Pro-endometriosis macrophage release of IL-33 is key for endometriosis pain and lesion formation

Animal handling and care was conducted in accordance to the United States National Institutes of Health Guide for the Care and Use of Laboratory Animals, and Australian National Health and Medical Research Council (NHMRC), and were in accordance with the laws of the United States and regulations of the Department of Agriculture. All experiments and procedures were approved by the Institutional Animal Care and Use Committee (IACUC) at South Australian Health and Medical Research Institute (protocol number SAM-23–040) for calcium imaging experiments), at Londrina State University (protocols numbers 2669.2019.54 and 047.2023 for ST2-KO experiments), and at Boston Children’s Hospital (protocols numbers 19-12-4054R and 00001816 for all remaining experiments). Healthy C57BL/6 mice (Stock #000664), B6.129-Trpv1 tm1(cre)Bbm /J (TRPV1cre, stock #017769), B6.129P2-Gt(ROSA)26 Sortm1(DTA)Lky /J (ROSA-DTA, stock #009669), C57BL/6-Tg(Csf1r-cre)1Mnz/J (Csf1r-cre, stock #029206), and B6;129S6-Gt(ROSA)26Sortm96(CAG-GCaMP6s)Hze/J (GCaMP6s, stock #024106) were purchased from Jackson Laboratories. C57BL/6N- Ramp1 tm1c(EUCOMM)Wtsi /H (RAMP1 fl/fl ) mice were kindly provided by Dr. Isaac Chiu. Csf1r-cre mice were bred with RAMP1 fl/fl to generate C57BL/6-Tg(Csf1r-cre)1Mnz/J +/− ;RAMP1 fl/fl that were then bred with RAMP1 fl/fl to generated mice with deletion of RAMP1 on macrophages (Csf1r ΔRamp1 ) BALB/c and ST2 −/− (ST2-KO, BALB/c background) were received from Ribeirão Preto Medical School (University of São Paulo, Ribeirão Preto, SP, Brazil). B6.129(Cg)-Scn10a tm2(cre)Jwo/Tjp J mice (Nav1.8cre, stock #036564) mice were a gift from W. Imlach, Monash University, Australia. Nav1.8cre mice were bred with GCaMP6s mice to generate Nav1.8-GCaMP6s or LM control mice. Female age-matched mice from 8 to 14 weeks of age were used for experiments involving transgenic mouse strains while 8-week-old mice were used for the experiments performed in C57BL6/J mice. Block randomization was used to randomize subjects into groups resulting in an equal sample size at all time-points. TRPV1 nociceptor ablation was performed as previously described ( 55 ). For subcutaneous RTX (cat #R400, Alomone), escalating doses of RTX (30, 70, 100 μg/kg on consecutive days) or vehicle (2% DMSO, 0.15% Tween-80 in sterile PBS) were used. Mice were under isoflurane anesthesia. For chemical depletion of macrophages, clodronate liposomes or vehicle control liposomes (Encapsula Nanaoscience) at 0.0625 mg/mouse ( 56 ) were injected intraperitoneally 24h before endometriosis induction. Mice were randomly assigned and housed in standard clear plastic cages with no more than 5 mice per cage in a 12:12h light/dark cycle with ad libitum access to water and food. Behavioral testing was performed between 9 a.m. and 5 p.m. by an investigator who was blind to treatment group in a room maintained at a temperature of 21±1°C. Blinding was maintained until all data analysis decisions were final. All efforts were made to minimize the number of animals used and their suffering. Euthanasia was performed by controlled CO 2 inhalation. Biopsies of endometriotic lesions were obtained from patients with a laparoscopic diagnosis of endometriosis at St. Marien Hospital, Amberg, and St. Hedwig Clinic, Regensburg, Germany. The samples were collected during days 3–5 of the menstrual cycle, corresponding to the proliferative phase. This study was approved by the Ethics Committee of the University of Regensburg, Germany (IRB approval numbers 21-2427-101, 23-3211-101, 22-2862-104). Written informed consent was obtained from all patients. For more information, please refer to the “human sample collection and demographic” section in the supplementary materials . Results are presented as mean ± SEM and followed ARRIVE 2.0 guidelines ( 57 ). Data were analyzed using the GraphPad Prism version 10.5 software. Two-way repeated measure analysis of variance (ANOVA), followed by Tukey’s post hoc , was used to analyze data from experiments with multiple time points (von Frey). One-way ANOVA followed by Tukey’s post hoc was used to analyze data from experiments with a single time point. To explore potential associations among IL-33 expression, glandular proliferative status, extent of fibrosis, duration of symptoms, revised American Society for Reproductive Medicine (rASRM) score, worst pain score, age at surgery, and age at menarche, a Spearman correlation analysis was conducted. Comparison between two groups was conducted using two-tailed Student’s t-test. Results are presented as Spearman’s correlation coefficients (ρ). For all analysis, differences were considered significant when p < 0.05.

We previously showed that macrophages respond to CGRP and stimulate endo-epi cell growth ( 11 ). We repeated these experiments and confirmed that CGRP-induced PEMs increase the growth of endo-epi cells in a co-culture system ( Fig 1A ). This effect was RAMP1-dependent as treatment with rimegepant, a RAMP1 antagonist, reduced PEM-induced endo-epi cell growth ( Fig 1 ). This demonstrates that PEMs release factors that support endometrial cell growth. To identify mediators involved in this cell growth, we performed bulk RNAseq of vehicle- or CGRP-stimulated macrophages as well as naïve endo-epi cells. Our bulk RNAseq data revealed 111 differentiated expressed genes (DEGs) in PEMs, including changes in Il33 transcripts ( Fig 1B ). Using CellPhoneDB to identify which transcripts might be involved in cell-cell signaling. We examined the endo-epi cell RNAseq data to assess the mRNA expression level (in transcripts per million, TPM) for each of the receptors for differentiated expressed genes. We found that of these receptors, the IL-33 receptor ST2 ( Il1rl1 )/IL-1Rap ( Il1rap ) was highly expressed by endo-epi cells ( Fig 1C , right side of table). Interestingly, ST2 was one of the few receptors expressed by endo-epi cells whose ligand was also increased in CGRP-mediated PEMs ( Fig 1B ), suggesting that PEM-induced endo-epi cell growth was via IL-33/ST2 signaling ( Fig 1C ). To confirm this, we performed a cell proliferation assay using vehicle-control- or IL-33-stimulated endo-epi cells. We found that mouse recombinant IL-33 induced endo-epi cell proliferation ( Fig 1D , left graph). This increased proliferation is ST2-dependent, as treatment with an anti-ST2 antibody impaired IL-33-induced cell proliferation, in a concentration-dependent manner, vs. IgG control ( Fig 1D , right graph). To understand the mechanisms involved in this cell growth, we performed bulk RNAseq of vehicle- or IL-33-stimulated endo-epi cells. Our bulk RNAseq analysis revealed a total of 408 DEGs ( Fig 1E and S1 ). Among the upregulated genes, we found genes related to cell proliferation such as cellular communication network factor 4 ( Ccn4 ), follistatin ( Fst ), and insulin-like growth factor-binding protein 7 ( Igfbp7 ); while downregulated genes such as dual specificity phosphatase 1 and 22 ( Dusp1 and Dusp22 ) are involved in controlling apoptosis ( Fig 1E ). Finally, we confirmed that in mouse endometriosis lesions, ST2 + cells were also present both in the epithelium and stroma with increased presence in the epithelium ( Fig 1F and G , left graph). Interestingly, we found that ST2 + epithelial cells were also proliferating as observed by co-localization with Ki-67 + cells ( Fig 1F and G , right graph). Altogether, we found that IL-33/ST2 signaling is involved in cell proliferation both in vitro and in vivo . Having observed a crucial role for IL-33/ST2 signaling in mouse lesions, we next wanted to determine whether IL-33 was also in human lesions. To that end, we examined 13 lesions from patients with a mean age of 34 years old ( Supplementary Table 1 ). We observed widespread immunohistochemical staining for IL-33 throughout the lesions, particularly in both the nucleus and cytoplasm of the epithelium of endometrial glands, in the stroma of fibrotic regions, and the endothelium of blood vessels ( Fig 2A ). We also found Ki67 + epithelial cells in human lesions ( Fig 2A ), though to a lesser extent than in mouse lesions. Nevertheless, this indicates that our mouse endometriosis lesions recapitulate features observed in human lesions ( Fig 1F ). Importantly, overall IL-33 staining is positively associated with the number of glands in lesions ( Fig 2B , top left graph). As fibrosis is an important hallmark of endometriosis lesions ( 33 ) and IL-33 is known to drive fibrosis in different organs ( 34 ), we next evaluated the relationship between IL-33 and fibrosis. Within each biopsy, the relative extension of fibrosis, characterized by typical dense clusters of elongated cells in the vicinity of the endometrium-like lesion, was scored on a scale from 0 to 4. We found that IL-33 staining in glands is positively associated with its staining in the fibrotic areas ( Fig 2B , top right graph). We also found that the IL-33 is positively associated with overall fibrotic score ( Fig 2B , bottom left graph), further suggesting that IL-33 can drive fibrosis. The extent of fibrosis demonstrated a trend toward a positive correlation with the duration since symptom onset (ρ = 0.671, p = 0.06, n = 8). Although this result did not reach statistical significance, it suggests that longer symptom duration may be associated with greater fibrotic development, supporting the hypothesis of a progressive fibrotic process in ectopic lesions. We also evaluated correlation with and among other patient demographic, symptom, and histological factors, but found no significant correlations ( Supplementary Table 2 ). Altogether, this indicates that IL-33 is correlated with a higher number of glands as well as a greater extent of fibrosis in human lesions. GWAS show that IL-33 is associated with the risk of developing endometriosis ( 24 , 25 ). To expand such findings, we then used the single cell disease relevance score (scDRS) to specifically pinpoint the cell type associated with such risk. scDRS is an algorithm that integrates single cell expression data with germline risk associations from large genome-wide association studies ( 35 ). Here we integrated GWAS data from a meta-analysis of 450,668 controls and 23,492 endometriosis cases with single cell data from endometrium and endometriosis lesions from a cohort of 21 patients ( 30 , 32 , 36 ). Focusing on immune cells, we found that IL-33 expression in several cells such as plasma cells, dendritic cells, effector CD8 T-cells, mast cells, and “M1” and “M2” macrophages is significantly correlated with endometriosis risk score ( Fig 2C , left graph). We then sought to determine whether IL-33 expression in these immune cells would be particularly associated with the risk of developing different types of endometriosis lesions (e.g., superficial vs. deep vs. endometrioma, etc.). We found that, except for endometriomas (ovarian lesions), IL-33 signal in macrophages was widespread for different types of endometriosis lesions, indicating that it might be an important driver of lesion formation or maintenance ( Fig 2C , right graph). IL-33 signal in other cell types such as plasma cells and dendritic cells showed a similar pattern, indicating these cells could also play a role in IL-33-driven lesion formation. To further expand these results, we next evaluated a transcriptional signature of IL-33 signaling (GO:0038172, see Supplementary Material ). We found that IL-33-mediated signaling as well as some of the downstream markers of IL-33/ST2 signaling were enriched in macrophages ( Fig 2D ), further indicating that IL-33 in macrophages could play a role in endometriosis lesion formation or maintenance. We previously demonstrated that TRPV1 + nociceptors control monocyte recruitment during endometriosis, thereby facilitating lesion growth and pain ( 11 ). However, whether these lesion-recruited macrophages produce IL-33 and are sensitive to nociceptor ablation was unknown. Therefore, we bred Trpv1 -cre mice with Rosa26 lox -STOP- lox DTA mice (ROSA-DTA) to create Trpv1 cre Dta nociceptor-ablated mice and Cre – littermate control (LM) mice. Using immunofluorescence of mouse lesions, we found that nociceptor-ablated mice had fewer F4/80 + macrophages as well as less IL-33 ( Fig 3A ). In agreement, we found a reduction in the pixel count as well as in the co-localization area between F4/80 and IL-33 upon ablation of TRPV1 + nociceptor ( Fig 3B ). This indicates that TRPV1 + nociceptors control the presence of F4/80 + IL-33 + macrophages in the lesion. We have previously observed that endometriosis lesions are supported by TRPV1 + nociceptors which release CGRP, which then acts on local macrophages ( 11 ). Our next step was to determine the extent to which IL-33 production was driven by this nociceptor to macrophage communication. For that, we used chemical and genetic approaches to ablate nociceptors ( Fig 4A ) or macrophages ( Fig 4B ) as well as four different FDA-approved drugs that target either CGRP or RAMP1 ( Fig 4C ) to measure IL-33 levels in the lesions. We first performed chemical ablation of nociceptors using subcutaneous injection of resiniferatoxin (RTX), an agonist for TRPV1 ( 39 , 40 ) ( Fig 4A , left graph). Upon chemical ablation of TRPV1 + nociceptors, we found a reduction in IL-33 levels in the lesion when compared to vehicle-treated mice ( Fig 4A , left graph). In agreement, genetic ablation of the same population using Trpv1 cre Dta mice reduced IL-33 levels as well ( Fig 4A , right graph), further indicating that nociceptors regulate IL-33 production. We next used FDA-approved drugs that target either CGRP (galcanezumab and fremanezumab, monoclonal antibodies) or RAMP1 (rimegepant or ubrogepant, small molecules). We found these drugs also reduced IL-33 levels ( Fig 4B ), supporting the notion that CGRP/RAMP1 signaling drives IL-33 production. Then, we chemically ablated peritoneal cavity macrophages using clodronate liposomes ( Fig 4B , left graph). Peritoneal-cavity-macrophage-depleted mice displayed lower levels of IL-33 when compared to vehicle-treated mice ( Fig 4C , left graph). Finally, to confirm that this reduction in IL-33 was a result of nociceptor CGRP to RAMP1 signaling in macrophages, we bred a macrophage-specific cre line ( Csf1r -cre) with Ramp1 fl/fl to delete Ramp1 from macrophages ( Fig 4C , right graph Csf1r ΔRamp1 ). Mice with Ramp1 -deleted macrophages showed reduced IL-33 production when compared to LM controls, demonstrating that nociceptor to macrophage communication via CGRP/RAMP1 signaling drives IL-33 production ( Fig 4C , right graph). Having observed that IL-33 induces endo-epi cell proliferation ( Fig 1D ) and that IL-33 staining correlates with the number of glands in human lesions ( Fig 3B , left graph), we next sought to determine whether IL-33 signaling is required for lesion formation. We first determined the dynamics of IL-33 production in the peritoneal cavity wash upon endometriosis induction by ELISA. We found that IL-33 peaks 3 days post induction and then returns to baseline ( Fig 5A , left bar graph), indicating that in the peritoneal cavity a burst of IL-33 is followed by a normalization of its levels during endometriosis induction. Interestingly, we detected IL-33 levels in the lesions at time points in which peritoneal cavity IL-33 levels were at baseline, e.g. 14 and 28dpi ( Fig. 5A , right bar graph). This indicates that IL-33 may be a crucial lesion regulator. To confirm that IL-33/ST2 signaling is required for lesion formation, we induced endometriosis using uterine horns of WT or mice lacking St2 ( Fig 5B ). When endometriosis was induced with uterine horns of St2 -KO mice, we observed a reduction in the number of lesions, lesion size, and burden ( Fig 5C ). We also found that remaining lesions had less IL-33 ( Fig 5D ). To expand these results, we next used a pre-treatment schedule (i.e., treatment before endometriosis induction) with an anti-IL-33 antibody ( Fig 5E ). We found that anti-IL-33 antibody treatment reduced evoked pain as observed by an increase in the mechanical threshold ( Fig 5F ) and also reduced both abdominal contortions and squashing ( Fig 5G ). Pre-treatment with anti-IL-33 antibody also reduced the number of lesions, lesion size and burden ( Fig 5H ). This indicates that IL-33/ST2 signaling is required for lesion formation, establishment, and endometriosis-associated pain. We next determined whether anti-IL-33 therapy could effectively intervene in ongoing disease (e.g., treatment starting at 29 dpi, when lesions are fully developed in our mouse model, Fig 6A ). Like the pre-treatment protocol, we found that post-treatment with anti-IL-33 antibody also reduced evoked pain as evidenced by an increase in the mechanical threshold ( Fig 6B ) and decreases in both abdominal contortions and squashing ( Fig 6C ). However, with this post-treatment protocol, we did not observe significant changes in lesion development ( Fig 6D ). Previous works have demonstrated that nociceptors express ST2 ( 20 , 22 , 34 , 41 ) and that injection of recombinant IL-33 induces pain ( 16 – 18 , 42 , 43 ), indicating that IL-33 directly activates nociceptors. Based on that data from the literature and our own data, IL-33 might have a dual role in endometriosis. At early time points, IL-33 participates in lesion establishment and once lesions are formed, IL-33 increases endometriosis pain. To address the question of whether IL-33 can mediate pain during endometriosis through direct activation of nerve fibers, we first performed immunofluorescence of mouse endometriosis lesions. We found that they contain ST2 + nerve fibers as observed by double staining of ST2 with PGP9.5 (pan neuronal marker, Fig 6E ). We then performed calcium imaging of DRG neurons from mice with and without endometriosis, to investigate the effect of IL-33 on sensory neuron’s activity. For that, we used Na V 1.8-GCaMP6s mice which express the calcium sensor GCaMP6s in Na V 1.8 + nociceptors ( Fig 6F ), as previously described ( 44 ). We found that IL-33 directly activates a proportion of DRG neurons cultured from both sham and endometriosis-lesion bearing mice ( Fig 6G and H ). Specifically, we found that endometriosis induced a >4-fold increase in the number of IL-33-responsive nociceptors with 122 out of 1006 responding DRG neurons in endometriosis mice vs 28 out of 912 responding DRG neurons in sham mice ( Fig 6I ). We also found that the overall magnitude of the response to IL-33 was higher in DRG neurons from mice with endometriosis, compared to those cultured from sham mice ( Fig 6J ). This indicates that: i) IL-33 directly activates nociceptors, ii) that nociceptors from mice with endometriosis are more prone to respond to IL-33, and iii) that the degree of IL-33-induced activation is higher in nociceptors from mice with endometriosis. For better visualization, traces of the IL-33-responsive DRG neurons are also shown in Fig S3 . Altogether, our data further supports the notion that IL-33 mediates pain through direct activation of ST2 + nociceptors and unveils a previously unknown mechanisms for pain generation during endometriosis.

To understand the endogenous role of IL-33 and its regulation during endometriosis, we first used unbiased techniques to discover that nociceptor-derived CGRP induces the expression and release of IL-33 by macrophages. We then employed a series of genetic and pharmacological manipulations to prevent IL-33 signaling. As IL-33 is the sole agonist for ST2 ( 12 , 34 , 45 ), we specifically determined the extent to which IL-33/ST2 signaling is required for lesion formation by using mice lacking St2 as tissue donors during endometriosis induction. We further confirmed that data by using treatment with monoclonal antibodies against IL-33 or ST2. To then understand how neuroimmune communication plays a role in IL-33 production, we applied a series of chemical and genetic approaches to ablate nociceptors and macrophages. Altogether, we unveil a hitherto unknown mechanism for neuroimmune-driven production of IL-33 as well as its potential dual role during endometriosis. Using scRNAseq and bulk RNAseq of mouse and human lesions, we identified a nociceptor to PEM pathway that drives IL-33 production. CGRP-driven production of IL-33 by PEMs is key to lesion growth and pain. This is in agreement with previous works showing a CGRP-mediated production of IL-33 by mesenchymal stromal cells ( 46 ) and innate lymphoid cells (ILC2s) ( 47 ). Interestingly, these IL-33 + mesenchymal stromal cells ( 48 ) and ILC2s ( 49 ) seem to be in close proximity to neurons, indicating that neuronal-derived molecules might be one of the regulators of IL-33 in these cells, as well. In humans, we found that IL-33 correlates with lesion fibrosis and gland numbers, indicating that IL-33 might be associated with both epithelial proliferation and fibrosis in humans. The latter observation is consistent with the ability of IL-33 to drive fibrosis in different organs ( 33 ). Furthermore, we also determined that IL33 transcripts in macrophages are associated with endometriosis risk in humans, as well as the different types of endometriosis lesions such as superficial, deep, and endometriomas as determined by scDRS. In these data, IL33 was also expressed in several other immune cell clusters such as dendritic cells, plasma cells, effector CD8 T-cells, suggesting that these cells could be also associated with endometriosis risk as well as the different types of lesions in patients. It remains to be determined whether ablating IL-33 from these cell types will play a role in endometriosis. A limitation of our scDRS approach is that the single cell data sets only capture the top 10–20% of genes based on absolute expression, and IL-33 expression, as is typical for cytokines, was only captured in a subset of cells. The observations made here, therefore, need to be taken with caution until further validation. Nevertheless, our combined human (IHC, scDRS) and pre-clinical data (behavior studies, cell proliferation, bulk RNAseq, calcium imaging assays, etc) demonstrate that IL-33 release by PEMs drives lesion formation and pain. We also provided evidence that suggest a dual role for IL-33 in endometriosis. IL-33/ST2 is required for lesion formation as our data demonstrates that uterine horn of mice lacking St2 generate fewer and smaller lesions. Furthermore, IL-33 induced a direct and potent activation of ST2 + nociceptors from endometriosis lesion-bearing mice. Therefore, targeting IL-33/ST2 signaling might be an effective way to treat endometriosis pain. Our data indicate that IL-33 plays multiple roles in endometriosis pathophysiology. During lesion induction, we found that IL-33 is rapidly released into the peritoneal cavity (peaks at 3 dpi) and then returns to baseline levels. This is consistent with studies showing that in addition to being constitutively produced and released, IL-33 can also be stored in the nucleus. Therefore, after tissue damage IL-33 can be released as an alarmin to initiate immune response ( 50 , 51 ), a process that likely occurs during retrograde menstruation. In vitro , we also show that IL-33 induces epithelial cell proliferation, corroborating studies showing IL-33 induces epithelial cell proliferation in skin cancer, colon cancer ( 52 ), during intestinal injury ( 52 ), and in nasal mucosal inflammation ( 53 ). Collectively, our data strongly support a role for IL-33/ST2 signaling in endometriosis lesion formation. In contrast, anti-IL-33 treatment of established lesions did not significantly reduce lesion size but substantially reduced pain. This points to an ongoing role for IL-33 in pain generation once lesions are fully formed. In agreement, we detected IL-33 levels in the lesions in time points where IL-33 levels in the peritoneal cavity wash had returned to baseline (e.g., 14 dpi). Even though ST2 staining is most prominent in the epithelium of lesions, we also found the presence of ST2 + PGP9.5 + nerve fibers, indicating the presence of IL-33-responding nerve fibers in the lesions. To support those observations, we found that DRG neurons of mice with endometriosis had an increased response to IL-33 when compared to sham mice. This further indicates that IL-33 directly activates nociceptors to trigger pain during endometriosis. Our data confirms data from the literature showing that ST2 is expressed by TRPV1 + sensory neurons ( 22 ) and that it exacerbates disease burden via nociceptor activation ( 20 – 22 ). Specifically for pain, our findings corroborates data showing that local injection of IL-33 is sufficient to trigger pain ( 16 – 18 , 54 ) while injection of suboptimal doses of IL-33 can also potentiate different stimuli to induce pain ( 16 ). Altogether, we suggest a dual role for IL-33/ST2 signaling during endometriosis. Initially, it is a major contributor to lesion formation but once lesions are fully formed an additional role is added in which throughout lesion lifetime IL-33/ST2 signaling continues to drive pain. In summary, we identified a neuroimmune-driven production of IL-33 production by PEMs that is key for lesion growth and pain during endometriosis in mice. We also found that IL-33 expression is associated with endometriosis risk in humans. Further, we provide evidence that IL-33 plays dual roles in endometriosis, contributing to both lesion formation as well as ongoing pain. Therefore, targeting IL-33/ST2 signaling might be an effective non-hormonal and non-opioid approach to treating endometriosis.

Endometriosis is a painful gynecological inflammatory disease that affects 10% of women ( 1 ) and transgender men ( 2 ) with annual health care and economic cost estimates ranging from $78–119 billion annually in the US alone ( 3 , 4 ). Histologically, endometriotic lesions are distinguished by the presence of endometrial-type epithelial glands and stroma, frequently accompanied by hemosiderin-laden macrophages, and fibrosis. Additional components such as blood vessels, nerve fibers, smooth muscle, and immune cell infiltrates are also commonly observed. Clinically, patients often suffer from debilitating and chronic pelvic or abdominal pain, a hallmark feature of the disease. Although hormonal therapies and nonsteroidal anti-inflammatory drugs (NSAIDs) offer symptom relief for some individuals, they are ineffective for a substantial fraction of patients and their use is limited by adverse effects and contraindications in patients with comorbid conditions. Surgical excision remains a standard intervention; however, recurrence of both disease and pain is frequent ( 5 , 6 ). This suboptimal effectiveness of current management strategies has been linked to increased reliance on opioids, with endometriosis patients demonstrating elevated risks of long-term opioid use, dependence, and overdose ( 7 ). Altogether, this highlights the need for novel therapeutic targets and medical interventions that can provide sustained benefits and reduce pain. Nociceptors are specialized peripheral somatosensory neurons that respond to noxious and/or injurious stimuli leading to pain ( 8 – 10 ). Upon activation, the peptidergic nociceptors subpopulation releases neuropeptides such as calcitonin gene-related peptide (CGRP) that signal through the calcitonin receptor-like receptor (CALCRL) in complex with receptor activity modifying protein 1 (RAMP1). The release of these neuropeptides by nociceptors shapes immune cell function in the innervated tissue in a context-dependent manner. This process is known as neuroimmune communication ( 8 – 10 ). We have previously demonstrated a role for neuroimmune communication during endometriosis ( 11 ). We found a nociceptor to macrophage communication via CGRP/RAMP1 signaling that polarizes macrophages to a pro-endometriosis phenotype, resulting in pro-endometriosis macrophages (PEMs). In vitro , we found that CGRP-induced release of mediators by PEMs leads to endometrial epithelial (endo-epi) cell growth and drives lesion formation and pain ( 11 ). However, the key mediators involved in this PEM-induced endometrial cell growth were unknown. Interleukin-33 (IL-33) is a member of the IL-1 family and the sole agonist of the previous orphan receptor ST2 ( 12 ). Both IL-33 and ST2 are expressed by virtually all immune cells ( 13 , 14 ). Seminal studies demonstrated that IL-33 induces pain and exacerbates disease burden, either by directly activating nociceptors or by potentiating the release of inflammatory stimulus ( 15 – 23 ). Specifically for endometriosis, genome-wide association studies (GWAS) show that variants proximal to IL-33 are associated with endometriosis risk ( 24 , 25 ). In mice, previous work has demonstrated that IL-33 contributes to the development of endometriosis lesions ( 26 , 27 ). But those studies used exogenous administration of IL-33 at high doses and/or surgery to induce endometriosis. Moreover, the key molecules involved in regulating IL-33 production were also not addressed. Surgery in the abdominal cavity changes the phenotype of peritoneal cells, confounding analysis of the pathophysiology of endometriosis ( 28 ). Additionally, exogenous administration of IL-33 at doses as low as 0.4 μg induces eosinophilia and production of Th2 cytokines ( 29 ), which again might be confound understanding the role of a cytokine in a disease model. Here, using unbiased approaches such as scRNAseq and bulk RNAseq, we discovered that nociceptor-derived CGRP induces the production of IL-33 by PEMs that is key for lesion growth and pain during endometriosis in mice. In agreement, our single cell disease relevance score (scDRS) that integrates human scRNAseq annotations with GWAS ( 30 – 32 ), identified IL-33 expression in macrophages as associated with endometriosis risk. We also provide evidence that suggests a dual role for IL-33 in endometriosis, in which, it is initially required for lesion formation and later for lesion maintenance only. Furthermore, we show evidence suggesting that IL-33 mediates endometriosis-associated pain, via a mechanism involving a direct and potent activation of ST2 + nociceptors. Therefore, targeting IL-33/ST2 signaling might be an effective way to treat endometriosis and associated pain.