The Dysregulated IL-23/T_H17 Axis in Endometriosis Pathophysiology

Endometriosis is characterized by the growth of endometrial-like tissue in ectopic sites and is commonly associated with dysmenorrhea, dyspareunia, chronic pelvic pain, and infertility (1, 2). Although endometriosis incidence and prevalence are difficult to establish due to the invasive nature of disease diagnosis, it is estimated that endometriosis afflicts 10% (190 million) of reproductive-aged women and an unknown percentage of gender-diverse individuals (3–5). Although the origin and pathogenesis of endometriosis remain largely unknown, we and others postulate that immune dysfunction promotes ectopic lesion growth and survival by impeding lesion clearance (1, 6). Specifically, proinflammatory IL-17 is well -established to contribute to the pathogenesis of several inflammatory diseases, including endometriosis (7–10). Indeed, we have demonstrated that endometriotic lesions produce IL-17 and that IL-17 is significantly increased in patient plasma compared with controls (10). Additionally, surgical removal of lesions significantly reduces IL-17 plasma concentration, suggesting that endometriotic lesions are likely a predominant source of circulating IL-17 (10). IL-17 is largely produced by T_H17 cells, which also coproduce IL-21 and IL-22, contributing to tissue inflammation and angiogenesis (10, 11). T_H17 cells act pathogenically in various inflammatory and autoimmune disorders, including psoriasis and rheumatoid arthritis, by secreting proinflammatory cytokines, which promote immune cell accumulation at inflammatory sites (11, 12). However, T_H17 cells are highly plastic, adopting differing fates in response to environmental cues, such as cytokines/chemokines (8, 13, 14). In fact, T_H17 cells may also play a beneficial (nonpathogenic) role against opportunistic pathogens and promote and/or build intestinal barrier integrity (11, 14–16). Specifically, IL-17A⁺IL-10⁻ T_H17 cells are termed “pathogenic” (pT_H17), whereas IL-17A⁺IL-10⁺ are termed “nonpathogenic” (npT_H17), promoting an anti-inflammatory/immunosuppressive phenotype (12, 17).

Studies show T_H17 differentiation is initiated from naive CD4⁺ T cells by a combination of cytokines TGF-β1, IL-1β, IL-6, IL-21, and IL-23 (11, 18, 19). IL-17 and IL-21 secreted by developing T_H17 cells further drive T_H17 cell fate and early proliferation, whereas IL-23 is fundamental for T_H17 cell expansion and phenotypic stabilization (11, 18–21). When exposed to IL-23, developing T_H17 cells produce significantly reduced anti-inflammatory IL-10, further evoking a proinflammatory, pathogenic phenotype (17, 18). In fact, T_H17 cells are incapable of inducing autoimmune disease (18) and unable to upregulate proinflammatory cytokines without exposure to IL-23 (17). Hence, complete acquisition of pathogenic potential in T_H17 cells is likely mediated by IL-23. This is supported by reports that IL-23–specific p19 subunit neutralization ameliorates/protects from autoimmune disease, including experimental autoimmune encephalomyelitis and autoimmune colitis, whereas loss of downstream IL-17 or IL-17 receptor A (IL-17RA) failed to reduce disease emergence/severity (11, 22). Thus, if IL-23 is a key contributor in driving T_H17 cells to a pathogenic profile, IL-23 may govern the balance between T_H17 and T-regulatory cells (T_REGs), which is critical to avoid immune dysfunction and maintain immune homeostasis (23–25).

Some theorize that in endometriosis, a tolerant, anti-inflammatory peritoneal microenvironment initially allows for early lesion development/establishment, preventing lesion clearance (26), although later lesion growth and maintenance are supported by a proinflammatory microenvironment, promoting inflammation, angiogenesis, and fibrosis (26). This scenario suggests a skewed T_H17/T_REG axis toward a tolerant (T_REG-dominant) microenvironment as lesions initially establish, followed by a proinflammatory (T_H17-dominant) microenvironment allowing for lesion survival, because downstream T_H17 cytokines, such as IL-17, are well known to promote lesion angiogenesis, proliferation, and inflammation (10, 27, 28). Results from our laboratory support this theory, finding decreased T_H17 cells in patient menstrual effluent compared with controls (29). Based on Sampson’s theory of retrograde menstruation, these data provide insights into early events in endometriosis lesion establishment, because decreased T_H17 cells in patient menstrual effluent may influence the local immune microenvironment and facilitate initial lesion establishment following retrograde menstruation. Although further studies are warranted, this theory is also supported by findings that T_H17 cells were significantly increased in the peritoneal fluid (PF) of patients with severe (III/IV stage) endometriosis compared with early (I/II stage) endometriosis (13).

IL-23, a heterodimeric cytokine largely secreted by activated dendritic cells and macrophages (30), has been established to dysregulate the T_H17/IL-17 axis in a variety of inflammatory and autoimmune disorders, including psoriasis and rheumatoid arthritis (19, 31, 32). Because evidence suggests dysregulated levels of T_H17 cells (13, 24, 33) and IL-17 (10, 28) in endometriosis, as well as an association of T_H17 cells with increasing disease severity (13), IL-23 is likely regulating the T_H17/IL-17 axis in endometriosis to exacerbate disease. Although downstream IL-17 has been previously evaluated for its therapeutic potential and role in various autoimmune and chronic inflammatory diseases (32), the role of upstream IL-23 has yet to be fully evaluated in endometriosis pathophysiology. Unlike targeting IL-17, targeting IL-23 reduces proinflammatory IL-17 production by pT_H17 cells, while preserving overall IL-17 production from other IL-17–producing cells, ameliorating safety concerns around eliminating protective effects of IL-17 (32). This study depicts a dysregulated IL-23/T_H17 axis in endometriosis patient samples, which was replicated in vitro. We also provide insights using an endometriosis mouse model, revealing IL-23 to significantly alter the local immune microenvironment and T_H17/T_REG axis, indicating a plausible role of IL-23 in endometriosis pathophysiology and highlighting IL-23 as a potential nonhormonal therapeutic target for endometriosis patients.

The ethics for the use of human samples in this study was approved by the Health Sciences Research Ethics Board at Kingston Health Sciences Center, Queen’s University (Kingston, ON, Canada), the Greenville Health System (Greenville, SC, USA), the University of North Carolina at Chapel Hill (Chapel Hill, NC, USA), and Wake Forest Baptist Health (Winston-Salem, NC, USA). Written, informed consent was attained from endometriosis patients and healthy controls prior to the collection and use of clinical of samples. Animal studies were approved by the Queen’s University Animal Care Committee at Queen’s University (Kingston, ON, Canada).

Endometriosis patients undergoing excision surgery due to endometriosis-associated infertility and/or pelvic pain were included in this study (Table I). Healthy, fertile controls without endometriosis were recruited for endometrial biopsy and plasma collection prior to elective tubal ligations. All patients and controls recruited for this study were free of hormonal therapy for a minimum of 3 mo prior to sample collection.

Total RNA was extracted from ectopic (n = 9) and eutopic (n = 9) endometrium of patients, and normal healthy endometrium (n = 9) using a total RNA purification kit (17200, Norgen Biotek Corp., Canada) as per the manufacturer’s protocol. RNA quality was assessed and cDNA synthesized as previously detailed (34). RT-qPCR was conducted with custom array plates (330171, Qiagen, Canada) using the LightCycler 480 real-time PCR system (Roche Molecular Systems Inc., Switzerland) with RT² SYBR Green qPCR Mastermix (330501, Qiagen, Canada) and 200 ng cDNA, as per the manufacturer’s protocol. Data analysis was conducted using the ΔΔCt method, and relative quantification was performed by geometric averaging of Gapdh and Actb reference genes. The custom RT-qPCR array was designed to target 11 genes of interest in the IL-23/T_H17/IL-17 axis: IL23A, IL12B, IL12A, IL23R, IL21, IL22, IL17A, IL17F, IL17RA, IL6, and TGFB1.

IL-23 concentration was measured in endometriosis patient tissue protein extracts (ectopic [n = 10] and eutopic [n = 16] endometrium) and patient plasma (n = 13) as compared with healthy controls (n = 19), as well as between stratified patient plasma samples (n = 7) over three time points using an IL-23 human ELISA kit (ab64708, Abcam, UK), as per the manufacturer’s protocol. A total of 20 mg of tissues were homogenized in PowerBead tubes (13113-50, Qiagen, Canada) containing protease inhibitor mixture (535140, Sigma-Aldrich, Canada) and tissue protein extraction reagent (78510, ThermoFisher, Canada). Protein concentration was determined using a Pierce BCA protein assay kit (23225, ThermoFisher) as per kit protocol. The samples were normalized to the lowest concentration and stored at −80°C prior to analysis. ELISA results were read in a SpectraMax iD5 microplate reader (Molecular Devices, USA) at an absorbance of 450 nm and a reference wavelength of 620 nm.

An endometrioma tissue microarray was constructed using patient tissue samples from Kingston General Hospital (Kingston, ON, Canada), as previously described (35). Matched patient ectopic and eutopic tissues (n = 17), as well as normal endometrium controls (n = 10), were subjected to immunohistochemistry (IHC) staining using a polyclonal IL-23 Ab (1:200; ab45420, Abcam, UK), and analysis using HALO image analysis software (Indica Labs, USA) was conducted as previously outlined (10, 34). Briefly, Ag retrieval was conducted using citrate buffer for 30 min prior to staining with primary anti–IL-23 Ab (30 min). Following this, samples were incubated with a biotinylated secondary polyclonal goat anti-rabbit IgG Ab (1:500; DakoCytomation, Denmark), stained with diaminobenzidine chromogen (DakoCytomation), and counterstained with Harris hematoxylin (Fisher Scientific, Canada) (10).

A total of 30 mL of whole blood was collected from healthy volunteers into BD Vacutainer tubes (CABDL366643L, VWR, USA). PBMCs were isolated using Lymphoprep (07811, StemCell Technologies, Canada) and SepMate-50 PBMC isolation tubes (85460, StemCell Technologies, Canada). Primary CD4⁺ T cells were isolated from PBMCs using an EasySep Human CD4⁺ T cell isolation kit (17952, StemCell Technologies, Canada) as per the manufacturer’s instructions. Primary CD4⁺ T cells were seeded at 1 × 10⁶ cells/well in 24-well plates and maintained in X-VIVO-15 medium (CA12001-988, VWR). Four treatment groups were used: 1) mixture and activation, 2) activation alone, 3) mixture alone, and 4) control medium. Activation consisted of 10 µg/ml plate-bound anti-human CD3 mAb (300437) and 1 µg/ml soluble anti-human CD28 mAb (302902) purchased from BioLegend (USA). Based on previously published protocols (36–39), the T_H17 mixture consisted of 5 ng/ml recombinant human TGF-β1 (rhTGF-β1; 240-B-002), 12.5 ng/ml rhIL-1β (201-LB-005), 25 ng/ml rhIL-6 (206-IL-010), 25 ng/ml rhIL-21 (8879-IL-010), 25 ng/ml rhIL-23 (1290-IL-010), 500 ng/ml anti–IFN-γ (506531, BioLegend), and 500 ng/ml anti-IL-4 (500837, BioLegend) purchased from R&D Systems unless stated otherwise. Primary CD4⁺ T cells were treated in triplicate and incubated at 37°C with 5% CO₂ for 4 d, as per established protocols (36, 37). On day 4, the cells were restimulated with 50 ng/ml of PMA (74042, StemCell Technologies, Canada) and 250 ng/ml of ionomycin (73722, StemCell Technologies, Canada) in combination with 2 µl/ml of protein transport inhibitor mixture (brefeldin A and monensin; 00-4980-93, ThermoFisher) and incubated at 37°C with 5% CO₂ for 5 h prior to analysis. Supernatant was analyzed using a commercially available multiplex cytokine/chemokine analysis (HD48-plex, EveTechnologies, Canada). To examine cell fate following in vitro polarization, flow cytometry was conducted on day 4.

An immortalized human endometriotic epithelial cell line derived from epithelial cells of peritoneal endometriosis (12Z cells; kindly gifted by Dr. Anna Starzinski-Powitz), endometrial epithelial carcinoma cells (EECCs; CRL-1671, ATCC), HUVECs (CRL-1730, ATCC), and human endometrial stromal cells (hESCs; T0533 ABM, Canada) were used in this study. EECCs were maintained in DMEM (D6429, Sigma-Aldrich, Canada) supplemented with 10% FBS (97068-085, VWR) and 1% penicillin and streptomycin (15140122, ThermoFisher). Cell line cultivation and 12Z, HUVEC, and hESC maintenance were conducted as previously described (34).

12Z cells, EECCs, HUVECs, and hESCs were seeded at 3 × 10⁵ cells/well onto 6-well plates and rested for 24 h prior to stimulation with PBS or rhIL-23 (1290-IL-010, R&D Systems, USA) at concentrations of 1, 10, 50, and 100 ng/ml. The cells were treated in triplicate and incubated at 37°C with 5% CO₂ for 24 h, and supernatant was analyzed via multiplex cytokine array (HD48-plex, EveTechnologies, Canada).

12Z cells, EECCs, HUVECs, and hESCs were seeded onto 96-well plates at 5 × 10³ cells/well and rested for 24 h. The cells were then treated in triplicate for 24 h with varying concentrations of rhIL-23 for analysis via a WST-1 proliferation assay using WST-1 reagent (0501594400, Sigma-Aldrich, Canada) and an apoptosis assay using Caspase-Glo 3/7 reagent (G8091, Promega, Canada), as per the manufacturer’s protocol. The results of WST-1 and Caspase-Glo assays were acquired as previously outlined (34).

To determine the influence of IL-23 on angiogenesis, a tubulogenesis assay was conducted using IBIDI µ-plates (81506, IBIDI, Germany), as per the manufacturer’s protocol. Briefly, each µ-plate well was coated with 10 µl of Matrigel (353230, Corning, USA) and incubated for 1 h in a humid chamber at 37°C with 5% CO₂. HUVECs were then seeded at 1 × 10⁴ cells/well in triplicate on top of polymerized Matrigel in 50 µl of medium containing either VEGF (25 ng/ml), PBS (vehicle), or rhIL-23 (1, 10, 50, or 100 ng/ml). The cells were incubated for 4 h at 37°C with 5% CO₂ prior to image acquisition. The images were captured at 4× objective using an Olympus CKX41 microscope (Olympus Life Science, USA) with a Lumenera Infinity 1-3 microscope camera. The images were analyzed for tube formation using WIMASIS software (WimTube: tube formation assay image analysis solution) (40).

Spleens from female C57BL/6 mice (n = 6) were collected and mechanically digested through 70-µm strainers before centrifugation at 400g for 5 min at 4°C, and the pellets were resuspended and counted. An EasySep isolation kit (19765, StemCell Technologies, Canada) was used to isolate naive CD4⁺ T cells, following the manufacturer’s protocol. CD4⁺ T cells were washed with 2% FACS buffer (PBS with 2% FBS) prior to resuspension in TexMACS complete medium (130-097-196, Miltenyi Biotec, USA), supplemented with 10% FBS, 1% penicillin and streptomycin, and 0.1 mM 2-ME (31350-010, ThermoFisher). The cells were counted and seeded at 2.5 × 10⁵ cells/well in 96-well plates. Four treatment groups were used: 1) mixture and activation, 2) mixture and activation supplemented with an additional 10 ng/ml recombinant mouse (rm)IL-23, 3) 10 ng/ml rmIL-23 (no mixture) and activation, and 4) medium and activation (control). Activation consisted of 5 µg/ml plate-bound anti-mouse CD3ε mAb (100340) and 5 µg/ml soluble anti-mouse CD28 mAb (102116) purchased from BioLegend. A commercially available CytoBox T_H17 mouse mixture (130-107-758, Miltenyi Biotec) was used for polarization of murine CD4⁺ T cells in vitro, as per the manufacturer’s instructions. To examine cell fate, flow cytometry was conducted on day 5 following restimulation for 5 h with 50 ng/ml PMA, 750 ng/ml ionomycin, and 2 µl/ml protein transport inhibitor mixture (last 4 h).

Seven-to-eight-week-old female C57BL/6 mice (n = 12; Charles River Laboratories, USA) were cohoused in conventional housing at three or four animals per cage. Endometriosis was induced (n = 12) as previously detailed (34). Briefly, uterine horns were harvested from donor mice (n = 3), and an epidermal punch tool was used to obtain 3-mm endometrium fragments. Recipient mice were anesthetized with 3.5% isoflurane, and a small incision was made in the abdominal wall to access the peritoneum, onto which two uterine fragments (3 mm³) were adhered using Vetbond tissue adhesive (1469SB, 3M, USA). Lesions were allowed to establish for 1 wk postoperative. Sham surgery was conducted (n = 12), in which mice underwent surgery, although no uterine fragments were adhered to the peritoneum. Postsurgical care was conducted as outlined (28). To examine the effects of IL-23 in vivo, both endometriosis and sham mice were treated with i.p. injections of 1 µg of rmIL-23 (n = 6/group; 589006, BioLegend) or PBS for controls (n = 6/group), three times a week for 3 weeks (Fig. 6A). The mice were euthanized 21 d postoperative by 5% isoflurane inhalation followed by cervical dislocation. PF, endometriosis-like lesions, uterine horns, and spleens were collected, processed, and stored prior to analysis as previously detailed (34). For flow cytometry, PF and splenocytes were thawed and resuspended in 10% FACS. The cells were centrifuged at 300g for 5 min at 4°C, and supernatant was decanted before resuspension in 100 µg/ml DNAse I (10104159001, Sigma-Aldrich, Canada), as per the manufacturer’s protocol. The cells were neutralized with 10% FACS, centrifuged at 300g for 5 min at 4°C, and supernatant was decanted before resuspension in 2% FACS for flow cytometry.

Paraformaldehyde-fixed, paraffin-embedded tissue blocks were sectioned at 4-µm thickness and deparaffinized using Citra Solv and rehydrated in graded concentrations of ethanol. Following Ag retrieval for 20 min using citrate buffer and blocking of endogenous peroxidase activity, sections were stained with polyclonal Abs for anti-mouse Ki67 (1:1500; ab15580, Abcam, UK) and anti-mouse CD31 (1:100; 77699S, New England Biolab, USA) to assess proliferation and vascularization, respectively, and digitally scanned for analysis as previously described (34). A computer-generated algorithm for Ki67 and CD31 was used to digitally analyze lesions. Custom trained cytonuclear and area quantification modules were used to quantify Ki67 and CD31 staining, respectively. Standard H&E staining was conducted as previously explained (41).

To assess whether T_H17 cells may be primed in vivo during lesion development, endometriosis was surgically induced in recipient mice (n = 16) as described (34). Endometriosis-induced mice either received daily i.p. injections of a T_H17 priming mixture (n = 8; 0.5 µg rmIL-6 [575704, BioLegend] and 1 µg rmTGF-β1 [763104, BioLegend]) or rested (n = 8) from days 0 to 3 prior to daily i.p. injections of 1 µg rmIL-23 (n = 4/group) or PBS (n = 4/group) for 10 d (Fig. 8A). The mice were euthanized 14 d postoperative, and PF, spleen, and endometriosis-like lesions were collected. Lesions were snap-frozen in liquid nitrogen prior to storage at −80°C for later analysis, and PF was processed for cryopreservation, as detailed (34), for later analysis via flow cytometry. The splenocytes were isolated as previously outlined (34) prior to subjection to ACK lysis buffer (A1049201, ThermoFisher) per the manufacturer’s instructions. The splenocytes were neutralized with RPMI 1640 (11875085, ThermoFisher) and centrifuged at 400g for 5 min at 4°C prior to resuspension and counting. The samples were seeded in 24-well plates at 1 × 10⁶ cells/well and incubated at 37°C with 5% CO₂ with 50 ng/ml PMA and 500 ng/ml ionomycin for 4 h to stimulate intracellular cytokine production. After 1 h of incubation, 2 µl/ml of protein transport inhibitor mixture was added. Flow cytometry was conducted to examine cell fate.

Cell suspensions were counted and aliquoted (5 × 10⁵ cells/sample) before staining with human TruStain FcX (422302) or TruStain FcX (101319) to block nonspecific Ag binding, as per the manufacturer’s instructions. All products were purchased from BioLegend unless otherwise stated. The cells were stained with extracellular Abs as detailed in Table II. To assess viability, human samples were stained with eBioscience fixability viability dye eFluor 780 (65-0865-18, ThermoFisher), whereas murine samples were stained with Zombie green fixable viability dye (423111). All cells were then fixed and permeabilized using the eBioscience Foxp3/transcription factor staining buffer set (00-5523-00, ThermoFisher), as per the manufacturer’s instructions. Following permeabilization, the cells were stained with intracellular Abs (Table II). The cells were washed with 2% FACS, and the data were acquired using CytoFLEX S (Beckman Coulter, USA) and analyzed via FlowJo version 10 software (BD, USA). Fluorescent minus one controls were used appropriately to determine gating strategy. To set a viability gate, an aliquot of heat-killed cells was combined 1:1 with live cells and stained with viability dye.

GraphPad Prism9 (USA) was used for all statistical analyses. When appropriate, an unpaired, nonparametric Student t test with Mann–Whitney correction was used to analyze two groups, and a one-way ANOVA with Tukey post hoc test was used to compare three or more groups. Error bars signify SD, and statistical significance was denoted by a p value ≤ 0.05.

To assess endometriotic lesions as a source of IL-23 and gene expression variations between matched patient ectopic and eutopic tissues, a RT-qPCR array was designed to target 11 key genes known to play a role in the IL-23/T_H17 axis (Fig. 1). We discovered IL23R, IL17A, IL17F, IL21, IL22, IL12B, and TGFB1 were significantly upregulated in patient ectopic tissues compared with both matched eutopic tissues and controls (Fig. 1A, 1B). Notably, IL23R (IL-23 receptor) was also significantly upregulated in patient eutopic tissues compared with controls (Fig. 1C).

Using ELISA, we found IL-23 protein was significantly increased in patient plasma compared with controls (Fig. 1D). Patient plasma was then stratified into preoperative, postoperative, and 3-mo post endometriosis lesion excision surgery time points to determine whether lesion presence alters circulating IL-23 levels. However, no significant differences were found between time points (Fig. 1E). IL-23 protein was also detected in patient ectopic lesion and eutopic samples, although no significant differences were revealed when comparing eutopic and ectopic tissues (Fig. 1F).

We aimed to depict spatial localization of IL-23 within a human tissue microarray (Fig. 2) containing matched patient ectopic (endometrioma) and eutopic tissue samples (n = 17; Fig. 2A, 2B) compared with control endometrium (n = 10; Fig. 2C). IL-23 protein expression was quantified using area quantification of percentage of positive stain of anti–IL-23 stain across the total core area (Fig. 2D) and independently within epithelial (luminal and glandular) and stromal tissue (Fig. 2E). Although there were no significant differences in IL-23 protein expression between matched patient (ectopic/eutopic) or control endometrium, IHC staining illustrates that IL-23 is predominately expressed in epithelial tissues compared with stroma of matched patient (ectopic/eutopic) and control endometrium.

In accordance with literature, we sought to investigate IL-23 involvement in human pT_H17 cell derivation using flow cytometry (Fig. 3A) and multiplex cytokine array. We found primary human CD4⁺ T cells treated in vitro with IL-23 mixture (rhIL-23, rhIL-21, rhTGF-β1, rhIL-1β, rhIL-6, anti–IFN-γ, and anti–IL-4) produced significantly increased the frequency and number of T_H17 cells (Fig. 3B, 3C). We then aimed to determine how T_H17 cells driven in the presence of IL-23 may influence the endometriotic microenvironment. Multiplex cytokine array results reveal that activated T_H17 cells driven in the presence of IL-23 produce significantly increased IL-17A/IL-17F and decreased IL-10, suggesting a pathogenic phenotype (Fig. 3D, 3E). Activated T_H17 cells driven in the presence of IL-23 also secreted significantly increased MCP-1 and CCL5/RANTES (Fig. 3F, 3G), as well as CXCL9 and macrophage-derived chemokine (MDC; data not shown).

We sought to determine the independent effects of IL-23 on cell lines representative of the endometriotic lesion microenvironment (Fig. 4). Specifically, selected cell lines were representative of endometriotic epithelial (12Z) cells and hESCs, as well as vasculature/endothelial cells (HUVECs) and EECCs. Multiplex cytokine array revealed that IL-23 treatment significantly alters production of platelet-derived growth factor (PDGF)-AA and PDGF-AB/BB by 12Z cells, EECCs, and HUVECs (Fig. 4A–C), as well as IFN-γ–inducible protein 10 (IP-10) and fibroblast growth factor-2 (FGF-2) production by EECCs and HUVECs, respectively (Fig. 4E, 4F). Although hESCs also produced notable PDGF levels in response to IL-23 treatment, there were no significant differences between treatment groups (data not shown). Additionally, rhIL-23 treatment had dose-dependent effects on hESCs, with significant increases in both IL-1β (Fig. 4D) and Fms-related tyrosine kinase 3 ligand (FLT-3L; Fig. 4G).

To assess IL-23 influence on cell proliferation and apoptosis, we conducted WST-1 and caspase 3/7 glo assays, respectively (Supplemental Fig. 1). rhIL-23 treatment significantly increased hESC proliferation in 1 and 100 ng/ml treatment groups, although there were no significant differences in 12Z cell, EECC, or HUVEC proliferation. There were also no differences in cell death of rhIL-23–treated 12Z cells, EECCs, HUVECs, or hESCs. Moreover, we sought to determine IL-23 influence on angiogenesis using a tubulogenesis assay with HUVECs. However, no significant differences were found in tube length or number of total branching points when comparing rhIL-23–treated groups to the PBS control (Supplemental Fig. 2).

The literature shows that IL-23 signaling is necessary to drive, maintain, and expand pT_H17 cell fate (18, 20, 42). Thus, we sought to model this in vitro, examining whether IL-23 alters production of pT_H17 cells from naive CD4⁺ T cells isolated from murine splenocytes using flow cytometry (Fig. 5A). The results depicted treatment with an IL-23 containing T_H17 derivation mixture significantly increased frequency and number of pT_H17 cells from murine naive CD4⁺ T cells compared with all other treatment groups (Fig. 5B, 5C).

To assess IL-23 influence on immune cell recruitment/immune dysfunction in endometriosis, C57BL/6 mice (n = 12) underwent endometriosis induction surgery, and mice were treated with rmIL-23 (n = 6) or PBS (controls, n = 6) via i.p. injections three times a week for 3 weeks. Murine splenocytes and PF were collected to assess systemic and local immune alterations, respectively, via flow cytometry (Fig. 6). The results depicted significant alterations to CD4⁺ T cells, myeloid cells, T_REGs, and pT_H17 cells in PF, although there were no significant differences in immune cell subsets within murine splenocytes. As speculated, IL-23 significantly altered the T_H17/T_REG axis in murine PF, increasing T_REGs and decreasing pT_H17 cells (Fig. 6J, 6K), although these differences were not statistically significant in murine splenocytes (Fig. 6D, 6E). Significant differences seen in immune cell subsets within PF were not detected in our sham mouse model (Supplemental Fig. 3), highlighting the influence of IL-23 specifically in the context of endometriosis. IL-23 also significantly increased immune cell recruitment to the peritoneal cavity, irrespective of disease context (endometriosis/sham; data not shown).

In addition to inflammation, lesion proliferation and vascularization are known hallmark features of endometriosis pathophysiology (6, 43, 44). Thus, we aimed to examine the role of IL-23 on lesion vascularization and proliferation using an established endometriosis mouse model (Fig. 7). H&E staining confirmed that murine endometriotic lesions were representative of endometriosis lesion histology (Fig. 7A, 7B). Additionally, lesions from IL-23–treated mice had large, multinucleated cells consistent with giant cell histology (Fig. 7C). To our knowledge, this is the first report identifying giant cells within murine endometriosis-like lesions. Although present in some PBS-treated mice, IL-23–treated mice lesions had significantly more giant cells (Fig. 7D). Lesions were also subjected to IHC for markers of proliferation (Ki67) and vascularization (CD31; Fig. 7E, 7F, 7H, 7I). IL-23–treated mice had trends of increased lesion proliferation and vascularization, although the differences were not statistically significant (Fig. 7G, 7J).

IL-23 is known to promote inflammation and pT_H17 cell fate in combination with other inflammatory cytokines (18, 45). Thus, we sought to determine whether treatment with T_H17 priming mixture and/or rmIL-23 influenced production of pT_H17 cells, and in turn the inflammatory milieu, within an endometriosis mouse model (Fig. 8). Splenocytes and PF were collected and analyzed via flow cytometry; however, no statistically significant differences in immune cell subsets were detected in treatment groups compared with controls.

IL-17, a proinflammatory and proangiogenic cytokine largely secreted by T_H17 cells, has been linked in endometriosis pathogenesis (10). Although the exact mechanisms are unknown, IL-17 and other downstream T_H17 cell cytokines may exacerbate endometriosis development by recruiting immune cells to the lesion site, such as neutrophils, macrophages, and lymphocytes, collectively enhancing lesion establishment (12, 28). Specifically, IL-17 is increased in patient plasma and PF, and this is correlated with disease severity (46). Additionally, IL-17 is produced by lesions, and postexcision surgery systemic IL-17 levels are significantly reduced (10), highlighting the potential link between IL-17 and endometriosis. T_H17 cells are also dysregulated in endometriosis (24, 29, 33) and associated with disease severity (13). Thus, whereas downstream IL-17 has been examined in endometriosis, a knowledge gap exists regarding the factor(s) driving T_H17/IL-17 axis dysregulation in endometriosis. IL-23 is known to govern T_H17 cell fate (20, 47), driving pT_H17 cells producing IL-17 and downregulating anti-inflammatory IL-10 (17, 18). Hence, as emerging evidence has underscored the role of IL-23 in driving this axis to produce IL-17 in various other diseases (30, 48), we sought to examine whether IL-23 may be driving T_H17/IL-17 axis dysregulation in endometriosis to exacerbate disease.

Our studies using patient samples reveal significant dysregulation in the IL-23/T_H17 axis in patient tissues. To our knowledge, this is the first report quantifying RNA and protein expression of IL-23 (p19 [IL23A] and p40 [IL12B] subunits) and key mediators in the IL-23/T_H17 axis in endometriosis patient ectopic (lesion) and matched eutopic endometrial tissues compared with healthy controls. RT-qPCR results show that IL17A, IL17F, IL21, IL22, IL12B (IL-23 p40 subunit), and TGFB1 are significantly upregulated in ectopic tissues compared with matched eutopic tissue and control samples. This is substantial because IL-17A, IL-17F, IL-21, and IL-22 are secreted by T_H17 cells, whereas IL-23 and TGF-β1 are important in driving and maintaining pT_H17 cell fate (20, 47). RT-qPCR results also that suggest patients may have increased sensitivity to IL-23 signaling, because IL23R was significantly upregulated in both ectopic and eutopic patient tissues compared with controls. Moreover, IL12B, encoding the p40 subunit shared between IL-23 and IL-12, was significantly increased in accordance with literature (49). This increase is unlikely to indicate IL-12 involvement because the p35 subunit (IL12A), specific to IL-12, was not significantly increased. Instead, this may indicate involvement of p40 homodimers or p40 monomers, which are increased in endometriosis PF (50).

Complimentary to RT-qPCR results, IL-23 protein was detected in endometriotic lesions and was significantly increased in patient plasma compared with controls. Although IL-23 protein expression was not significantly different across patient ectopic (ovarian endometrioma) or eutopic tissues compared with controls, IL-23 was significantly increased in luminal and glandular epithelium compared with stroma. Because endometrioma lesions are more cystic in nature, they are comprised of less luminal and glandular epithelium compared with noncystic lesions. Thus, because we illustrate that IL-23 is predominately localized to epithelial tissue, IL-23 protein is likely reduced in endometrioma lesions compared with other lesion subtypes. However, further research is required to quantify differential IL-23 protein expression across lesion subtypes. Others do report significantly increased IL-23 protein in endometriosis patient PF, ectopic lesions, and serum, which is correlated with increased IL-17 levels (51–54). Moreover, IL-23 is significantly increased in both patient follicular fluid and serum in later stages of disease (III and IV) compared with earlier stages (I and II) (52), highlighting that IL-23 may be associated with disease severity. However, some report no significant differences in IL-23 in patient serum when not stratified by disease stage (52, 53). Although we did not capture significant differences in IL23A (p19 subunit) in patient tissues at the mRNA level, dysregulation of other key members within IL-23/T_H17 axis combined with significantly increased IL-23 protein levels in patient plasma indicate involvement in endometriosis pathophysiology.

In our in vitro studies, IL-23 mixture treatment in combination with CD3/CD28 activation significantly promoted T_H17 cell differentiation from human and murine naive CD4⁺ T cells. This mixture was selected because evidence shows the endometriosis milieu contains and is influenced by TGF-β1, IL-1β, IL-6, IL-21, and IL-23 cytokines (52, 55–58), which are also vital in driving pT_H17 cell fate (17, 59). Thus, we aimed to partially recapitulate this environment in vitro. The results depict that naive human CD4⁺ T cells driven in the presence of this combination treatment produce significantly increased IL-17A/F and decreased IL-10, characteristic of pT_H17 cells, confirming that T_H17 cells exposed to IL-23 mixture acquire a pathogenic phenotype consistent with literature (12, 20, 47). Additionally, human naive CD4⁺ T cells driven with this combination treatment produce significantly increased MCP-1, CCL5, CXCL9, and MDC. Because these factors play important chemotactic roles, IL-23–driven and activated T_H17 cells may promote immune cell recruitment and produce proinflammatory factors, such as IL-17, to further exacerbate lesion inflammation, proliferation, and angiogenesis. However, although there are reports of pT_H17 cells producing CCL5 (12), it is unclear to what abundance T_H17 cells may produce MCP-1, CXCL9, and MDC. Further studies are required to examine coproduction of these factors by T_H17 cells.

Broadly, IL-23 is produced in response to damage-associated molecular patterns (DAMPs) (60). Because the literature suggests that DAMPs/alarmins are released during endometriosis lesion establishment (56), IL-23 may not only act in endometriosis to promote and support pT_H17 cell fate but also act directly on the lesion to perpetuate inflammatory signals in response to DAMPs. Thus, we sought to determine the direct effects of IL-23 treatment on cells representative of the endometriotic microenvironment, specifically 12Z cells, EECCs, HUVECs, and hESCs. The results depicted that in vitro rhIL-23 stimulation of cell lines significantly altered production of PDGF-AA from 12Z cells, PDGF-AB/BB levels from both EECCs and HUVECs, and FGF-2 production from HUVECs. Additionally, in vitro rhIL-23 treatment produced significant dose-dependent increases of IP-10 by EECCs and IL-1β and FLT-3L by hESCs. rhIL-23 treatment also significantly increased proliferation of hESCs, a major component of the endometriotic lesion. Collectively, our in vitro data suggest that IL-23 stimulation significantly alters production of factors known to play a role in lesion establishment and maintenance and may promote lesion growth/survival.

Our mouse model of endometriosis has been extensively used to gain mechanistic insights into the complex pathophysiology of endometriosis (34, 61–63). Indeed, we identify significantly increased numbers of large, multinucleated cells in lesions obtained from IL-23–treated mice compared with controls. Histologically, these cells appear to be giant cells of monocyte origin; however, further characterization is necessary. Although we captured trends of increased lesion proliferation and vascularization in IL-23–treated mice, these differences were not statistically significant. This may be reflective of our mouse model, time course, and dose/route of IL-23 administration, because IL-23 has known effects on cell proliferation and vascularization (60, 64–66). We also used this mouse model to determine whether IL-23 treatment alters systemic and/or local immune cell subsets and in turn may promote immune dysfunction associated with endometriosis. Indeed, IL-23 treatment significantly reduced CD4⁺ T cells and myeloid cells in murine PF compared with controls, possibly due to homing of CD4⁺ T cells and myeloid cells to the lesion. Additionally, IL-23 treatment dysregulated the local T_H17/T_REG axis, where pT_H17 cells were significantly reduced and T_REGs were significantly increased in PF of IL-23–treated mice. This T_H17/T_REG axis dysregulation is noteworthy because balance between T_H17 and T_REGs is necessary to maintain immune homeostasis and avoid dysfunction (23, 67–69). These significant alterations in immune cell subsets within murine PF were not observed in our sham mouse model, highlighting that these differences were specific to an endometriosis-like microenvironment. Finally, because the literature suggests that TGF-β1 and IL-6 are required for initial derivation of T_H17 cells, whereas IL-23 is necessary for full induction and maintenance of pT_H17 cell fate (70, 71), we used a mouse model to assess the influence of IL-23 alone and in combination with a T_H17 priming mixture (TGF-β1 and IL-6) during lesion establishment. Our short time-course study revealed no significant differences in murine immune cell subsets between groups. Thus, further research is required to establish the specific combination of cytokines and timeframe necessary to capture murine T_H17 differentiation in vivo. Our in vivo data reveal that IL-23 treatment significantly influences local immune dysfunction in endometriosis and promotes immune cell recruitment to the peritoneal cavity, which may exacerbate local immune dysfunction to promote lesion survival and the well known inflammatory milieu in endometriosis.

We acknowledge the limitations of this study and others investigating endometriosis due to limited access acquiring well categorized patient samples stratified by disease stage, subtype, and menstrual cycle. Invasive diagnostic measures, heterogenous symptom presentation, and stigmatization of reproductive disorders also complicate this, commonly resulting in long diagnostic delays and misdiagnosis of endometriosis patients. Due to limitations in obtaining fresh patient samples, we were unable to isolate primary cells from endometriosis lesions, and thus we acknowledge known limitations of immortalized cell lines. However, we supplement these data with data from primary human CD4⁺ T cells. Moreover, although EECCs are widely used in endometriosis literature to examine the molecular mechanisms of endometriosis pathophysiology (72–74), we acknowledge the limitations of this cell line. Specifically, because EECCs originate from the endometrium of a carcinoma patient, EECCs may not be exactly representative of “normal” eutopic endometrium (75). Finally, while we use our syngeneic immunocompetent endometriosis mouse model to provide insights into the role of IL-23 in endometriosis pathophysiology, this model is not able to completely recapitulate the chronic nature and heterogenous presentation of human disease, because rodents do not spontaneously develop endometriosis.

In conclusion, while our data suggest IL-23/T_H17 axis dysregulation in endometriosis, it is important to note that no existing evidence in literature has mechanistically linked this axis to endometriosis pathology. In this article, we report significant IL-23/T_H17 axis dysregulation in endometriosis patients, and this was captured in in vitro studies and our in vivo mouse model. Because there is currently no effective long-term, noninvasive, or nonhormonal therapy option for endometriosis patients who aim to conceive, there is an urgent need to explore novel therapeutics. Future directions should investigate the correlation between IL-23 and IL-17 levels in circulation and locally in patient PF, because others report a significant positive correlation of IL-23 and IL-17 levels in endometriosis patient plasma and PF (51). Similarly, examining the correlation between IL-23 and IL-10 levels in circulation and PF may provide insights into local and systemic immunomodulatory effects of IL-23. Ultimately, IL-23 may provide a promising therapeutic avenue to reduce the large burden of endometriosis, because IL-23–targeting therapeutics have been approved for human use and/or are currently in development for various chronic inflammatory diseases, including rheumatoid arthritis and psoriasis (32, 76–80).

The authors have no financial conflicts of interest.

We thank Oliver Jones at Queen’s CardioPulmonary Unit (Queen’s University) for all histology processing. We also thank Wei Wang at Kingston Health Sciences Centre (Queen’s University) and Yuexin Yang in the Department of Pathology and Molecular Medicine (Queen’s University) for immunohistochemistry processing. Finally, we thank Xiao Zhang at Queen’s Laboratory for Molecular Pathology (Queen’s University) for aid with slide scanning.