntissues with mucosal and glandular epithelium and areas \nof stroma. The endometrium had typical late secretory/\nmenstrual morphology with expanded stroma containing \nscattered inflammatory cells and secretory and inactive-\ntype glands (Fig.  1A–D, upper panels). Immunostaining \nof ME-derived tissue sections reveals a range of stromal \ncells stained with antibodies to CD10, a clinically used \nmarker of endometrial stroma [16, 17], and an abundance \nof uNK cells (stained with antibodies to CD56 (NCAM), \nan archetypical marker of NK cells [Fig.  1A–D, lower \npanels]).\nSingle‑cell RNA sequencing (scRNA‑Seq) of digested \nfreshly processed ME reveals the presence \nof a heterogeneous mixture of immune and non‑immune \ncells\nWe have analyzed ME samples from 33 subjects, \nincluding age-matched healthy controls (N  = 9), endo -\nmetriosis cases (N  = 11), and subjects with chronic \nsymptoms suggestive of endometriosis but not yet diag -\nnosed (N  = 13) (see Table  1). ME samples from either \nwhole ME (unfractionated) or ME samples enriched for \ntissues (“ME-tissue”) were digested with collagenase I \nand DNase I, depleted of neutrophils, and processed for \nscRNA-Seq, as described in the methods. As shown in \nFig.  2, a graph-based clustering approach using Seurat \n\nPage 6 of 16Shih et al. BMC Medicine          (2022) 20:315 \ndistinguishes multiple cell clusters shown on the UMAP \n(uniform manifold approximation and projection) plot. \nThere is striking diversity of the cell types defined by \nthe cluster analysis. A major group of uterine NK cells \nis designated cluster uNK1, with a small associated \ncluster designated uNK2. Sets of clusters related to \nCD8+ and CD4+  T cells are shown in the central por -\ntion of the plot. Endometrial stromal cells and epithe -\nlial cells are identified in major clusters in the right side \nof the UMAP plot. Subclusters of endometrial stromal \ncells are described below in detail. Based on [31], Epi -\nthelial1 appears to be a mix of lumenal and glandular \nepithelial cells, Epithelial2 is comprised of ciliated epi -\nthelial cells, and Epithelial3 is a separate set of CD326-\nexpressing cells that do not overlap with Epithelial1 or \nEpithelial2. Distinct clusters of B cells and myeloid cells \ncan also be delineated, along with a small cluster of \nplasmacytoid dendritic cells (pDC). The positive gene \nmarkers used to generate the cell clusters shown in \nFig. 2 are included in Additional file  2. Overall, the vari -\nous cell clusters are well represented whether unfrac -\ntionated whole ME or tissue-enriched ME is processed \nfor scRNA-Seq. Some differences in cell subset fre -\nquencies can be observed; in particular, epithelial cells \nwere enhanced when tissue-enriched ME was utilized \nfor sample processing (see Additional file 3 ).\nCell clusters from ME containing endometrial tissue \ndiffer between endometriosis cases and healthy controls; \nrelative depletion of uterine NK cells and enrichment of B \ncells in endometriosis cases\nWe compared the relative frequency of the various \ncell clusters in the freshly processed ME obtained \nfrom the diagnosed endometriosis cases (N  = 11) \ncompared with controls (N  = 9), as shown in Fig.  3. \nBy inspection of Fig.  3, it is apparent that both clus -\nters of uNK cells (uNK1 and uNK2) are markedly \ndepleted in the cases vs. controls (average percent -\nage of uNK approximately 8% in cases, 28% in con -\ntrols), as well as an increase in the proportion of B \ncells in cases (~9%) vs. controls (~3%). The odds \nratios and confidence intervals for these two cell \nenrichment patterns are shown in Fig.  4, along with \nthe patterns of enrichment of all the other major cell \nclusters. While there is some variation among many \nof the different cell clusters, a formal analysis shows \nthe most striking differences are observed for uNK \ncells, which are enriched in controls (and depleted in \ncases; uNK1, P  < 10E− 16; uNK2, P  < 10E− 16), along \nFig. 2 Cellular composition of digested ME based on scRNA-Seq. UMAP plot for all 33 digested menstrual effluent (ME) samples (controls = 9; \nendometriosis cases = 11; symptomatic cases = 13). Several well-delineated cell clusters include a large cluster of uterine NK cells (uNK1), as well as \nclearly separated stromal cells, epithelial cells, and B cells. Several clusters each of T cells and myeloid cells are also defined, as well as a small cluster \nof plasmacytoid dendritic cells (pDC). A small cluster of approximately 60 unknown cells is in the lower right corner. The positive gene markers used \nto generate the cell clusters shown are included in Additional file 2.\n\nPage 7 of 16\nShih et al. BMC Medicine          (2022) 20:315 \n \nwith a relative enrichment in the proportion of B \ncells in the cases diagnosed with endometriosis (and \nrelatively depleted in controls; P  < 10E− 16). Note \nthat the stromal cell cluster is not significantly differ -\nent between cases and controls (P  > 0.05).\nWe also explored whether the various proportions of \ncell clusters of the ME preparations from the “sympto -\nmatic” but undiagnosed group of subjects (N = 13) are \ndifferent from ME preparations from controls. This is \nclearly the case, as shown in Additional file  4. Here, we \nshow the relative enrichment of uNK cells is maintained \nin controls in comparison to the symptomatic group \n(uNK1, P < 10E−16; uNK2, P = 0.0025), similar to that \nobserved with ME from cases. B cells also show a sig -\nnificant relative enrichment in symptomatic as well as \ndiagnosed cases, compared with controls (symptomatic \nvs. control, P = 5.8 ×  10−6), similar to that observed with \nME from cases (Additional file  4). Perhaps not surpris -\ningly, these significant differences in symptomatic cases \nvs. controls are less striking than the differences in endo -\nmetriosis cases vs. controls, given the likely heterogeneity \nof the symptomatic group.\nDecidualized stromal cell subclusters are reduced \nin endometriosis\nPrevious studies have reported reduced decidualization \ncapacity in endometrial stromal cells grown from biop -\nsies of patients with endometriosis [32]. We have also \nFig. 3 Distinct cellular composition differences in digested ME from endometriosis cases vs. controls are revealed by scRNA-Seq. The data taken \nfrom the UMAP plot in Fig. 2 is separated into two groups: controls (n = 9, providing 14,327 cells) and endometriosis cases (n = 11, providing 11,924 \ncells). The most striking difference is the increased fractions of uterine NK cells (uNK1 and uNK2) in the endometrial tissues of controls as compared \nto cases. In contrast, B cells are significantly enriched in cases. A formal analysis of enrichment is given in Fig. 4 and confirms the significant \nenrichment of uNK cells and B cells in controls and cases, respectively. The positive gene markers used to generate the cell clusters shown are \nincluded in Additional file 2\nFig. 4 Analysis of enrichment of cell subsets in ME comparing \nendometriosis cases and controls. These data are taken from data \nshown in Fig. 3. The  Log2 odd ratios (OR) with cell subsets enriched \nin controls on the left and cell subsets enriched in cases on the \nright. It is apparent that uterine NK (uNK) cells, both uNK1 and \nuNK2, are significantly enriched in controls, while B cells show the \ngreatest enrichment in cases. Note: Epithelial cells are excluded from \nthis analysis because their enrichment was affected by the tissue \npreparation method used. As described in the methods section, \nthese data are corrected for covariates including 10X library batch, \nsample preparation (whole ME or ME-tissue), nUMI per cell, percent \nmitochondrial reads, and cell cycle phase\n\nPage 8 of 16Shih et al. BMC Medicine          (2022) 20:315 \nobserved impaired decidualization using stromal cells \ngrown directly from ME [11, 15]. Therefore, we examined \nwhether this trend could be observed in fresh stromal \ncells analyzed by scRNA-Seq. The stromal cell numbers \nor percentages did not significantly differ between the \ncontrol and endometriosis groups, as shown in Figs.  3 \nand 4. However, subclustering of the stromal cell cluster \nclearly identified 5 subclusters of interest within the stro-\nmal cell population (Fig.  5A). We have designated these \nsubclusters based on the dominant transcripts expressed \nin each of these subclusters, as shown in the violin plots \nin Fig.  5B. Two of the five subclusters (2 and 4) are not \ndifferent between cases and controls (the top genes of \nsubclusters 2 and 4 are described in Additional file  5). \nThe subclusters showing significant enrichment in either \ncases or controls (subclusters 1, 3, and 5) are indicated \nby the  Log2 (odds ratios, [OR]) below the UMAP plot \n(Fig. 5C).\nIt is striking that an apparently decidualized stromal \ncell subcluster (expressing IGFBP1 mRNA) is signifi -\ncantly enriched in controls compared with endometriosis \ncases (Fig.  5A, B). In addition to IGFBP1, the top differ -\nentially expressed genes in this subcluster (compared to \nother stromal cell subclusters) include LEFTY2, DCN, \nLUM, MDK, C1QTNF6, APOE/D, DCN, and other pro -\ngesterone sensitive and decidualization/fertility gene \nmarkers (see left panel (subcluster 3) in Fig.  6 and Addi-\ntional file  6 [33–112]. This suggests that a phenotype of \n“decidualization” can be measured directly in stromal \ncells derived from fresh ME and is associated with con -\ntrol vs. disease phenotype. A modest enrichment of a \nsubcluster expressing IL11 was observed in cases, as indi-\ncated in Fig.  5A–C. In addition to IL11, this subcluster \nis associated with transcripts for MMP3, MMP1, MMP9, \nSERPINB2, S100A6, and CXCL8 , among other genes \nassociated with inflammation, fibrosis, and senescence, \nFig. 5 Analysis of the stromal cell subclusters. A UMAP plot of the five stromal cell subclusters are shown. B Violin plots showing the defining \ngene expression per subcluster for subclusters 1-5. C  Log2 (odds ratio) shows that subcluster 3 (IGFBP1+) is significantly enriched in controls  (Log2 \nOR = − 1.3, case vs. control). In contrast, subcluster 1 (IL11+) and subcluster 5 (MGP+) are enriched in diagnosed subjects. The top transcripts \ncharacterizing these three distinct stromal cell subclusters are summarized in Fig. 6 and emphasize the enrichment of the decidualized stromal \ncells—subcluster 3 (IGFBP1+)—in controls\n\nPage 9 of 16\nShih et al. BMC Medicine          (2022) 20:315 \n \nas well as endometriosis, as shown in the middle panel \n(subcluster 1) of Fig. 6 (and Additional file 6). A third sub-\ncluster, designated by high expression of the gene encod -\ning matrix Gla protein (MGP ), is also enriched in the \nstromal cells of cases (Fig.  5A–C). This subset expresses \nnumerous extracellular matrix genes that have been \nassociated with presence of perivascular stromal cells, \nsenescence, and cell adhesion/cell spreading, including \nFN1 (which encodes fibronectin-1), a known risk locus \nfor endometriosis [113]. Figure  6 (right panel (subcluster \n5) and Additional file  6) also shows the list of top genes \nexpressed in this subset. Additional file  7 demonstrates \nthat the IGFBP1+ and MGP + subclusters map to stro -\nmal cells subsets defined in the decidua found in the first \ntrimester of pregnancy by Vento-Tormo et al. [26].\nWe also examined the differences between cases and \ncontrols in the two uNK subclusters present in digested \nendometrial tissues in ME (uNK1 and uNK2, see Fig.  2). \nWe noted a distinct subcluster of uNK cells (uNK2) that \nis characterized by the expression of genes associated \nwith cell proliferation such as MKI67 which encodes \nKi67 and TOP2A which encodes topoisomerase 2A (see \nAdditional file  8 for a full uNK subcluster analysis). As \ndiscussed below, this cluster also mapped nearly exactly \n(97%) with a proliferative subset of uNK cells that has \nbeen defined by scRNA-Seq in decidua obtained during \nthe first trimester of pregnancy [26]. This is consistent \nwith the proliferation of uNK cells and overall accumula -\ntion of uNK cells in the course of decidualization in con -\ntrol subjects vs. cases, as shown in Figs. 3 and 4.\nFinally, in order to address the possibility that the use \nof hormone treatment by one endometriosis subject (see \nTable 1) might have affected our results, we performed a \nreanalysis of cases and controls after eliminating this sub-\nject. The results are not significantly changed.\nSpecifically, the overall spatial distribution of the \nUMAP is not different when the cells from the endome -\ntriosis subject on hormones were removed from the data \n(Additional file  9), and the  log2 odd ratios (OR) for the \ncell subsets enriched in controls and cell subsets enriched \nin cases are not different (Additional file  10). Addition -\nally, the clustering of the endometrial stromal cells is not \ndifferent (Additional file  11) from the original set (with \nall cases). Comparing all differentially expressed genes \nwith/without the hormone sample in stromal cell sub -\nsets and correlating log fold changes of significant genes \nwithin each subcluster shows they are highly correlated, \nR> 0.997 for all samples (data not shown).\nDiscussion\nThese studies show for the first time that the phenotype \nof eutopic endometrial tissue shed into the menstrual \neffluent is distinct in patients with endometriosis com -\npared to control subjects. There are three major observa -\ntions. First, the endometrial stromal cells show a relative \ndeficiency of progesterone-sensitive gene markers asso -\nciated with endometrial stromal cell decidualization \nin patients with endometriosis (e.g., IGFBP1, LEFTY2, \nLUM, DCN, etc.). This is consistent with previous studies \nshowing impaired decidualization of cultured endome -\ntrial stromal cells obtained from endometrial and ectopic \nendometriosis biopsies [32, 114], as well as from men -\nstrual effluent [11, 15]. Secondly, there is a striking reduc-\ntion in the proportion of uNK cells in the ME-derived \nFig. 6 Distinct subclusters of decidualized stromal cells and pro-inflammatory stromal cells distinguish ME from controls and endometriosis cases. \nUpper panel: A summary of genes enriched in the stromal cell subclusters which are significantly enriched in cases (subclusters 1 [IL11+] and 5 \n[MGP+]) or controls (subcluster 3 [IGFBP1+]). Note: Subclusters 2 and 4 were not significantly different in cases vs. controls; see Additional file 6 \nfor the listing of genes differentially expressed in these clusters. Lower panel: Characteristic features of stromal cell subcluster gene markers. The \ndecidualized stromal cell subcluster (IGFBP1+, subcluster 3) is prominently enriched in genes that are associated with decidualization and uterine \nreceptivity and are progesterone responsive. In contrast, the non-decidualized stromal cell subsets that are enriched in cases (MGP+ [subcluster \n5] and IL11+ [subcluster 1]) are variably enriched in estrogen responsive genes, and remarkably enriched in genes associated with inflammation, \nfibrosis, and cellular senescence. Note: MGP+ (subcluster 5) is also enriched in cell adhesion and cell spreading gene markers\n\nPage 10 of 16Shih et al. BMC Medicine          (2022) 20:315 \nendometrial tissue of patients with endometriosis com -\npared with controls. This was suggested by our previous \nstudies of free cells present in ME using flow cytometry \nmethods [11], but it is clearly a major distinguishing \nfeature of the eutopic endometrium of endometriosis \npatients. Thirdly, our data suggest an enrichment of B \ncells in the eutopic endometrium of patients with endo -\nmetriosis, a finding that is consistent with the hypothesis \nthat chronic inflammation and/or chronic endometritis is \na predisposing factor in the development of endometrio -\nsis [115].\nA deficiency in the decidualization capacity of stro -\nmal cells cultured from biopsies of the eutopic endo -\nmetrium has been reported previously [32] and is also \nfound in ME-derived stromal cells collected at the time \nof menstruation [11, 15]. Our scRNA-Seq data clearly \nshows the reduction of the IGFBP1+-expressing decidu -\nalized stromal cell subcluster in endometriosis cases vs. \ncontrols (Fig.  5C). The relationship of this finding to the \npathogenesis of endometriosis is not established. One \npossibility is that this differentiation deficiency leaves \nbehind non-decidualized endometrial stromal cells that \nexhibit proinflammatory, pro-fibrotic, and/or senescent \nphenotypes. These “pathogenic” cells may then initi -\nate or promote lesions following retrograde transfer \ninto the peritoneal cavity. The enrichment of an IL11+-\nexpressing stromal cell subcluster in the endometriosis \nME samples that express many estrogen-responsive, pro-\ninflammatory, pro-fibrotic, and senescence gene mark -\ners (shown in Fig.  6 and Additional file  6) provides some \nsupport for this possibility, but this needs confirmation \nin larger datasets. The significant increase in the MGP + \nstromal subcluster in endometriosis (Fig.  5C) is also of \npotential interest. As shown in Fig.  6 (right panel), the \nMGP+ stromal cell subcluster expresses many genes that \nare associated with the extracellular matrix, including \nFN1 (encoding fibronectin-1) which has been associated \nwith an increased risk for endometriosis in GWAS stud -\nies [116]. Interestingly, most of the top markers found in \nthe IL11+ and the MGP + subclusters are either associ -\nated with senescence or induce senescence (e.g., IL11 and \nSERPINB2, see Fig. 6 and Additional file 6). Inflammation \nand senescence are key features of endometriosis and \nreduced uterine receptivity and infertility [117–119].\nAnother possibility is that the overall environment of \nthe eutopic endometrium predisposes to reduced stro -\nmal cell decidualization, independent of any direct role \nor effect on stromal cell subsets in the disease. A chronic \ninflammatory endometrial environment might lead to, \nor be associated with, other changes that put individu -\nals at risk for endometriosis. For example, the presence \nof chronic endometritis has been reported to be a sig -\nnificant risk factor for endometriosis [115, 120]; chronic \nendometritis is also associated with reduced stromal cell \ndecidualization [121]. Interestingly, the presence of B \ncells in endometrial tissue, particularly plasma cells, is a \nrequirement for the clinical diagnosis of chronic endo -\nmetritis [115]. We note the significant increase in B cells \nin shed endometrium of endometriosis patients (Figs.  3 \nand 4) and symptomatic subjects (Additional file 4) when \ncompared to controls. This may reflect an inflammatory \nstate, as B cells play an important role in mediating or \nregulating inflammatory and autoimmune diseases [122]. \nThe numbers of B cells available for detailed analysis have \nnot allowed us to fully understand the phenotype of these \ncells; this is an area for future study.\nWe have demonstrated that uNK cells are remark -\nably depleted in the ME-derived endometrial tissues of \npatients with endometriosis (Figs. 3 and 4 and Additional \nfile 4). This may reflect compromised decidualization in \nthese subjects. uNK cells are characteristic of decidual -\nizing tissues [123] and are also prominent in the decidua \nof early pregnancy [26]. To our knowledge, this is the first \nreport of proliferating uNK cells found in ME. Crosstalk \nbetween stromal cells and uNK cells is a feature that pro -\nmotes decidualization and uterine receptivity/placen -\ntal vascular remodeling [124]. uNK cells do not appear \nto play a major role in decidualization in uNK deficient \nIL15 knockout mice [125]. It remains unclear whether \nuNK cells or stromal cells are the primary driver of the \ndecidualization impairment in endometriosis. However, \nuNK cells do play a role in the maintenance of decidual \nintegrity as reported by Ashkar et  al. [126]. Brighton \nand co-workers emphasized the important role of uNK \ncells in clearing senescent decidual cells in the cycling \nhuman endometrium and their clearance is proposed \nto be important for optimal fertility [9]. A lack of uNK \ncells in the endometrium may contribute to increased \nnumbers of senescent cells observed in the stromal sub -\nclusters among endometriosis subjects and may contrib -\nute to endometriosis-associated infertility. However, it is \nplausible that a lack of decidualizing endometrial stromal \ncells (with concomitant reduced production of IL-15 and \nuNK chemo attractants) reduces the infiltration and pro -\nliferation of uNK in decidualizing zones. Defective uNK \ncell function has recently been proposed in the setting of \nendometriosis with infertility [127]. We did not observe \nIL15 expression by ME-stromal cells; this is not surpris -\ning as IL-15 expression by stromal cells peaks before \nthe mid-secretory phase. Interestingly, we did observe \nenhanced expression of IL2RB (which encodes a compo -\nnent of the IL-15 receptor) in the uNK cells of controls \ncompared with the endometriosis group. Since uNK cells \nare reported to play a role in infertility [123, 128], and \ninfertility is a common feature of endometriosis, further \nanalysis of the uNK subset will clearly be of interest.\n\nPage 11 of 16\nShih et al. BMC Medicine          (2022) 20:315 \n \nTaken together, these observations suggest a set of \ninteractions that may drive the development and/or \nprogression of endometriosis at several levels, as sum -\nmarized in Fig.  7. Stromal cell decidualization may be \ninhibited by a number of factors, including chronic \ninflammation, stress, and/or progesterone resistance. \nThis may divert stromal cells into a more proinflamma -\ntory/senescent state. Furthermore, deficient decidualiza -\ntion may also compromise the infiltration of uNK cells \ninto the decidua and therefore reduce the clearance of \nsenescent cells. Clearly, host genetic variation may influ -\nence these processes at every level of these interactions.\nIt is encouraging that many of our findings in patients \nwith pathologically confirmed endometriosis are also \npresent in a proportion of subjects with chronic symp -\ntoms that are suggestive of endometriosis, even in the \nabsence of a confirmed tissue diagnosis. The delay in \ndiagnosis of endometriosis is widely recognized as a \nmajor barrier in the management of this disease, with \ndelays of up to a decade in some subjects before the dis -\nease is recognized [129]. We recognize the limitation that \nthe symptomatic group lacks a diagnosis, and therefore, \nwe cannot assess the predictive ability of our results. \nTo address this limitation, a clinical trial is underway to \nenroll symptomatic subjects who are being evaluated \nby diagnostic laparoscopy as part of their standard care \nby collaborating surgeons; scRNA-Seq profiles of their \nME collected prior to surgery will be validated based on \nthe results of their laparoscopic diagnosis. Such a study \ndesign will be required to establish the positive and nega-\ntive predictive value of menstrual tissue analysis in a real-\nworld clinical setting where an endometriosis screening \ntest might be applied.\nThe method of using methanol fixation has been widely \nused [18], along with other approaches such as cryo -\npreservation, for scRNA-Seq. Despite its advantages, it \nis possible that methanol fixation may change scRNA-\nSeq outcomes and relative cell numbers. Therefore, it \nwill be important to replicate these data using a vari -\nety of methods. Due to the cost and complexity of the \nanalysis, an scRNA-Seq approach is unlikely to become \na diagnostic test for endometriosis. However, we pro -\npose that the data obtained from scRNA-Seq can be \nleveraged to develop future diagnostic and/or screen -\ning tests. What should such screening tests involve? It \nwill likely include an assessment of gene expression pat -\nterns among ME-derived stromal cells or uNK cells (or \nspecific stromal cell and uNK subsets). An initial analysis \nof stromal cell clusters suggests several potentially use -\nful gene expression differences among cases vs. controls \n(Additional file 12). Differences are also observed in uNK \ncells (Additional file 13) or indeed may be found in other \ncell types as well. On the other hand, if it can be adapted \nto a clinical diagnostic test, scRNA-Seq analysis of these \ntissues is likely to be the most informative approach, \nperhaps having more global utility to establish complex \nFig. 7 A disease model for endometriosis. Defective endometrial stromal cell decidualization may be driven by multiple factors including \ninflammation, chronic endometritis, stress, and/or progesterone resistance. This, in turn, may direct stromal cell differentiation in the direction of \nchronic inflammation and senescence, with accompanying senescence-associated secretory phenotypes (SASPs), which include pro-inflammatory \nmediators and proteases. The senescence phenotype may also impair decidualization. Reduced decidualization may also compromise the \ninfiltration and proliferation of uNK cells, which are likely to be important for senescent cell removal. Further analysis of other cells in menstrual \neffluent will be important to provide further support for this model\n\nPage 12 of 16Shih et al. BMC Medicine          (2022) 20:315 \nand heterogeneous disease subtypes, as well as predict -\ning or following response to therapy. Additional pheno -\ntypes that can be uncovered using scRNA-Seq analysis on \nlarger populations may yet yield additional biomarkers \nthat can be incorporated into a more targeted multivari -\nate biomarker analysis for diagnostic purposes.\nIn any case, the integration of our findings into a unified \npicture of the pathogenesis of endometriosis will require \nadditional scRNA-Seq studies of larger heterogeneous \npopulations, at different stages of disease development and \ninclude deeper analysis of T cells, B cells, myeloid cells, and \nepithelial cells. Abnormalities of the eutopic endometrium \nare widely recognized features of endometriosis [6– 9], and \nthis can provide diagnostic value, regardless of whether \nretrograde menstruation plays a causative role. In addition, \nit is likely that scRNA-Seq approaches of the endometrium \nvia ME may allow for improved classification of clinically \nmeaningful disease subsets and as a means for assessing \npatients’ responses to therapies, as well as uterine-associ-\nated fertility status. For example, many of the genes that \nexhibit changes in the stromal cell subclusters are associ -\nated with either estrogen or progesterone responsiveness \n(Fig. 6), and these differences could be used to guide or \nassess responses to hormonal therapies and for assessing \naspects of uterine receptivity/fertility.\nOn the other hand, if disease causation is due to retro -\ngrade menstruation of abnormal endometrial tissues, ME \nanalysis provides an opportunity to explore new therapies. \nFor example, based on the enrichment of pro-senescent \ngenes in endometriosis endometrial stomal cells (vs. control \ncells) and the deficit of uNK cells in endometriosis subjects, \nwe propose investigating senescence as a feature of endo-\nmetriosis. If this can be demonstrated, it may have poten-\ntial therapeutic implications, since various senotherapeutics \n(senolytic and senomorphic agents) have now been shown \nto improve chronic inflammatory diseases in pre-clinical \nmodels and human clinical trials [130, 131]. This is signifi-\ncant since none of the current medical therapies for endo-\nmetriosis have been shown to alter disease progression.\nConclusions\nIn summary, these scRNA-Seq data of ME collected from \nendometriosis cases and healthy controls represent a first \nattempt to globally characterize the cellular diversity of \nendometrium that is shed at the time of menstruation. \nMore detailed studies in larger datasets are clearly required, \nparticularly regarding diversity in T cells, B cells, and mye-\nloid cells, as well as epithelial cells. We propose that a com-\nprehensive assessment of cellular phenotypes in ME tissues \nwill open a new window on both diagnosis as well as pre -\nventive treatment for patients at risk for endometriosis as \nwell as other uterine and reproductive disorders.\nAbbreviations\nBSA: Bovine serum albumin; DTT: Dithiothreitol; FBS: Fetal bovine serum; GEO: \nGene Expression Omnibus; H&E: Hematoxylin and eosin; MASC: Mixed-effects \nmodeling of associations of single cells; ME: Menstrual effluent; nUMI: Number \nof unique molecular identifiers; OR: Odds ratios; pDC: Plasmacytoid dendritic \ncells; SASP: Senescence-associated secretory phenotype; scRNA-Seq: Single-\ncell RNA sequencing; SD: Standard deviation; SSC: Saline sodium citrate; \nUMAP: Uniform manifold approximation and projection; uNK: Uterine natural \nkiller.\nSupplementary Information\nThe online version contains supplementary material available at https:// doi. \norg/ 10. 1186/ s12916- 022- 02500-3.\nAdditional file 1. Comparison of UMAP plots of ME digests using fresh \ncells vs. methanol fixation. Virtually identical UMAP plots are observed \nusing either fresh or methanol fixed cells from a single subject. The ME \nsamples were prepared by tissue enrichment and tissue digestion, fol-\nlowed by scRNA-Seq analysis, as described in the methods section.\nAdditional file 2. Markers for UMAP plots of major cell clusters.\nAdditional file 3. Cell cluster composition of ME is similar when ME is \nanalyzed after enrichment for endometrial tissues and digested or ana-\nlyzed as digested whole ME. Comparison of UMAP plots is shown for ME \nsamples prepared by tissue enrichment of menstrual effluent (“ME-Tissue”; \n6 diagnosed subjects and 5 controls) or when tissue digestion is applied \nto unfractionated ME (“whole ME” , 5 diagnosed subjects, and 4 controls). \nThe various cell types are generally well represented between the two \napproaches to ME preparation. Of note there is an increased yield of epi-\nthelial cells in ME samples enriched for tissue. The positive gene markers \nused to generate the cell clusters shown are included in Additional file 2.\nAdditional file 4. Cell clusters of ME samples distinguish endometriosis \ncases and symptomatic cases vs. controls. The combined UMAP plot \nshown in Fig. 1 is split into controls (n = 9), cases (n = 11), and subjects \nwith suggestive symptoms of endometriosis but without laparoscopic tis-\nsue diagnosis – the “symptomatic” group (n = 13). Comparisons of uterine \nNK (uNK) cell and B cell frequencies in the symptomatic group show a \ntrend that is similar to cases vs. controls. The positive gene markers used \nto generate the cell clusters shown are included in Additional file 2.\nAdditional file 5. Top 10 genes in groups 2 and 4 stromal cell subclusters \nfrom Figs. 5 and 6. \nAdditional file 6. References for genes described in Fig. 6. \nAdditional file 7. Stromal cell subclusters in ME samples map to stromal \ncell clusters found in first trimester decidua. We have compared the map-\nping of stromal subclusters reported by Vento-Tormo [26], based on the \nanalysis of decidua in the first trimester, with the mapping of stromal cell \nsubclusters we have described in menstrual effluent (ME). Note that our \nIGFBP1+ subcluster maps almost identically to the decidualizing stromal \ncell subset dS2 defined by Vento-Tormo [26]. In addition, our MGP+ \nsubset shows a substantial overlap with dP2 and dP1 of Vento Tormo \n[26], subsets which are attributed to the perivascular stromal cells in first \ntrimester decidua.\nAdditional file 8.  uNK subclusters reveal a proliferating uNK subcluster \nenriched in control ME. We have examined subclusters of uterine NK (uNK) \ncells in our dataset and identified a subcluster whose gene expression \npatterns reflect cell proliferation, with substantial enrichment of MKI67 \nand TOP2A. This subset is over 98% matched to a proliferative uNK cell \nsubcluster defined in the decidua of first trimester pregnancy [26]. This \nsubset corresponds to our subset uNK2 that is enriched in controls (Fig. 3).\nAdditional file 9. The UMAP plot derived from a reanalysis of endome-\ntriosis cases (n=10) and controls (n=9) after removal of one subject on \nhormones. For this reanalysis, a total of 1112 singlet cells were eliminated \nfrom the ME-tissue run from one affected subject on hormones. Only the \nsinglets were analyzed for this revised figure which shows only subtle \nchanges in the details of the UMAP shown in Fig. 3 of the main text.\n\nPage 13 of 16\nShih et al. BMC Medicine          (2022) 20:315 \n \nAdditional file 10. A reanalysis of endometriosis cases and controls after \nremoval of one subject on hormones. The  Log2 odd ratios (OR) with cell \nsubsets enriched in controls (n=9) on the left and cell subsets enriched \nin cases (n=10)  on the right. As is the case for Fig. 4 in the main text, it is \napparent that uterine NK (uNK) cells, both uNK1 and uNK2, are signifi-\ncantly enriched in controls, while B cells show the greatest enrichment \nin cases. As noted in the methods section, these data are corrected for \ncovariates including 10X library batch, sample preparation (whole ME or \nME-tissue), nUMI per cell, percent mitochondrial reads and phase.\nAdditional file 11. A reanalysis of stromal cells from endometriosis \ncases (n=10) and controls (N=9) after removal of one subject on hor-\nmones. Not surprisingly, this revised figure shows alterations in the spatial \ndistribution of the UMAP compared to Figure 5 in the main text, but no \nsignificant differences in the cell subset distribution comparing cases and \ncontrols.\nAdditional file 12. Stromal cells exhibit distinguishing gene markers \ndifferentially regulated in ME from endometriosis cases (n=11) and con-\ntrols (n=9). Violin plots of the top 10 genes that distinguish endometriosis \ncases and controls within the total stromal cell population in ME. The \ndata suggest that IL11 and other transcripts may be useful in distinguish-\ning stromal cells isolated from ME obtained from endometriosis case vs \ncontrol subjects.\nAdditional file 13.  uNK cells exhibit distinguishing gene markers dif-\nferentially regulated in ME from endometriosis cases and controls. Violin \nplots of the top 10 genes that distinguish endometriosis cases (n=11) \nand controls (n=9) in an analysis of the uNK1 and uNK2 cell subsets as a \nwhole. Note that IFITM2 and DNAJA1 expression are substantially higher \nin ME obtained from endometriosis cases and may provide a useful \ndiagnostic target based on uNK cells that could be purified from tissues \nisolated from ME.\nAcknowledgements\nWe are grateful to the Endometriosis Foundation of America and to Dr. Tamer \nSeckin for providing early support for this work and the for the ongoing sup-\nport from the Northwell Health Innovation Award as well as the constant sup-\nport provided to PKG by the family of Robert S. Boas. We also thank Anthony \nLiew, Cassie Pond, and Maruf Chowdhury who provided valuable technical \nsupport for this project. We are especially grateful to the many extraordinary \npatients and volunteers without whose participation this project could not \nhave been accomplished.\nAuthors’ contributions\nConceptualization: CNM, PKG. Recruitment and enrollment: KE, MDF. Data \ncollection: AJS, RPA, KE, HK, MDF. Formal analysis: PKG, CNM, AJS, RPA. Sample \nprocessing: RP , PKC, HV, RH, AN. Pathology slide review: AMT. Library construc-\ntion: HK. Funding acquisition: CNM, PKG. Supervision of scRNA-Seq: ATL. Writ-\ning—original draft: PKG, CNM, AJS, RPA. Writing—review and editing: HK, PKG, \nCNM, RPA, AJS. All authors read and approved the final manuscript.\nFunding\nThis work was supported by the Northwell Health Innovations Award and the \nEndometriosis Foundation of America.\nAvailability of data and materials\nThe original data and materials presented in the study are available from the \ncorresponding authors upon reasonable request. scRNA-Seq data is available \nat National Center for Biotechnology Information/Gene Expression Omnibus \n(GEO) (accession number GSE203191).\nDeclarations\nEthics approval and consent to participate\nAll procedures for the collection of samples from research subjects were \nperformed with the approval of the institutional review board (IRB) of the \nFeinstein Institutes/Northwell Health IRB#13-376A. All participants signed \ninformed consent prior to enrollment and study participation.\nConsent for publication\nAll authors give their consent for publication of this manuscript.\nCompeting interests\nThe authors declare that they have no competing interests.\nAuthor details\n1 Robert S. Boas Center for Genomics and Human Genetics, Feinstein Institutes \nfor Medical Research, Northwell Health, 350 Community Drive, Manhasset, \nNY 11030, USA. 2 Donald and Barbara Zucker School of Medicine, 500 Hofstra \nBlvd, Hempstead, NY, USA. 3 Department of Pathology, North Shore University \nHospital, Northwell Health, 300 Community Drive, Manhasset, NY, USA. \nReceived: 18 February 2022   Accepted: 27 July 2022\nReferences\n 1. International working group of Aagl EE, Wes, Tomassetti C, Johnson NP , \nPetrozza J, Abrao MS, et al. An International Terminology for Endome-\ntriosis, 2021. J Minim Invasive Gynecol. 2021;28(11):1849–59.\n 2. Jensen JR, Coddington CC. Evolving spectrum: the pathogenesisof \nendometriosis. Clin Obstet Gynecol. 2010;2:379–88.\n 3. Zondervan KT, Becker CM, Koga K, Missmer SA, Taylor RN, Vigano P . \nEndometriosis. Nat Rev Dis Primers. 2018;4(1):9.\n 4. Halme J, Hammond MG, Hulka JF, Raj SG, Talbert LM. Retrograde men-\nstruation in healthy women and in patients with endometriosis. Obstet \nGynecol. 1984;64(2):151–4.\n 5. Zondervan KT, Becker CM, Missmer SA. Endometriosis. N Engl J Med. \n2020;382(13):1244–56.\n 6. Brosens I, Brosens JJ, Benagiano G. The eutopic endometrium in \nendometriosis: are the changes of clinical significance? Reprod BioMed \nOnline. 2012;24(5):496–502.\n 7. Bulun SE. Endometriosis. N Engl J Med. 2009;360(3):268–79.\n 8. Vallve-Juanico J, Houshdaran S, Giudice LC. The endometrial immune \nenvironment of women with endometriosis. Hum Reprod Update. \n2019;25(5):564–91.\n 9. Liu H, Lang JH. Is abnormal eutopic endometrium the cause of \nendometriosis? The role of eutopic endometrium in pathogenesis of \nendometriosis. Med Sci Monit. 2011;17(4):RA92–9.\n 10. van der Molen RG, Schutten JH, van Cranenbroek B, ter Meer M, Donckers J, \nScholten RR, et al. Menstrual blood closely resembles the uterine immune \nmicro-environment and is clearly distinct from peripheral blood. Hum \nReprod. 2014;29(2):303–14.\n 11. Warren LA, Shih A, Renteira SM, Seckin T, Blau B, Simpfendorfer K, et al. \nAnalysis of menstrual effluent: diagnostic potential for endometriosis. \nMol Med. 2018;24(1):1.\n 12. Schmitz T, Hoffmann V, Olliges E, Bobinger A, Popovici R, Nossner E, \net al. Reduced frequency of perforin-positive CD8+ T cells in menstrual \neffluent of endometriosis patients. J Reprod Immunol. 2021;148:103424.\n 13. Hosseini S, Shokri F, Tokhmechy R, Savadi-Shiraz E, Jeddi-Tehrani M, \nRahbari M, et al. Menstrual blood contains immune cells with inflam-\nmatory and anti-inflammatory properties. J Obstet Gynaecol Res. \n2015;41(11):1803–12.\n 14. Sabbaj S, Hel Z, Richter HE, Mestecky J, Goepfert PA. Menstrual blood \nas a potential source of endometrial derived CD3+ T cells. PLoS One. \n2011;6(12):e28894.\n 15. Nayyar A, Saleem MI, Yilmaz M, DeFranco M, Klein G, Elmaliki KM, et al. \nMenstrual Effluent Provides a Novel Diagnostic Window on the Patho-\ngenesis of Endometriosis. Front Reprod Health. 2020;2(3):1–14.\n 16. McCluggage WG, Sumathi VP , Maxwell P . CD10 is a sensitive and diagnosti-\ncally useful immunohistochemical marker of normal endometrial stroma \nand of endometrial stromal neoplasms. Histopathology. 2001;39(3):273–8.\n 17. Potlog-Nahari C, Feldman AL, Stratton P , Koziol DE, Segars J, Merino MJ, \net al. CD10 immunohistochemical staining enhances the histological \ndetection of endometriosis. Fertil Steril. 2004;82(1):86–92.\n 18. Chen J, Cheung F, Shi R, Zhou H, Lu W, Consortium CHI. PBMC fixation \nand processing for Chromium single-cell RNA sequencing. J Transl Med. \n2018;16(1):198.\n\nPage 14 of 16Shih et al. BMC Medicine          (2022) 20:315 \n 19. Kang HM, Subramaniam M, Targ S, Nguyen M, Maliskova L, McCarthy \nE, et al. Multiplexed droplet single-cell RNA-sequencing using natural \ngenetic variation. Nat Biotechnol. 2017;36(1):89–94.\n 20. Hao Y, Hao S, Andersen-Nissen E, Mauck WM 3rd, Zheng S, Butler \nA, et al. Integrated analysis of multimodal single-cell data. Cell. \n2021;184(13):3573–87 e29.\n 21. Hafemeister C, Satija R. Normalization and variance stabilization of \nsingle-cell RNA-seq data using regularized negative binomial regres-\nsion. Genome Biol. 2019;20(1):296.\n 22. Korsunsky I, Millard N, Fan J, Slowikowski K, Zhang F, Wei K, et al. Fast, \nsensitive and accurate integration of single-cell data with Harmony. Nat \nMethods. 2019;16(12):1289–96.\n 23. Wang F, Qualls AE, Marques-Fernandez L, Colucci F. Biology and pathol-\nogy of the uterine microenvironment and its natural killer cells. Cell Mol \nImmunol. 2021;18(9):2101–13.\n 24. Queckborner S, von Grothusen C, Boggavarapu NR, Francis RM, Davies \nLC, Gemzell-Danielsson K. Stromal Heterogeneity in the Human Prolif-\nerative Endometrium-A Single-Cell RNA Sequencing Study. J Pers Med. \n2021;11(6):448.\n 25. Andreatta M, Corria-Osorio J, Muller S, Cubas R, Coukos G, Carmona SJ. \nInterpretation of T cell states from single-cell transcriptomics data using \nreference atlases. Nat Commun. 2021;12(1):2965.\n 26. Vento-Tormo R, Efremova M, Botting RA, Turco MY, Vento-Tormo M, \nMeyer KB, et al. Single-cell reconstruction of the early maternal-fetal \ninterface in humans. Nature. 2018;563(7731):347–53.\n 27. Dinh HQ, Lin X, Abbasi F, Nameki R, Haro M, Olingy CE, et al. Single-cell \ntranscriptomics identifies gene expression networks driving dif-\nferentiation and tumorigenesis in the human fallopian tube. Cell Rep. \n2021;35(2):108978.\n 28. Lee RD, Munro SA, Knutson TP , LaRue RS, Heltemes-Harris LM, Farrar \nMA. Single-cell analysis identifies dynamic gene expression networks \nthat govern B cell development and transformation. Nat Commun. \n2021;12(1):6843.\n 29. Collin M, McGovern N, Haniffa M. Human dendritic cell subsets. Immu-\nnology. 2013;140(1):22–30.\n 30. Chambers SEJ, Pathak V, Pedrini E, Soret L, Gendron N, Guerin CL, et al. \nCurrent concepts on endothelial stem cells definition, location, and \nmarkers. Stem Cells Transl Med. 2021;10(Suppl 2):S54–61.\n 31. Garcia-Alonso L, Handfield LF, Roberts K, Nikolakopoulou K, Fernando \nRC, Gardner L, et al. Mapping the temporal and spatial dynam-\nics of the human endometrium in vivo and in vitro. Nat Genet. \n2021;53(12):1698–711.\n 32. Barragan F, Irwin JC, Balayan S, Erikson DW, Chen JC, Houshdaran S, \net al. Human endometrial fibroblasts derived from mesenchymal \nprogenitors inherit progesterone resistance and acquire an inflamma-\ntory phenotype in the endometrial niche in endometriosis. Biol Reprod. \n2016;94(5):118.\n 33. Satterfield MC, Hayashi K, Song G, Black SG, Bazer FW, Spencer TE. Pro-\ngesterone regulates FGF10, MET, IGFBP1, and IGFBP3 in the endome-\ntrium of the ovine uterus. Biol Reprod. 2008;79(6):1226–36.\n 34. Young CH, Snow B, DeVore SB, Mohandass A, Nemmara VV, Thompson \nPR, et al. Progesterone stimulates histone citrullination to increase \nIGFBP1 expression in uterine cells. Reproduction. 2021;162(2):117–27.\n 35. Ujvari D, Jakson I, Babayeva S, Salamon D, Rethi B, Gidlof S, et al. \nDysregulation of in vitro decidualization of human endometrial stromal \ncells by insulin via transcriptional inhibition of forkhead box protein O1. \nPLoS One. 2017;12(1):e0171004.\n 36. Fei W, Kijima D, Hashimoto M, Hashimura M, Oguri Y, Kajita S, et al. A \nfunctional role of LEFTY during progesterone therapy for endometrial \ncarcinoma. Cell Commun Signal. 2017;15(1):56.\n 37. Takano M, Lu Z, Goto T, Fusi L, Higham J, Francis J, et al. Transcriptional \ncross talk between the forkhead transcription factor forkhead box O1A \nand the progesterone receptor coordinates cell cycle regulation and \ndifferentiation in human endometrial stromal cells. Mol Endocrinol. \n2007;21(10):2334–49.\n 38. Ono YJ, Terai Y, Tanabe A, Hayashi A, Hayashi M, Yamashita Y, et al. \nDecorin induced by progesterone plays a crucial role in suppressing \nendometriosis. J Endocrinol. 2014;223(2):203–16.\n 39. Halari CD, Nandi P , Jeyarajah MJ, Renaud SJ, Lala PK. Decorin production \nby the human decidua: role in decidual cell maturation. Mol Hum \nReprod. 2020;26(10):784–96.\n 40. Tamm-Rosenstein K, Simm J, Suhorutshenko M, Salumets A, Metsis \nM. Changes in the transcriptome of the human endometrial Ishikawa \ncancer cell line induced by estrogen, progesterone, tamoxifen, and \nmifepristone (RU486) as detected by RNA-sequencing. PLoS One. \n2013;8(7):e68907.\n 41. Salgado RM, FR, Zorn TMT. Modulation of small leucine-rich pro -\nteoglycans (SLRPs) expression in the mouse uterus by estradiol and \nprogesterone. Reprod Biol Endocrinol. 2011;9(9):22.\n 42. Lucariello A, TE, Boccia O, Perna A, Sellitto C, Castald MA, et al. Small \nleucine rich proteoglycans are differently distributed in normal and \npathological endometrium. In Vivo. 2015;29:217–22.\n 43. Martin SS, MS-SM, Ferreira S, de Oliveira F, Aplin FJD, Abrahamsohn P , \net al. Small leucine-rich proteoglycans (SLRPs) in uterine tissues dur -\ning pregnancy in mice. Reproduction. 2003;125:585–95.\n 44. Do Carmo S, Seguin D, Milne R, Rassart E. Modulation of apolipo -\nprotein D and apolipoprotein E mRNA expression by growth arrest \nand identification of key elements in the promoter. J Biol Chem. \n2002;277(7):5514–23.\n 45. Kao LC, TS, Lobo S, Imani B, Yang JP , Gemeyer A, et al. Global gene \nprofiling in human endometrium during the window of implanta-\ntion. Endocrinology. 2002;143:2119–38.\n 46. Altmae S, Koel M, Vosa U, Adler P , Suhorutsenko M, Laisk-Podar T, \net al. Meta-signature of human endometrial receptivity: a meta-\nanalysis and validation study of transcriptomic biomarkers. Sci Rep. \n2017;7(1):10077.\n 47. Omar M, Laknaur A, Al-Hendy A, Yang Q. Myometrial progesterone \nhyper-responsiveness associated with increased risk of human uter -\nine fibroids. BMC Womens Health. 2019;19(1):92.\n 48. Rytkonen KT, Erkenbrack EM, Poutanen M, Elo LL, Pavlicev M, Wagner \nGP . Decidualization of human endometrial stromal fibroblasts is a \nmultiphasic process involving distinct transcriptional programs. \nReprod Sci. 2019;26(3):323–36.\n 49. Giudice LC, Milkowski DA, Lamson G, Rosenfeld RG, Irwin JC. Insulin-\nlike growth factor binding proteins in human endometrium: steroid-\ndependent messenger ribonucleic acid expression and protein \nsynthesis. J Clin Endocrinol Metab. 1991;72(4):779–87.\n 50. Tarantino S, Verhage HG, Fazleabas AT. Regulation of insulin-like \ngrowth factor-binding proteins in the baboon (Papio anubis) uterus \nduring early pregnancy. Endocrinology. 1992;130(4):2354–62.\n 51. Jasienska G, Ellison PT, Galbarczyk A, Jasienski M, Kalemba-Drozdz \nM, Kapiszewska M, et al. Apolipoprotein E (ApoE) polymorphism \nis related to differences in potential fertility in women: a case of \nantagonistic pleiotropy? Proc Biol Sci. 2015;282(1803):20142395.\n 52. Garcia AJ, Tom C, Guemes M, Polanco G, Mayorga ME, Wend K, \net al. ERalpha signaling regulates MMP3 expression to induce \nFasL cleavage and osteoclast apoptosis. J Bone Miner Res. \n2013;28(2):283–90.\n 53. Keller NR, S-RE, Eisenberg E, Osteen KG. Progesterone exposure pre -\nvents matrix metalloproteinase-3 (MMP-3) stimulation by interleukin-\n1a in human endometrial stromal cells. J Clin Endocrinol Metab. \n2000;85:11–1619.\n 54. Yamashita CM, Dolgonos L, Zemans RL, Young SK, Robertson J, Bri-\nones N, et al. Matrix metalloproteinase 3 is a mediator of pulmonary \nfibrosis. Am J Pathol. 2011;179(4):1733–45.\n 55. Luddi A, Marrocco C, Governini L, Semplici B, Pavone V, Luisi S, et al. \nExpression of Matrix Metalloproteinases and Their Inhibitors in \nEndometrium: High Levels in Endometriotic Lesions. Int J Mol Sci. \n2020;21(8):2840.\n 56. Chen C, Li C, Liu W, Guo F, Kou X, Sun S, et al. Estrogen-induced \nFOS-like 1 regulates matrix metalloproteinase expression and the \nmotility of human endometrial and decidual stromal cells. J Biol Chem. \n2020;295(8):2248–58.\n 57. Lockwood CJ, KG, Hausknecht VA, Papp C, Schatz F. Matrix metal-\nloproteinase and matrix metalloproteinase inhibitor expression in \nendometrial stromal cells during progestin-initiated decidualization \nand menstruation-related progestin withdrawal. Endocrinology. \n1998;139:4607–13.\n 58. Singer CF, Marbaix E, Kokorine I, Lemoine P , Donnez J, Eeckhout Y, et al. \nParacrine stimulation of interstitial collagenase (MMP-1) in the human \nendometrium by interleukin 1alpha and its dual block by ovarian \nsteroids. Proc Natl Acad Sci U S A. 1997;94(19):10341–5.\n\nPage 15 of 16\nShih et al. BMC Medicine          (2022) 20:315 \n \n 59. Ghosh K, Capell BC. The senescence-associated secretory pheno-\ntype: critical effector in skin cancer and aging. J Invest Dermatol. \n2016;136(11):2133–9.\n 60. Basisty N, Kale A, Jeon OH, Kuehnemann C, Payne T, Rao C, et al. A pro-\nteomic atlas of senescence-associated secretomes for aging biomarker \ndevelopment. PLoS Biol. 2020;18(1):e3000599.\n 61. von Rango U, Alfer J, Kertschanska S, Kemp B, Muller-Newen G, Heinrich \nPC, et al. Interleukin-11 expression: its significance in eutopic and \nectopic human implantation. Mol Hum Reprod. 2004;10(11):783–92.\n 62. Ng B, Cook SA, Schafer S. Interleukin-11 signaling underlies fibrosis, \nparenchymal dysfunction, and chronic inflammation of the airway. Exp \nMol Med. 2020;52(12):1871–8.\n 63. Ng B, Dong J, Viswanathan S, Widjaja AA, Paleja BS, Adami E, et al. \nFibroblast-specific IL11 signaling drives chronic inflammation in murine \nfibrotic lung disease. FASEB J. 2020;34(9):11802–15.\n 64. Chen H, Chen H, Liang J, Gu X, Zhou J, Xie C, et al. TGF-beta1/IL-11/\nMEK/ERK signaling mediates senescence-associated pulmonary fibrosis \nin a stress-induced premature senescence model of Bmi-1 deficiency. \nExp Mol Med. 2020;52(1):130–51.\n 65. Dimitriadis E, Stoikos C, Stafford-Bell M, Clark I, Paiva P , Kovacs G, et al. \nInterleukin-11, IL-11 receptoralpha and leukemia inhibitory factor are \ndysregulated in endometrium of infertile women with endometriosis \nduring the implantation window. J Reprod Immunol. 2006;69(1):53–64.\n 66. Gubbels Bupp MR, Jorgensen TN, Kotzin BL. Identification of candidate \ngenes that influence sex hormone-dependent disease phenotypes in \nmouse lupus. Genes Immun. 2008;9(1):47–56.\n 67. Schroder WA, Le TT, Major L, Street S, Gardner J, Lambley E, et al. A \nphysiological function of inflammation-associated SerpinB2 is regula-\ntion of adaptive immunity. J Immunol. 2010;184(5):2663–70.\n 68. Hsieh HH, Chen YC, Jhan JR, Lin JJ. The serine protease inhibitor serpinB2 binds \nand stabilizes p21 in senescent cells. J Cell Sci. 2017;130(19):3272–81.\n 69. Park SR, Lee JW, Kim SK, Yu WJ, Lee SJ, Kim D, et al. The impact of fine \nparticulate matter (PM) on various beneficial functions of human endo-\nmetrial stem cells through its key regulator SERPINB2. Exp Mol Med. \n2021;53(12):1850–65.\n 70. Zhang X, Christenson LK, Nothnick WB. Regulation of MMP-9 expres-\nsion and activity in the mouse uterus by estrogen. Mol Reprod Dev. \n2007;74(3):321–31.\n 71. Ahmad N, Chen S, Wang W, Kapila S. 17beta-estradiol Induces MMP-9 \nand MMP-13 in TMJ Fibrochondrocytes via Estrogen Receptor alpha. J \nDent Res. 2018;97(9):1023–30.\n 72. Marbaix EDJ, Courtoy PJ, Eeckhout Y. Progesterone regulates the activity \nof collagenase and related gelatinases A and B in human endometrial \nexplants. Proc Nati Acad Sci USA. 1992;89:11789–93.\n 73. Steenport M, Khan KM, Du B, Barnhard SE, Dannenberg AJ, Falcone DJ. \nMatrix metalloproteinase (MMP)-1 and MMP-3 induce macrophage \nMMP-9: evidence for the role of TNF-alpha and cyclooxygenase-2. J \nImmunol. 2009;183(12):8119–27.\n 74. Su L, Dong Y, Wang Y, Wang Y, Guan B, Lu Y, et al. Potential role of senes-\ncent macrophages in radiation-induced pulmonary fibrosis. Cell Death \nDis. 2021;12(6):527.\n 75. Hong EJ, Park SH, Choi KC, Leung PC, Jeung EB. Identification of estrogen-regu-\nlated genes by microarray analysis of the uterus of immature rats exposed to \nendocrine disrupting chemicals. Reprod Biol Endocrinol. 2006;4:49.\n 76. Ghezzo F, Berta GN, Beccaro M, D’Avolio A, Racca S, Conti G, et al. Cal-\ncyclin gene expression modulation by medroxyprogesterone acetate. \nBiochem Pharmacol. 1997;54(2):299–305.\n 77. Xia C, Braunstein Z, Toomey AC, Zhong J, Rao X. S100 proteins as an \nimportant regulator of macrophage inflammation. Front Immunol. \n2017;8:1908.\n 78. Landi C, Bargagli E, Carleo A, Refini RM, Bennett D, Bianchi L, et al. Bron-\nchoalveolar lavage proteomic analysis in pulmonary fibrosis associated \nwith systemic sclerosis: S100A6 and 14-3-3epsilon as potential biomark-\ners. Rheumatology (Oxford). 2019;58(1):165–78.\n 79. Slomnicki LP , Lesniak W. S100A6 (calcyclin) deficiency induces senes-\ncence-like changes in cell cycle, morphology and functional character-\nistics of mouse NIH 3T3 fibroblasts. J Cell Biochem. 2010;109(3):576–84.\n 80. Haim K, Weitzenfeld P , Meshel T, Ben-Baruch A. Epidermal growth factor \nand estrogen act by independent pathways to additively promote \nthe release of the angiogenic chemokine CXCL8 by breast tumor cells. \nNeoplasia. 2011;13(3):230–43.\n 81. Armstrong GM, Maybin JA, Murray AA, Nicol M, Walker C, Saunders PTK, \net al. Endometrial apoptosis and neutrophil infiltration during men-\nstruation exhibits spatial and temporal dynamics that are recapitulated \nin a mouse model. Sci Rep. 2017;7(1):17416.\n 82. Russo RC, Garcia CC, Teixeira MM, Amaral FA. The CXCL8/IL-8 \nchemokine family and its receptors in inflammatory diseases. Expert \nRev Clin Immunol. 2014;10(5):593–619.\n 83. Konno R, Y-OH, Fujiwara H, Uchide I, Shibahara H, Okwada M, et al. \nRole of immunoreactions and mast cells in pathogenesis of human \nendometriosis -morphologic study and gene expression analysis. \nHum Cell. 2003;16(3):141–9.\n 84. Luckow Invitti A, Schor E, Martins Parreira R, Kopelman A, Kamer -\ngorodsky G, Goncalves GA, et al. Inflammatory cytokine profile of \ncocultivated primary cells from the endometrium of women with \nand without endometriosis. Mol Med Rep. 2018;18(2):1287–96.\n 85. Acosta JC, O’Loghlen A, Banito A, Guijarro MV, Augert A, Raguz S, \net al. Chemokine signaling via the CXCR2 receptor reinforces senes-\ncence. Cell. 2008;133(6):1006–18.\n 86. Carleton JB, Berrett KC, Gertz J. Multiplex enhancer interference \nreveals collaborative control of gene regulation by estrogen receptor \nalpha-bound enhancers. Cell Syst. 2017;5(4):333–44 e5.\n 87. Heckmann BL, Zhang X, Xie X, Liu J. The G0/G1 switch gene 2 \n(G0S2): regulating metabolism and beyond. Biochim Biophys Acta. \n2013;1831(2):276–81.\n 88. Barradas M, Gonos ES, Zebedee Z, Kolettas E, Petropoulou C, Delgado \nMD, et al. Identification of a candidate tumor-suppressor gene \nspecifically activated during Ras-induced senescence. Exp Cell Res. \n2002;273(2):127–37.\n 89. Hu WP , Tay SK, Zhao Y. Endometriosis-specific genes identified by \nreal-time reverse transcription-polymerase chain reaction expression \nprofiling of endometriosis versus autologous uterine endometrium. J \nClin Endocrinol Metab. 2006;91(1):228–38.\n 90. Andrade PM, Silva ID, Borra RC, Lima GR, Baracat EC. Estrogen and \nselective estrogen receptor modulator regulation of insulin-like \ngrowth factor binding protein 5 in the rat uterus. Gynecol Endocrinol. \n2002;16(4):265–70.\n 91. Nguyen X-X, Muhammad L, Nietert PJ, Feghali-Bostwick C. IGFBP-5 \nPromotes Fibrosis via Increasing Its Own Expression and That of \nOther Pro-fibrotic Mediators. Front Endocrinol. 2018;9:eaaf7533.\n 92. Kim KS, Seu YB, Baek SH, Kim MJ, Kim KJ, Kim JH, et al. Induc-\ntion of cellular senescence by insulin-like growth factor binding \nprotein-5 through a p53-dependent mechanism. Mol Biol Cell. \n2007;18(11):4543–52.\n 93. Sanada F, Taniyama Y, Muratsu J, Otsu R, Shimizu H, Rakugi H, et al. \nIGF binding protein-5 induces cell senescence. Front Endocrinol \n(Lausanne). 2018;9:53.\n 94. Salih DAM, TG, Holding C, Szestak TAM, Gonzalez MI, Carter EJ, et al. \nInsulin-like growth factor-binding protein 5 (Igfbp5) compromises \nsurvival, growth, muscle development, and fertility in mice. PNAS. \n2004;101:4314–9.\n 95. Sheikh MS, Shao ZM, Chen JC, Fontana JA. Differential regulation \nof matrix Gla protein (MGP) gene expression by retinoic acid and \nestrogen in human breast carcinoma cells. Mol Cell Endocrinol. \n1993;92(2):153–60.\n 96. Dressman MA, Walz TM, LC, Barnes L, Buchholtz S, Kwon I, et al. Genes \nthat co-cluster with estrogen receptor alpha in microarray analysis of \nbreast biopsies. Pharmacogenomics J. 2001;1:135–41.\n 97. Han L, Li X, Zhang G, Xu Z, Gong D, Lu F, et al. Pericardial interstitial \ncell senescence responsible for pericardial structural remodeling in \nidiopathic and postsurgical constrictive pericarditis. J Thorac Cardio -\nvasc Surg. 2017;154(3):966–75 e4.\n 98. Kumari R, Jat P . Mechanisms of cellular senescence: cell cycle arrest \nand senescence associated secretory phenotype. Front Cell Dev Biol. \n2021;9:645593.\n 99. Stilley JA, Birt JA, Nagel SC, Sutovsky M, Sutovsky P , Sharpe-Timms KL. \nNeutralizing TIMP1 restores fecundity in a rat model of endometriosis \nand treating control rats with TIMP1 causes anomalies in ovarian \nfunction and embryo development. Biol Reprod. 2010;83(2):185–94.\n 100. Wang J, Jarrett J, Huang CC, Satcher RL Jr, Levenson AS. Identification \nof estrogen-responsive genes involved in breast cancer metastases \nto the bone. Clin Exp Metastasis. 2007;24(6):411–22.\n\nPage 16 of 16Shih et al. BMC Medicine          (2022) 20:315 \n•\n \nfast, convenient online submission\n •\n  \nthorough peer review by experienced researchers in your field\n• \n \nrapid publication on acceptance\n• \n \nsupport for research data, including large and complex data types\n•\n  \ngold Open Access which fosters wider collaboration and increased citations \n \nmaximum visibility for your research: over 100M website views per year •\n  At BMC, research is always in progress.\nLearn more biomedcentral.com/submissions\nReady to submit y our researc hReady to submit y our researc h  ?  Choose BMC and benefit fr om: ?  Choose BMC and benefit fr om: \n 101. Kunzmann S, Ottensmeier B, Speer CP , Fehrholz M. Effect of proges-\nterone on Smad signaling and TGF-beta/Smad-regulated genes in \nlung epithelial cells. PLoS One. 2018;13(7):e0200661.\n 102. Thweatt R, Lumpkin CK Jr, Goldstein S. A novel gene encoding a \nsmooth muscle protein is overexpressed in senescent human fibro-\nblasts. Biochem Biophys Res Commun. 1992;187(1):1–7.\n 103. Vafashoar F, Mousavizadeh K, Poormoghim H, Haghighi A, Pashangza-\ndeh S, Mojtabavi N. Progesterone aggravates lung fibrosis in a mouse \nmodel of systemic sclerosis. Front Immunol. 2021;12:742227.\n 104. Schafer MJ, White TA, Iijima K, Haak AJ, Ligresti G, Atkinson EJ, et al. Cel-\nlular senescence mediates fibrotic pulmonary disease. Nat Commun. \n2017;8:14532.\n 105. Kim TH, Yoo JY, Choi KC, Shin JH, Leach RE, Fazleabas AT, et al. Loss of \nHDAC3 results in nonreceptive endometrium and female infertility. Sci \nTransl Med. 2019;11(474).\n 106. DeNardo DG, Kim HT, Hilsenbeck S, Cuba V, Tsimelzon A, Brown PH. \nGlobal gene expression analysis of estrogen receptor transcription fac-\ntor cross talk in breast cancer: identification of estrogen-induced/acti-\nvator protein-1-dependent genes. Mol Endocrinol. 2005;19(2):362–78.\n 107. Cao W, Mah K, Carroll RS, Slayden OD, Brenner RM. Progesterone \nwithdrawal up-regulates fibronectin and integrins during menstrua-\ntion and repair in the rhesus macaque endometrium. Hum Reprod. \n2007;22(12):3223–31.\n 108. Chen G, Liu L, Sun J, Zeng L, Cai H, He Y. Foxf2 and Smad6 co-regulation \nof collagen 5A2 transcription is involved in the pathogenesis of intrau-\nterine adhesion. J Cell Mol Med. 2020;24(5):2802–18.\n 109. Chan JM, Ho SH, Tai IT. Secreted protein acidic and rich in cysteine-\ninduced cellular senescence in colorectal cancers in response to \nirinotecan is mediated by P53. Carcinogenesis. 2010;31(5):812–9.\n 110. Urushiyama H, Terasaki Y, Nagasaka S, Terasaki M, Kunugi S, Nagase \nT, et al. Role of alpha1 and alpha2 chains of type IV collagen in early \nfibrotic lesions of idiopathic interstitial pneumonias and migration of \nlung fibroblasts. Lab Investig. 2015;95(8):872–85.\n 111. Lee Y, Shivashankar GV. Analysis of transcriptional modules during \nhuman fibroblast ageing. Sci Rep. 2020;10(1):19086.\n 112. Teo YV, Rattanavirotkul N, Olova N, Salzano A, Quintanilla A, Tarrats \nN, et al. Notch signaling mediates secondary senescence. Cell Rep. \n2019;27(4):997–1007 e5.\n 113. Matalliotaki C, Matalliotakis M, Rahmioglu N, Mavromatidis G, Matal-\nliotakis I, Koumantakis G, et al. Role of FN1 and GREB1 gene polymor-\nphisms in endometriosis. Mol Med Rep. 2019;20(1):111–6.\n 114. Klemmt PA, Carver JG, Kennedy SH, Koninckx PR, Mardon HJ. Stromal \ncells from endometriotic lesions and endometrium from women with \nendometriosis have reduced decidualization capacity. Fertil Steril. \n2006;85(3):564–72.\n 115. Cicinelli E, Trojano G, Mastromauro M, Vimercati A, Marinaccio M, \nMitola PC, et al. Higher prevalence of chronic endometritis in women \nwith endometriosis: a possible etiopathogenetic link. Fertil Steril. \n2017;108(2):289–95 e1.\n 116. Sapkota Y, Steinthorsdottir V, Morris AP , Fassbender A, Rahmioglu N, \nDe Vivo I, et al. Meta-analysis identifies five novel loci associated with \nendometriosis highlighting key genes involved in hormone metabo-\nlism. Nat Commun. 2017;8:15539.\n 117. Tomari H, Kawamura T, Asanoma K, Egashira K, Kawamura K, Honjo K, \net al. Contribution of senescence in human endometrial stromal cells \nduring proliferative phase to embryo receptivitydagger. Biol Reprod. \n2020;103(1):104–13.\n 118. Lin X, Dai Y, Tong X, Xu W, Huang Q, Jin X, et al. Excessive oxidative stress \nin cumulus granulosa cells induced cell senescence contributes to \nendometriosis-associated infertility. Redox Biol. 2020;30:101431.\n 119. Yu CX, Song JH, Li YF, Tuo Y, Zheng JJ, Miao RJ, et al. Correlation \nbetween replicative senescence of endometrial gland epithelial cells \nin shedding and non-shedding endometria and endometriosis cyst \nduring menstruation. Gynecol Endocrinol. 2018;34(11):981–6.\n 120. Takebayashi A, Kimura F, Kishi Y, Ishida M, Takahashi A, Yamanaka A, et al. \nThe association between endometriosis and chronic endometritis. PLoS \nOne. 2014;9(2):e88354.\n 121. Wu D, Kimura F, Zheng L, Ishida M, Niwa Y, Hirata K, et al. Chronic endo-\nmetritis modifies decidualization in human endometrial stromal cells. \nReprod Biol Endocrinol. 2017;15(1):16.\n 122. Miyagaki T, Fujimoto M, Sato S. Regulatory B cells in human inflamma-\ntory and autoimmune diseases: from mouse models to clinical research. \nInt Immunol. 2015;27(10):495–504.\n 123. Sojka DK, Yang L, Yokoyama WM. Uterine natural killer cells. Front \nImmunol. 2019;10:960.\n 124. Zhang Y, Wang Y, Wang XH, Zhou WJ, Jin LP , Li MQ. Crosstalk between \nhuman endometrial stromal cells and decidual NK cells pro-\nmotes decidualization in vitro by upregulating IL25. Mol Med Rep. \n2018;17(2):2869–78.\n 125. Bany BM, Scott CA, Eckstrum KS. Analysis of uterine gene expression \nin interleukin-15 knockout mice reveals uterine natural killer cells do \nnot play a major role in decidualization and associated angiogenesis. \nReproduction. 2012;143(3):359–75.\n 126. Ashkar AA, Black GP , Wei Q, He H, Liang L, Head JR, et al. Assess-\nment of requirements for IL-15 and IFN regulatory factors in uterine \nNK cell differentiation and function during pregnancy. J Immunol. \n2003;171(6):2937–44.\n 127. Jorgensen H, Fedorcsak P , Isaacson K, Tevonian E, Xiao A, Beste M, et al. \nEndometrial cytokines in patients with and without endometriosis \nevaluated for infertility. Fertil Steril. 2022;117(3):629–40.\n 128. Seshadri S, Sunkara SK. Natural killer cells in female infertility and recur-\nrent miscarriage: a systematic review and meta-analysis. Hum Reprod \nUpdate. 2014;20(3):429–38.\n 129. As-Sanie S, Black R, Giudice LC, Gray Valbrun T, Gupta J, Jones B, et al. \nAssessing research gaps and unmet needs in endometriosis. Am J \nObstet Gynecol. 2019;221(2):86–94.\n 130. Kirkland JL, Tchkonia T. Senolytic drugs: from discovery to translation. J \nIntern Med. 2020;288(5):518–36.\n 131. Wissler Gerdes EO, Zhu Y, Tchkonia T, Kirkland JL. Discovery, develop-\nment, and future application of senolytics: theories and predictions. \nFEBS J. 2020;287(12):2418–27.\nPublisher’s Note\nSpringer Nature remains neutral with regard to jurisdictional claims in pub-\nlished maps and institutional affiliations.","source_license":"CC0","license_restricted":false}