Preferential innervation of endometriosis by hyperexcitable Ret/GFRα1+ nociceptors associates with target GDNF and clinical pain

Endometriosis is a prevalent condition characterized by chronic pelvic pain that is frequently 24 refractory to treatment. While the mechanisms underlying this pain remain poorly defined, clinical evidence often indicates that lesion innervation, but not disease stage (e.g. number and 26 depth of lesions), correlate with pelvic pain severity. However, characterization of lesion- innervating neurons is incomplete, revealing an opportunity to identify novel, disease-modifying 28 therapeutics. Here, we coupled functional analyses of lesion-innervating neurons in a mouse model with concurrent identification and characterization of lesion-innervating neurons from 30 pain-phenotyped endometriosis patients. Following the confirmation of abdominal-directed pain- like behaviors in the mouse model, electrophysiological analysis revealed that lesion-innervating 32 dorsal root ganglion (DRG) neurons are hyperexcitable compared to matched controls. These neurons are predominantly small-diameter and bind Isolectin B4, an established marker of the 34 GDNF Family Ligand receptor, Ret. GDNF is concentrated within the stromal layer of both mouse and human lesions, adjacent to axons expressing the GDNF co-receptor, GFRα1. 36 Critically, clinical pain correlates with lesion GDNF level, axonal density, and neuronal GFRα1 levels. These data provide evidence that endometrial lesions may recruit the Ret-positive 38 subpopulation of nociceptors where they become sensitized and increase patient pain. 40 42 44 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 3

46 Endometriosis is a common disease characterized by chronic pelvic pain1-6 which impacts approximately 70% of patients, 37% of which report severe symptoms7,8. The pain is resistant to 48 many analgesics including opioids and anti-inflammatory agents9-12. Hormonal therapies are often implemented and reduce endometriosis-associated pain in a subset of patients13,14. 50 However, 8-40% of patients do not respond to hormonal therapies, and these therapies are often stopped due to side effects9,13, raising the need for new treatment options with improved 52 efficacy and reduced side effect burden. The disease is defined by ectopic growth of endometrial-like tissue forming lesions within the abdominal cavity, most commonly on the 54 peritoneal wall7,20-22. Laparoscopic surgery to identify and resect lesions reduces pain-related symptoms in about 65% of patients. However, lesions frequently recur (28% at 18 months, 40% 56 at 9 nine years), necessitating repeat procedures9,15,16. The chronic nature of endometriosis- associated pelvic pain leads to healthcare costs of 69.4 billion US dollars/year. The high impact 58 on the patients’ quality of life highlights the need for better understanding of the mechanisms underlying endometriosis to improve treatments, particularly those that target the pain of 60 endometriosis6,17,18. Endometriosis is a complex inflammatory, and endocrine-dependent disease composed of 62 multiple interacting cellular compartments, including ectopic endometrial cells (gland-associated epithelial and stromal cells), immune populations, vasculature, and innervating axons from the 64 dorsal root ganglia (DRG) and sympathetic ganglia19-23. Histologically, lesions are defined by endometrial glands containing an epithelial layer surrounded by a dense stromal cell layer, 66 resembling glands in the eutopic endometrium24. While substantial research has focused on hormonal and immune contributions to disease development, mechanisms underlying the 68 patient pain experience remain poorly defined25-30. Clinical endometriosis staging describes the severity of the disease based on lesion number, depth, and anatomical distribution31. Despite 70 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 4 clear presentation differences between patients, disease stage does not reliably predict pain severity32-34. This disconnect suggests that additional factors contribute to the pain of 72 endometriosis. DRG sensory neurons include nociceptors that transmit noxious peripheral signals to the central 74 nervous system, including endocrine and inflammatory signals in tissue35-37. Clinical studies suggest that endometrial lesions are hyper-innervated relative to normal peritoneum22,38,39. 76 Critically, the density of innervation correlates to patient pain39-41 and is reduced by hormonal treatment used to manage the pain42. Surgical resection of lesions reduces pain transiently 78 before lesions recur13,43,44, linking symptoms to innervated ectopic tissue. Further, ablation of the uterosacral-nerve can reduce patient pain45. However, these observations remain largely 80 correlative and do not functionally assess the neurons that innervate lesions since these tissues are not accessible in the clinic. As a result, the excitability, molecular identity, and recruitment 82 mechanisms of lesion-innervating neurons remain poorly defined. Recent advances targeting DRG neurons to reduce pain in other conditions highlight the importance of characterizing the 84 sensory neuron component of this disease46. To address current limitations, preclinical mouse models of endometriosis have been developed 86 to enable mechanistic investigation of the disease47-49. Although mice do not menstruate,

of endometrial tissue into the peritoneal cavity results in the formation of ectopic 88 lesions with key features of the human disease, including implantation site and histological similarities47. In these models, ablation of nociceptors reduces pain-like behaviors, supporting a 90 functional role for sensory neurons in endometriosis-associated pain30. Given that lesions are ectopic tissues that become innervated during or after formation, it is possible that processes 92 that drive DRG sensory neuron innervation could contribute to endometriosis pain, and possibly the disease itself30. Innervation of peripheral tissues is a regulated process in which axons are 94 guided by local molecular cues50-52. During development, target-released glial cell line-derived (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 5 neurotrophic factor (GDNF) promotes the growth of sensory neurons that express the Ret 96 receptor tyrosine kinase and the GDNF co-receptor GFRα153-57. Further, in the context of injury, GDNF/GFRα1/Ret signaling in nociceptors contributes to pain-like behaviors58,59. Given these 98 dual roles in target innervation and modulation of neuronal excitability, it is possible that this growth factor system is mechanistically related to disease progression and pain severity in 100 endometriosis. In conjunction with the previously described presence of GDNF and other growth factors in endometriosis60,61, this signaling system is poised to be disease modifying. 102 In this study, we investigate the DRG neurons which innervate endometriosis lesions. In the mouse model, we find that lesion-innervating DRG neurons are small-diameter and bind IB4, a 104 marker for the GDNF Family Ligand (GFL) receptor Ret, consistent with their identification as nociceptors. We show that these neurons are hyperexcitable compared to neurons innervating 106 the adjacent peritoneal wall. In patient lesions, GDNF is expressed in the stromal layer of endometrial glands nearby innervating axons that are frequently GFRα1-positive. Finally, we 108 find that the density of innervation and abundance of GDNF and GFRα1 positively correlate with patient-reported pain. These data suggest novel, non-hormonal targets for the management of 110 endometriosis pain. 112

Induction of endometriosis in a mouse model drives chronic abdominal pain-like 114 behaviors. To investigate mechanistic changes due to endometriosis, we first adopted a preclinical mouse model47. In this validated model, eutopic endometrium of a donor animal is 116 injected into the peritoneal cavity of a host animal (Endo). Littermate control animals were injected with media without tissue (Sham). The disease model was allowed to develop over 118 eight weeks (Fig. 1A) which resulted in large, vesicle-like lesions that frequently formed on the (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 6 ventral peritoneal wall, similar to human vesicular endometrial-like lesions31,62-64 (Fig. 1B). 120 Histologic analysis of mouse and human lesions demonstrate that samples from both species present with endometrial-like glands (G) enveloped by a dense stromal cell compartment (Str.), 122 immediately adjacent to the gland-epithelial layer. Inflammation/fibrosis (Infl.) is also present at the border of the gland-related stroma and in nearby tissue, as expected60 (Fig. 1B). In the 124 animal model, we found during careful gross postmortem dissection that ~78.6% of Endo animals formed visually detectable lesions at an average of ~2 lesions/animal. (Fig. 1C). Endo 126 animals were apparently healthy and indistinguishable from Shams with no difference in weight over the development of the model (Fig. 1D). Behavioral assessments at five weeks and eight 128 weeks post induction revealed abdominal mechanical hypersensitivity of Endo animals compared to Shams (Fig. 1E-F), as previously reported47. Spontaneous pain-like behaviors also 130 emerged in Endo animals including writhing65 and “abdominal squashing” characterized by pressing of the abdomen against the floor grate47 (Fig. 1G-H). No difference in grooming/licking 132 directed to the abdomen was found between groups (Fig. 1I). Given the clinical prevalence of dysmenorrhea16, we grouped pain-like behaviors across the mouse estrous cycle determined by 134 visual inspection of the external genitalia66. We found no effects of estrous phase on mechanical sensitivity but observed an increase in the number of “abdominal squashing” bouts in Endo 136 animals in metestrus relative to those in diestrus at eight weeks post transfer. No behavioral differences were detected across estrous phases in Sham animals, although we are not 138 sufficiently powered to detect changes across all phases (Supp. Fig. 1). Together, these data indicate that the mouse endometriosis model recapitulates key aspects of the human pain 140 phenotype, including spontaneous behaviors, and might therefore provide insights into the human disease. 142 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 7 Electrophysiological properties of DRG sensory neurons innervating peritoneal lesions 144 in mouse. Given the clinical data suggesting that innervation density in lesions is related to patient-reported pain39-41, and because the lesions contain inflammatory and hormonal factors 146 that can sensitize neurons67, we sought to functionally characterize lesion-innervating neurons. At the conclusion of behavioral testing, we retrogradely labeled DRG neurons that innervate 148 lesions by micro-injecting WGA into peritoneal lesions (Lesion). As a control, we injected WGA into the wall of the peritoneum where lesions were absent (Endo Wall; Fig. 2A-B). In addition, 150 we injected WGA into the peritoneal wall of littermate sham animals at a location where lesions often form (Sham Wall). Animals were allowed to recover for three to seven days to allow 152 sufficient time for retrograde transport of the dye to the DRG cell bodies of neurons innervating the injection site. WGA injected into lesions remains in lesions without spreading to adjacent 154 non-lesion peritoneum (Supp. Fig. 2). Concurrently, we dissected and dissociated DRGs for in vitro characterization of neuronal excitability by patch clamp electrophysiology of WGA-labeled 156 neurons (Fig. 2C). We restricted electrophysiology analyses to small-diameter neurons (≤30 µm, putative nociceptors68-70; Fig. 2D). We found that the resting membrane potential (RMP) of 158 Lesion-innervating neurons was significantly more depolarized at rest compared to neurons innervating the Sham Wall (Fig. 2E). Interestingly, Endo Wall-innervating neurons showed a 160 trend toward being depolarized relative to Sham Wall-innervating neurons, although the differences were not statistically significant (Fig. 2E). Among the lesion-innervating neurons, we 162 observed several instances of spontaneous firing at rest, while this was never seen in the controls (Fig. 2F). We found no difference in the incidence of single/repetitive firing of Lesion-164 innervating neurons compared to control neurons (Fig. 2G). For experiments evaluating depolarization-evoked excitability, all neurons were held at -60 mV. Consecutive step injections 166 revealed that the rheobase (minimum amount of current necessary to fire an action potential (AP)) of Lesion-innervating neurons was significantly lower compared to control neurons (Fig. 168 2H, representative image Fig. 2H′). Evaluation of AP dynamics revealed that the peak amplitude (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 8 of the AP was significantly reduced in Lesion-innervating neurons relative to control neurons 170 (Fig. 2I, representative image Fig. 2I′). Additional electrophysiological properties were also tested with no significant differences in Lesion-innervating neurons compared to Sham Wall-172 innervating neurons (Suppl. Fig. 2). These results suggest that Lesion-innervating neurons are more excitable and better poised to readily fire APs relative to neurons that innervate the 174 peritoneal wall. During recordings it became apparent that the number of small-diameter neurons in the Lesion-176 innervating group was greater than the number of small-diameter neurons in the peritoneal wall- innervating groups. While electrophysiological properties were all measured in small-diameter 178 neurons (Fig. 2), analysis of all WGA-labeled neurons in culture revealed an overall smaller average diameter of Lesion-innervating neurons compared to control neurons (Fig. 3A). Further, 180 the proportion of all neurons in culture that were small-diameter (≤30 µm) among Lesion- innervating neurons was significantly greater compared to peritoneal wall-innervating groups 182 (Fig. 3B). Given these results, we hypothesized that hyperexcitable lesion-innervating neurons represent a distinct subpopulation of putative nociceptors. 184 Lesions are innervated by small-diameter, IB4/Ret positive neurons. To determine the 186 identity of WGA labeled neurons, we performed immunohistochemistry on DRG tissue from Sham Wall (Fig. 4A-A′′) and Lesion (Fig. 4B-B′′) traced animals. We found that micro-injections 188 into these locations resulted in WGA labeling in lower thoracic (T11-13) and upper lumbar (L1-2) DRG levels, consistent with previously mapped innervation patterns71,72. Isolectin B4 (IB4)-190 binding is classically used to identify a subset of DRG neurons that are small-diameter and Ret positive, indicating responsiveness to GDNF family ligands54,59,73. Considering our in vitro 192 evidence that lesion-innervating neurons are small-diameter (Fig. 3), we tested whether DRG neurons retrogradely labeled from lesions also bound IB4. The proportion of WGA-labeled 194 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 9 neurons did not differ between groups (Fig. 4C). However, the average diameter of Lesion- innervating neurons was significantly smaller than the average diameter of Sham Wall-196 innervating neurons (Fig. 4D). A distribution analysis of diameters binned by size demonstrates that Lesion-innervating neurons skew towards small-diameter, putative nociceptors, whereas 198 Sham Wall-innervating neurons have a more widespread size distribution (Fig. 4E). Finally, consistent with the differences in sizes, quantification of IB4-positive neurons reveals that 200 Lesion-innervating neurons bind IB4 significantly more frequently than Sham Wall-innervating neurons (Fig. 4F). Together, these data support the in vitro observation that Lesion-innervating 202 DRG neurons are predominantly small-diameter neurons belonging to the Ret expressing subpopulation of nociceptors and suggest that GFL receptors (GFRs) and their associated 204 ligands may play a role in the innervation of endometriosis lesions. We and others have previously shown, and confirm here using a Ret-reporter mouse line74, that 206 IB4-labeled neurons almost exclusively also express Ret, particularly in the small-diameter population (Supp. Fig. 3)54,75. To determine if the Ret+ population of small-diameter DRG 208 neurons are functionally distinct from Ret- small-diameter neurons in naïve animals, we performed patch clamp electrophysiology of small-diameter DRG neurons in the Ret-reporter 210 mouse. We found that Ret+ neurons have a higher rheobase (less excitable) than Ret- neurons. Action potential waveform analysis found that Ret+ neurons have a lower amplitude and peak 212 compared to Ret- neurons (Supp. Fig. 3). Together, these data demonstrate that although small-diameter Ret+ DRG neurons are inherently less excitable than Ret- DRG neurons, when 214 this population innervates lesions they show increased excitability compared to small-diameter neurons that innervate the peritoneal wall and are often Ret-. Overall, our data is consistent with 216 a potential role for GFL/Ret signaling in lesion innervation that might contribute to sensitization and pain in this endometriosis mouse model. 218 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 10 Endometriosis patient-reported pain is correlated with stromal GDNF levels and neuronal 220 GFRα1 levels. To explore the potential role of GDNF/Ret signaling in clinical endometriosis, we recruited patients scheduled for minimally invasive gynecological surgery to diagnose and 222 resect endometriosis lesions. During a study visit prior to scheduled surgery, patients completed the Endometriosis Health Profile-30 Pain Scale (EHP-30) and reported their endometriosis-224 related pelvic pain intensity over the last 30 days using a scale ranging from 0 (no pain) to 10 (worst pain imaginable). Fourteen participants who had a self-reported dynamic range of pain 226 and confirmed endometriosis lesions that were removed during the surgery were included in this study. Consistent with some prior reports, clinically assigned disease stage was not related to 228 patient reported pain (Table 1)32-34. We sought to identify molecular markers that correlate with patient pain. We first compared 230 lesions from patients who reported high/severe pelvic pain (pain ratings = 7-9; N=6) to lesions from patients reporting mild/moderate pain (pain ratings = 3-6; N=6) by performing unbiased 232 bulk proteomics assaying >5,000 proteins. We found multiple growth factors enriched in lesions from high/severe pain participants including GDNF, NOTCH1, ARTN, NT3, and BDNF (Fig. 5A). 234 Although these differences are not statistically significant, a principal component analysis reveals a cluster of high-pain samples with high expression of GDNF (Supp. Fig. 4). 236 Interestingly, binning patients into groups of mild (pain intensity = 3-4; N=2), moderate (intensity = 5-6; N=4), high (intensity = 7; N=3), and severe pain (intensity = 8-9; N=3) reveals a clear 238 trend where GDNF levels are positively associated with clinical pain scores (Fig. 5B). Analysis of publicly available endometriosis lesion datasets indicates that GDNF is likely expressed in 240 stromal cells within endometriosis lesions although smooth muscle cells and B cells also express this gene76. We leveraged our proteomic dataset to assess relationships between 242 GDNF and stromal cell markers (COL4A1, CD10, ESR1 and PGR), some of which are used clinically to diagnose endometriosis77,78. We found that the levels of most stromal factors are 244 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 11 tightly associated with GDNF, but B cell and smooth muscle cell markers do not correlate with GDNF levels (Supp. Fig. 4). Normalization of GDNF levels to COL4A1 (pan-stromal cell marker, 246 Fig. 5C) as well as CD10 (diagnostic endometriosis stroma marker; Fig. 5D) suggests that the association between pelvic pain and GDNF levels might be due to stromal cell expression of 248 GDNF. Further, we found by immunohistochemistry that GDNF is restricted to endometrial glands, most commonly in the stromal cell compartment of patient lesions (Fig. 5E) and of the 250 mouse model lesions (Supp. Fig. 5). Finally, pathway enrichment analysis implicates both stromal cell expansion and inflammatory pathways to be enriched in lesions from patients 252 reporting high/severe pain (Supp. Fig. 4). Together, these data support a possible relationship between lesion GDNF and endometriosis pain and strongly associate GDNF with the lesion 254 stromal cell compartment. Publicly available human DRG datasets confirm that GFRα1 is highly expressed in subsets of 256 neurons consistent with nociceptors79. In our proteomic dataset, we found that GFRα1 expression, but not GFRα2/3, correlates with the abundance of Peripherin, a marker for the 258 peripheral nervous system including both sensory and sympathetic axons (Supp. Fig. 4). Immunohistochemical analysis confirmed a strong correlation between GFRα1 and Peripherin in 260 axons innervating endometrial-associated stroma and glands (Fig. 5F-G). Associations in the proteomics dataset reveal that GFRα1 levels correlate with TRPV1 levels (sensory axon 262 specific) but not TH levels (sympathetic axon specific)80,81. This suggests that the GFRα1+ axons innervating lesions are sensory neurons (Supp. Fig. 4) which is consistent with prior 264 reports that demonstrate a reduction in sympathetic axons in lesions compared to healthy peritoneal tissue82. These data, combined with the expression of GDNF in lesions, inspired 266 further testing and validation to determine if a GDNF-GFRα1 axis is modified in lesions based on pain score. 268 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 12 Patient endometriosis pain is related to levels of GDNF-GFRα1 in lesions. After identifying the putative sources of GDNF and GFRα1 in lesions, we next sought to validate whether 270 expression of GDNF and GFRα1 in endometriosis lesions is related to clinical pain scores. Given the apparent containment of GDNF to the stromal layer of endometrial glands, we 272 restricted our histological quantification to endometrial glands. We found a strong trend between cell density-normalized GDNF expression level mean fluorescence intensity (MFI) and clinical 274 pain, consistent with the proteomic finding that GDNF levels are related to pain intensity (Fig. 6A-C), though this was not statistically significant. Importantly, the size of the gland and the 276 density of cells in the region analyzed did not correlate with pain (Supp. Table 1). Analysis of Peripherin in endometrial glands supports prior reports that patient pain is related to lesion 278 axonal density39-41 (Fig. 6D-F), here localized specifically to glands. Robust innervation density is also found in the lesions dissected from the animal model, demonstrating a similar pattern of 280 innervation as human (Supp. Fig. 5). Finally, we quantified the amount of GFRα1 fluorescence and found that GFRα1 also displays a strong trend to be correlated with patient reported pain 282 (Fig. 6G-I). GDNF, Peripherin and GFRα1 MFI were converted to a composite z-score to analyze GDNF-GFRα1 signaling in our dataset83. This score was significantly correlated to 284 clinical pain (Fig. 6J) and displayed a positive trend to associate with patient’s EHP-30 pain score (Fig. 6K). Together, these data provide correlative evidence in human samples indicating 286 that stromal GDNF and axonal GFRα1 expression levels are related to clinical pain intensity supporting the mechanistic findings in the animal model. 288

290 We report data suggesting distinct properties and identities of DRG neurons that innervate endometrial-like lesions in both a mouse model of the disease and pain-phenotyped participants 292 who have endometriosis (Fig. 7). We confirm that the mouse model recapitulates prominent (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 13 features of the pain experienced by humans including spontaneous, pelvic-directed pain-like 294 behaviors (Fig. 1)47. We demonstrate, for the first time, that mouse endometrial-like lesions are innervated by DRG neurons that are hyperexcitable compared to control neurons (Fig. 2) and 296 which we identified to be the Ret/IB4 subpopulation of small-diameter nociceptors (Figs. 3-4)54. These findings in the mouse model prompted us to examine the innervation of endometrial 298 lesions from human patients who self-reported their pelvic pain. We identified GFL signaling components in human lesions (Fig. 5) and revealed that levels of GDNF, and its co-receptor 300 GFRα1 on lesion-innervating axons, were predictive of patient-reported pain (Fig. 6). These data are consistent with the hypothesis that GDNF expressed in lesions drives neuronal cell-302 type specific innervation of lesions followed by neuronal sensitization, increased excitability and pelvic pain. 304 Human endometriosis lesions consist of various tissues/cell types including implantation site tissue (peritoneum), endometrial-like tissue (epithelial cells, stromal cells), vascular tissues, 306 immune cells, and innervating axons19,20. The molecular landscape in endometrial-like lesions is consequently complex20,27-29,84-86. We reasoned that evaluating the proteome of lesions from 308 patients who reported a dynamic range of pelvic pain might reveal novel therapeutic strategies targeting pain. Our data support GDNF as a strong candidate that may link patient pain and 310 lesion biology to neuronal recruitment. It is well established that GDNF-GFRα1 signaling regulates the targeting and survival of specific sensory neuron populations during 312 development54-56. It is therefore possible that similar mechanisms are reactivated in endometriosis, similar to neurotrophic factor-mediated targeting of axons to tumors in certain 314 cancers87,88. According to publicly available human datasets, GDNF is expressed at low levels in the stromal cell compartment of the eutopic endometrium in the uterus (putative source of 316 endometrial cells that form lesions)76. It is therefore possible that after/during ectopic attachment of endometrial cells to the peritoneal wall, GDNF is upregulated60 (Fig. 5) and capable of 318 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 14 increasing GFRα1-positive sensory neuron innervation (Supp. Fig. 5). The identification of this neuronal subtype in human lesions is consistent with the mouse data, which revealed that DRG 320 neurons innervating lesions are often small-diameter and IB4-positive, designating them as Ret+ nociceptors (Fig. 4 and Supp. Fig. 3). Functional experiments to reduce GDNF in lesion-322 fated stromal cells would therefore be a valuable tool to determine the requirement for GDNF to drive pain-like behaviors, innervation, and lesion formation, and would provide further rationale 324 for a therapeutic approach targeting GFL/GFRα1/Ret signaling. The prevalence of this growth factor signaling axis in endometrial lesions in mouse endometrial-326 like lesions prompted us to examine the relationship between GFRα1 and GDNF expression in human endometriosis lesions and clinical pain (Fig. 6). In human endometriosis lesions, we find 328 that the level of GDNF, putatively expressed by stromal cells of the endometrial gland, and the level of the GFRα1 expressed by DRG neurons that innervate lesions, are positively correlated 330 with clinical pain intensity and show a positive correlation trend with endometriosis health profile. These data support further examination of the role of the GDNF/GFRα1/Ret signaling 332 pathway in endometriosis pain. Together, these data support the hypothesis that the level of GFLs in lesions impact patient pain by increasing recruitment of nociceptors to lesions. 334 In addition to GDNF, we also identified several other neurotrophic factors that are non- significantly enriched in lesions from patients who reported more pain. BDNF and NT3 are 336 notable due to their roles in endometriosis pathogenesis89,90 while ARTN is of interest because it preferentially binds to the GFL co-receptor GFRα3 which is expressed on a different subset of 338 DRG sensory neurons than GFRα153,56 (Fig. 5). It is possible that ARTN/GFRα3 signaling represents a parallel pathway to GDNF/GFRα1 in regulating lesion innervation and pain. 340 Our results are consistent with prior studies that reported correlations between lesion innervation and patient pain39-41. Our study replicates this finding and extends it by identifying a 342 specific subpopulation of small-diameter GFRα1-positive nociceptors (human; Figs. 6) and (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 15 corresponding IB4-positive nociceptors (mouse; Fig. 4) that innervates lesions (Figs. 2, 5, 6) 344 suggesting selective recruitment of nociceptors that express Ret. While this identifies a neuronal type in lesions, the functional properties of these neurons have previously not been explored. 346 We demonstrate in the mouse model that lesion-innervating DRG neurons are hyperexcitable compared to control neurons (Fig. 2). There are two notable explanations for the 348 hyperexcitability of lesion-innervating DRG neurons. The first is that the inherent properties of nociceptors that innervate lesions render these neurons more excitable compared with control 350 neurons. Since IB4 binding in DRG neurons specifically marks the subpopulation of Ret-positive nociceptors, we used a Ret-reporter mouse to compare the excitability of Ret+ neurons to Ret- 352 neurons to determine if there are inherent differences in excitability between these populations74. We found that Ret+ neurons have a higher rheobase (less excitable) compared to 354 Ret- neurons (Supp. Fig. 3), consistent with a prior report which investigated the excitability of IB4+ vs. IB4- small-diameter neurons91. In contrast to these findings, lesion-innervating neurons 356 that are frequently IB4+ (Fig. 4), have a significantly lower rheobase (more excitable) compared to Wall-innervating neurons. These data therefore favor a second interpretation, that lesion-358 innervating neurons become sensitized by the highly inflamed microenvironment of lesions after recruitment. This sensitization might be through GDNF directly58 and/or other pathways such as 360 tumor necrosis factor (TNF), interleukin-6 (IL-6) and interferon (IFN) signaling which are enriched in lesions from patients who report higher pain (Supp. Fig. 4). Although 362 inflammatory/sensitizing factors would increase the excitability of these neurons, they would not change their identity (i.e. small-diameter nociceptors), an effect that may be evident in AP 364 waveform similarities between Ret+ and Lesion-innervating neurons (Supp. Fig. 3 and Fig. 2, respectively). We therefore propose a model in which GDNF in lesions dictates the recruitment 366 of GFRα1-positive nociceptors, which are subsequently sensitized by inflammatory mediators known to be present in lesions, driving clinical pain. 368 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 16 Together, the identification of GDNF and GFRα1 positive axons in human lesions suggests that therapeutics that either 1) target excitability of sensory neurons46, and specifically the Ret+ 370 subpopulation of nociceptors, or 2) block the innervation of lesions by Ret+ nociceptors, may represent viable strategies for addressing endometriosis-associated pain. Indeed, targeting 372 GDNF signaling offers a distinct advantage as a disease modifying therapy particularly to reduce post-surgical recurrence of endometriosis pain by potentially preventing hyper-374 innervation of lesions by this nociceptor population. Although functional studies are required to validate this approach, clinical trials utilizing similar growth factor ligand blockade for other 376 diseases strengthen the feasibility of the strategy92. There are several limitations to our study. Animal models of endometriosis are useful tools but 378 this disease cannot be exactly recapitulated in this model organism because the tested mice (C57Bl/6) do not menstruate, and they have a reproductive tract which does not allow for the 380 prevailing hypothesis of the source of endometrial-like cells in the peritoneum in endometriosis7,93. It is interesting that we found more pain-like behaviors in Endo animals in 382 metestrus compared to diestrus because similar cycle-dependent pain is observed in many patients (Supp. Fig. 1)14. Further, we found increased levels of components of sex-hormone 384 signaling pathways in lesions from patients with high pain scores in our human proteomic dataset (Supp. Fig. 4). Investigations which explore cycle-related differences may better 386 support the relevance of the mouse model to the human disease and encourage mechanistic testing of sex-hormone levels in relation to pain-like behaviors. While lesions that form in the 388 mouse are consistent with some types of lesions seen in human endometriosis (Fig. 1), they do not represent all histologic forms of the disease62,63. Given the relatively early timeframe of 390 animal experiments compared to human disease, which is often not diagnosed for years after symptom onset7, the mouse model lesions are most concordant with superficial peritoneal 392 endometriosis. This aligns with human endometriosis disease trajectory where superficial (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 17 peritoneal endometriosis is the dominant histology in adolescents (i.e. early in the disease)94,95. 394 Regardless, GDNF expression in the gland-stromal compartment and innervation of endometrial glands in both species were similar (Fig. 5, Supp. Fig. 5). Another limitation is that the scope of 396 this study is primarily restricted to lesions and lesion-innervating neurons, despite data indicating that other anatomical sites (e.g. eutopic endometrium, peritoneal fluid) are also 398 affected by the disease26,30,76,96-100. While we restricted our analyses to lesions and innervating neurons, a strength of this study is that we observed that peritoneal wall innervating neurons 400 from Endo animals exhibited a trend toward a more depolarized resting membrane potential compared to Sham littermate controls (Fig. 2). This finding is consistent with prior clinical data 402 that found higher transcript levels of nociceptor-associated ion channels (TRPV1, TRPA1) in the peritoneal wall of patients with endometriosis compared to healthy controls98. These data 404 suggest that neuronal sensitization may not be restricted to lesion-innervating populations but may instead reflect broader changes to the peritoneum and/or to the peripheral nervous system. 406 Such widespread alterations could contribute to the diffuse and persistent nature of pelvic pain in endometriosis43,44,101 and highlight the importance of considering both lesion-specific and 408 system-level mechanisms. Finally, the clinical sample size evaluated in this dataset (N=12 proteomics, N=6 IHC) is limited to pilot study observations. Endometriosis is a highly 410 heterogenous disease and a larger dataset is necessary to confirm the correlation between GDNF expression level and clinical pain, and to control for factors which might alter GDNF 412 levels such as medications, menstrual cycle and age16. The results of this study are first-in-kind linking functional data from an animal model of the 414 disease with neuronal identification in clinical, pain-defined samples. The data indicate that the Ret-positive subpopulation of nociceptors might be actively recruited into lesions through a 416 stromal cell GDNF and sensory neuron GFRα1 signaling axis. We demonstrate that the neurons innervating endometriosis lesions in mice are highly excitable and are a distinct neuronal 418 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 18 population of nociceptors. In addition, higher levels of GDNF and GFRα1 are associated with higher levels of patient-reported pain, connecting the observed molecular features of the lesion 420 microenvironment to neuronal function and clinical symptoms. Together, these findings define a mechanistic framework for how lesion innervating DRG neurons contribute to endometriosis-422 associated pain and identify lesion GDNF and GFRα1-positive nociceptors as tractable targets for therapeutic intervention. 424 426 428 430 432 434 436 438 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 19

440 1 Astruc, A. et al. Advanced Insights into Human Uterine Innervation: Implications for Endometriosis and Pelvic Pain. J Clin Med 13 (2024). 442 https://doi.org/10.3390/jcm13051433 2 Gruber, T. M. & Mechsner, S. Pathogenesis of Endometriosis: The Origin of Pain and 444 Subfertility. Cells 10, 1381 (2021). 3 Ellis, K., Munro, D. & Clarke, J. Endometriosis Is Undervalued: A Call to Action. Front 446 Glob Womens Health 3, 902371 (2022). https://doi.org/10.3389/fgwh.2022.902371 4 As-Sanie, S. et al. Assessing research gaps and unmet needs in endometriosis. Am J 448 Obstet Gynecol 221, 86-94 (2019). https://doi.org/10.1016/j.ajog.2019.02.033 5 Nunez-Badinez, P. et al. Preclinical models of endometriosis and interstitial 450 cystitis/bladder pain syndrome: an Innovative Medicines Initiative-PainCare initiative to improve their value for translational research in pelvic pain. Pain 162, 2349-2365 (2021). 452 https://doi.org/10.1097/j.pain.0000000000002248 6 Maddern, J., Grundy, L., Castro, J. & Brierley, S. M. Pain in Endometriosis. Frontiers in 454 cellular neuroscience 14, 590823 (2020). https://doi.org/10.3389/fncel.2020.590823 7 Zondervan, K. T., Becker, C. M. & Missmer, S. A. Endometriosis. The New England 456 journal of medicine 382, 1244-1256 (2020). https://doi.org/10.1056/NEJMra1810764 8 Sinaii, N. et al. Differences in characteristics among 1,000 women with endometriosis 458 based on extent of disease. Fertil Steril 89, 538-545 (2008). https://doi.org/10.1016/j.fertnstert.2007.03.069 460 9 Olive David, L. & Schwartz Lisa, B. Endometriosis. New England Journal of Medicine 328, 1759-1769 (1993). https://doi.org/10.1056/NEJM199306173282407 462 10 Lamvu, G. et al. Patterns of Prescription Opioid Use in Women With Endometriosis: Evaluating Prolonged Use, Daily Dose, and Concomitant Use With Benzodiazepines. 464 Obstet Gynecol 133, 1120-1130 (2019). https://doi.org/10.1097/aog.0000000000003267 11 Chiuve, S. E. et al. Chronic opioid use and complication risks in women with 466 endometriosis: A cohort study in US administrative claims. Pharmacoepidemiol Drug Saf 30, 787-796 (2021). https://doi.org/10.1002/pds.5209 468 12 Leyland, N., Estes, S. J., Lessey, B. A., Advincula, A. P . & Taylor, H. S. A Clinician's Guide to the Treatment of Endometriosis with Elagolix. J Womens Health (Larchmt) 30, 470 569-578 (2021). https://doi.org/10.1089/jwh.2019.8096 13 Selçuk, I. & Bozdağ, G. Recurrence of endometriosis; risk factors, mechanisms and 472 biomarkers; review of the literature. J Turk Ger Gynecol Assoc 14, 98-103 (2013). https://doi.org/10.5152/jtgga.2013.52385 474 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 20 14 Committee, A. P. Treatment of pelvic pain associated with endometriosis. Fertility and Sterility 90, S260-S269 (2008). 476 https://doi.org/https://doi.org/10.1016/j.fertnstert.2008.08.057 15 Wright, J., Lotfallah, H., Jones, K. & Lovell, D. A randomized trial of excision versus 478 ablation for mild endometriosis. Fertility and Sterility 83, 1830-1836 (2005). https://doi.org/https://doi.org/10.1016/j.fertnstert.2004.11.066 480 16 Dunselman, G. A. et al. ESHRE guideline: management of women with endometriosis. Hum Reprod 29, 400-412 (2014). https://doi.org/10.1093/humrep/det457 482 17 Soliman, A. M. et al. Health Care Utilization and Costs Associated with Endometriosis Among Women with Medicaid Insurance. J Manag Care Spec Pharm 25, 566-572 484 (2019). https://doi.org/10.18553/jmcp.2019.25.5.566 18 Soliman, A. M., Yang, H., Du, E. X., Kelley, C. & Winkel, C. The direct and indirect costs 486 associated with endometriosis: a systematic literature review. Hum Reprod 31, 712-722 (2016). https://doi.org/10.1093/humrep/dev335 488 19 Imperiale, L., Nisolle, M., Noël, J. C. & Fastrez, M. Three Types of Endometriosis: Pathogenesis, Diagnosis and Treatment. State of the Art. J Clin Med 12 (2023). 490 https://doi.org/10.3390/jcm12030994 20 Yan, J. et al. Single-cell analysis reveals insights into epithelial abnormalities in ovarian 492 endometriosis. Cell Reports 43, 113716 (2024). https://doi.org/https://doi.org/10.1016/j.celrep.2024.113716 494 21 McKinnon, B. D., Bertschi, D., Bersinger, N. A. & Mueller, M. D. Inflammation and nerve fiber interaction in endometriotic pain. Trends in Endocrinology & Metabolism 26, 1-10 496 (2015). https://doi.org/https://doi.org/10.1016/j.tem.2014.10.003 22 Tokushige, N., Markham, R., Russell, P . & Fraser, I. S. Nerve fibres in peritoneal 498 endometriosis. Human Reproduction 21, 3001-3007 (2006). https://doi.org/10.1093/humrep/del260 500 23 Berkley, K. J., Dmitrieva, N., Curtis, K. S. & Papka, R. E. Innervation of ectopic endometrium in a rat model of endometriosis. Proc Natl Acad Sci U S A 101, 11094-502 11098 (2004). https://doi.org/10.1073/pnas.0403663101 24 Kamergorodsky, G. et al. Histologic classification of specimens from women affected by 504 superficial endometriosis, deeply infiltrating endometriosis, and ovarian endometriomas. Fertility and Sterility 92, 2074-2077 (2009). 506 https://doi.org/https://doi.org/10.1016/j.fertnstert.2009.05.086 25 Huhtinen, K. et al. Endometrial and Endometriotic Concentrations of Estrone and 508 Estradiol Are Determined by Local Metabolism Rather than Circulating Levels. The Journal of Clinical Endocrinology & Metabolism 97, 4228-4235 (2012). 510 https://doi.org/10.1210/jc.2012-1154 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 21 26 Mori, T. et al. Local estrogen formation and its regulation in endometriosis. Reproductive 512 Medicine and Biology 18, 305-311 (2019). https://doi.org/https://doi.org/10.1002/rmb2.12285 514 27 Huang, X. et al. Single-cell transcriptome analysis reveals endometrial immune microenvironment in minimal/mild endometriosis. Clinical and Experimental Immunology 516 212, 285-295 (2023). https://doi.org/10.1093/cei/uxad029 28 Tan, Y. et al. Single-cell analysis of endometriosis reveals a coordinated transcriptional 518 programme driving immunotolerance and angiogenesis across eutopic and ectopic tissues. Nature Cell Biology 24, 1306-1318 (2022). https://doi.org/10.1038/s41556-022-520 00961-5 29 Garcia-Alonso, L. et al. Mapping the temporal and spatial dynamics of the human 522 endometrium in vivo and in vitro. Nature genetics 53, 1698-1711 (2021). https://doi.org/10.1038/s41588-021-00972-2 524 30 Fattori, V. et al. Nociceptor-to-macrophage communication through CGRP/RAMP1 signaling drives endometriosis-associated pain and lesion growth in mice. Science 526 Translational Medicine 16, eadk8230 (2024). https://doi.org/doi:10.1126/scitranslmed.adk8230 528 31 Abrao, M. S. et al. AAGL 2021 Endometriosis Classification: An Anatomy-based Surgical Complexity Score. J Minim Invasive Gynecol 28, 1941-1950.e1941 (2021). 530 https://doi.org/10.1016/j.jmig.2021.09.709 32 Balasch, J. et al. Visible and non-visible endometriosis at laparoscopy in fertile and 532 infertile women and in patients with chronic pelvic pain: a prospective study. Hum Reprod 11, 387-391 (1996). https://doi.org/10.1093/humrep/11.2.387 534 33 Bajaj, P., Bajaj, P ., Madsen, H. & Arendt-Nielsen, L. Endometriosis is associated with central sensitization: a psychophysical controlled study. The Journal of Pain 4, 372-380 536 (2003). https://doi.org/https://doi.org/10.1016/S1526-5900(03)00720-X 34 dell'Endometriosi, G. I. p. l. S. Relationship between stage, site and morphological 538 characteristics of pelvic endometriosis and pain. Hum Reprod 16, 2668-2671 (2001). https://doi.org/10.1093/humrep/16.12.2668 540 35 Dourson, A. J. et al. Early Life Nociception is Influenced by Peripheral Growth Hormone Signaling. The Journal of neuroscience : the official journal of the Society for 542 Neuroscience, 4410-4427 (2021). https://doi.org/10.1523/jneurosci.3081-20.2021 36 Dourson, A. J. et al. Macrophage memories of early-life injury drive neonatal nociceptive 544 priming. Cell Reports 43 (2024). https://doi.org/10.1016/j.celrep.2024.114129 37 Fadaka, A. O. et al. The intersection of endocrine signaling and neuroimmune 546 communication regulates muscle inflammation-induced nociception in neonatal mice. Brain Behav Immun (2024). https://doi.org/10.1016/j.bbi.2024.12.148 548 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 22 38 Liang, Y . & Yao, S. Potential role of estrogen in maintaining the imbalanced sympathetic and sensory innervation in endometriosis. Molecular and Cellular Endocrinology 424, 42-550 49 (2016). https://doi.org/https://doi.org/10.1016/j.mce.2016.01.012 39 Morotti, M., Vincent, K., Brawn, J., Zondervan, K. T. & Becker, C. M. Peripheral changes 552 in endometriosis-associated pain. Human Reproduction Update 20, 717-736 (2014). https://doi.org/10.1093/humupd/dmu021 554 40 Mechsner, S. et al. A pilot study to evaluate the clinical relevance of endometriosis- associated nerve fibers in peritoneal endometriotic lesions. Fertil Steril 92, 1856-1861 556 (2009). https://doi.org/10.1016/j.fertnstert.2008.09.006 41 Godin, S. K., Wagner, J., Huang, P . & Bree, D. The role of peripheral nerve signaling in 558 endometriosis. FASEB Bioadv 3, 802-813 (2021). https://doi.org/10.1096/fba.2021- 00063 560 42 Medina, M. G. & Lebovic, D. I. Endometriosis-associated nerve fibers and pain. Acta Obstet Gynecol Scand 88, 968-975 (2009). https://doi.org/10.1080/00016340903176826 562 43 Zanelotti, A. & Decherney, A. H. Surgery and Endometriosis. Clin Obstet Gynecol 60, 477-484 (2017). https://doi.org/10.1097/grf.0000000000000291 564 44 Guo, S.-W. Recurrence of endometriosis and its control. Human Reproduction Update 15, 441-461 (2009). https://doi.org/10.1093/humupd/dmp007 566 45 Lichten, E. M. & Bombard, J. Surgical treatment of primary dysmenorrhea with laparoscopic uterine nerve ablation. J Reprod Med 32, 37-41 (1987). 568 46 Stewart, R. G. et al. Modulation of human dorsal root ganglion neuron firing by the Nav1.8 inhibitor suzetrigine. Proceedings of the National Academy of Sciences 122, 570 e2503570122 (2025). https://doi.org/doi:10.1073/pnas.2503570122 47 Fattori, V. et al. Nonsurgical mouse model of endometriosis-associated pain that 572 responds to clinically active drugs. Pain 161, 1321-1331 (2020). https://doi.org/10.1097/j.pain.0000000000001832 574 48 Maddern, J. et al. A syngeneic inoculation mouse model of endometriosis that develops multiple comorbid visceral and cutaneous pain like behaviours. Pain 163, 1622-1635 576 (2022). https://doi.org/10.1097/j.pain.0000000000002552 49 Tejada, M. A. et al. Rodent Animal Models of Endometriosis-Associated Pain: Unmet 578 Needs and Resources Available for Improving Translational Research in Endometriosis. International journal of molecular sciences 24 (2023). 580 https://doi.org/10.3390/ijms24032422 50 Fitzgerald, M. Developmental biology of inflammatory pain. British journal of anaesthesia 582 75, 177-185 (1995). https://doi.org/10.1093/bja/75.2.177 51 Alvares, D. & Fitzgerald, M. Building blocks of pain: the regulation of key molecules in 584 spinal sensory neurones during development and following peripheral axotomy. Pain 82, S71-S85 (1999). https://doi.org/https://doi.org/10.1016/S0304-3959(99)00140-2 586 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 23 52 Fitzgerald, M. The development of nociceptive circuits. Nature Reviews Neuroscience 6, 507-520 (2005). https://doi.org/10.1038/nrn1701 588 53 Airaksinen, M. S. & Saarma, M. The GDNF family: Signalling, biological functions and therapeutic value. Nature Reviews Neuroscience 3, 383-394 (2002). 590 https://doi.org/10.1038/nrn812 54 Molliver, D. C. et al. IB4-binding DRG neurons switch from NGF to GDNF dependence in 592 early postnatal life. Neuron 19, 849-861 (1997). https://doi.org/10.1016/s0896- 6273(00)80966-6 594 55 Chen, Z. et al. p75 Is Required for the Establishment of Postnatal Sensory Neuron Diversity by Potentiating Ret Signaling. Cell Rep 21, 707-720 (2017). 596 https://doi.org/10.1016/j.celrep.2017.09.037 56 Jankowski, M. P., Miller, L. & Koerber, H. R. Increased Expression of Transcription 598 Factor SRY-box-Containing Gene 11 (Sox11) Enhances Neurite Growth by Regulating Neurotrophic Factor Responsiveness. Neuroscience 382, 93-104 (2018). 600 https://doi.org/10.1016/j.neuroscience.2018.04.037 57 Cacalano, G. et al. GFRalpha1 is an essential receptor component for GDNF in the 602 developing nervous system and kidney. Neuron 21, 53-62 (1998). https://doi.org/10.1016/s0896-6273(00)80514-0 604 58 Queme, L. F., Weyler, A. A., Cohen, E. R., Hudgins, R. C. & Jankowski, M. P. A dual role for peripheral GDNF signaling in nociception and cardiovascular reflexes in the mouse. 606 Proc Natl Acad Sci U S A 117, 698-707 (2020). https://doi.org/10.1073/pnas.1910905116 59 Albers, K. M., Woodbury, C. J., Ritter, A. M., Davis, B. M. & Koerber, H. R. Glial cell-line-608 derived neurotrophic factor expression in skin alters the mechanical sensitivity of cutaneous nociceptors. The Journal of neuroscience : the official journal of the Society 610 for Neuroscience 26, 2981-2990 (2006). https://doi.org/10.1523/jneurosci.4863-05.2006 60 Asally, R., Markham, R. & Manconi, F. The Expression and Cellular Localisation of 612 Neurotrophin and Neural Guidance Molecules in Peritoneal Ectopic Lesions. Molecular neurobiology 56, 4013-4022 (2019). https://doi.org/10.1007/s12035-018-1348-6 614 61 Kasheh Farahani, Z., Taherianfard, M., Naderi, M. M. & Ferrero, H. Assessing Pain Behavioral Responses and Neurotrophic Factors in the Dorsal Root Ganglion, Serum 616 and Peritoneal Fluid in Rat Models of Endometriosis. J Family Reprod Health 14, 259- 268 (2020). https://doi.org/10.18502/jfrh.v14i4.5210 618 62 Laufer, M. R. Identification of Clear Vesicular Lesions of Atypical Endometriosis: A New Technique. Journal of Pediatric and Adolescent Gynecology 10, 169 (1997). 620 https://doi.org/https://doi.org/10.1016/S1083-3188(97)70242-0 63 Justice, T. D. & Nakajima, S. T. All Clear or So We Thought: A Novel Approach to the 622 Diagnosis of Clear Vesicular Endometriosis. Fertility and Sterility 99, S35-S36 (2013). https://doi.org/10.1016/j.fertnstert.2013.01.081 624 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 24 64 Zhang, P. & Wang, G. Progesterone Resistance in Endometriosis: Current Evidence and Putative Mechanisms. International journal of molecular sciences 24 (2023). 626 https://doi.org/10.3390/ijms24086992 65 Gawade, S. P . Acetic acid induced painful endogenous infliction in writhing test on mice. 628 J Pharmacol Pharmacother 3, 348 (2012). https://doi.org/10.4103/0976-500x.103699 66 Byers, S. L., Wiles, M. V., Dunn, S. L. & Taft, R. A. Mouse estrous cycle identification tool 630 and images. PloS one 7, e35538 (2012). https://doi.org/10.1371/journal.pone.0035538 67 Omenge, H. J., Kempa, P. A., Jeong, J. W., So, K. A. & Kim, T. H. Endometriosis-632 associated Pain: Mechanism, Neuroimmune Signature, and Translational Precision Strategies. Endocrine reviews (2026). https://doi.org/10.1210/endrev/bnag019 634 68 McIlvried, L. A. et al. Intrinsic adaptive plasticity in mouse and human sensory neurons. J Gen Physiol 157 (2025). https://doi.org/10.1085/jgp.202313488 636 69 Gold, M. S., Dastmalchi, S. & Levine, J. D. Co-expression of nociceptor properties in dorsal root ganglion neurons from the adult rat in vitro. Neuroscience 71, 265-275 638 (1996). https://doi.org/https://doi.org/10.1016/0306-4522(95)00433-5 70 Han, C. et al. Human Nav1.8: enhanced persistent and ramp currents contribute to 640 distinct firing properties of human DRG neurons. Journal of neurophysiology 113, 3172- 3185 (2015). https://doi.org/10.1152/jn.00113.2015 642 71 Struller, F. et al. Peritoneal innervation: embryology and functional anatomy. Pleura Peritoneum 2, 153-161 (2017). https://doi.org/10.1515/pp-2017-0024 644 72 Takahashi, Y., Chiba, T., Kurokawa, M. & Aoki, Y. Dermatomes and the central organization of dermatomes and body surface regions in the spinal cord dorsal horn in 646 rats. Journal of Comparative Neurology 462, 29-41 (2003). https://doi.org/https://doi.org/10.1002/cne.10669 648 73 Leitner, M. L. et al. Analysis of the retrograde transport of glial cell line-derived neurotrophic factor (GDNF), neurturin, and persephin suggests that in vivo signaling for 650 the GDNF family is GFRalpha coreceptor-specific. The Journal of neuroscience : the official journal of the Society for Neuroscience 19, 9322-9331 (1999). 652 https://doi.org/10.1523/jneurosci.19-21-09322.1999 74 Jain, S. et al. RET is dispensable for maintenance of midbrain dopaminergic neurons in 654 adult mice. The Journal of neuroscience : the official journal of the Society for Neuroscience 26, 11230-11238 (2006). https://doi.org/10.1523/jneurosci.1876-06.2006 656 75 Golden, J. P . et al. RET Signaling Is Required for Survival and Normal Function of Nonpeptidergic Nociceptors. The Journal of Neuroscience 30, 3983 (2010). 658 https://doi.org/10.1523/JNEUROSCI.5930-09.2010 76 Marečková, M. et al. An integrated single-cell reference atlas of the human 660 endometrium. Nature genetics 56, 1925-1937 (2024). https://doi.org/10.1038/s41588- 024-01873-w 662 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 25 77 Potlog-Nahari, C. et al. CD10 immunohistochemical staining enhances the histological detection of endometriosis. Fertil Steril 82, 86-92 (2004). 664 https://doi.org/10.1016/j.fertnstert.2003.11.059 78 Shin, H. J., Gil, M. & Lee, I. S. Association of Elevated Expression Levels of COL4A1 in 666 Stromal Cells with an Immunosuppressive Tumor Microenvironment in Low-Grade Glioma, Pancreatic Adenocarcinoma, Skin Cutaneous Melanoma, and Stomach 668 Adenocarcinoma. J Pers Med 12 (2022). https://doi.org/10.3390/jpm12040534 79 Bhuiyan, S. A. et al. Harmonized cross-species cell atlases of trigeminal and dorsal root 670 ganglia. bioRxiv (2023). https://doi.org/10.1101/2023.07.04.547740 80 Shiers, S. I. et al. Nageotte nodules in human DRG reveal neurodegeneration in painful 672 diabetic neuropathy. bioRxiv (2024). https://doi.org/10.1101/2024.08.22.609215 81 Korobkin, A. A. et al. Developmental Changes in the Expression of TRPV1 Channels in 674 Autonomic Nervous System Neurons. Neuroscience and Behavioral Physiology 43, 743- 747 (2013). https://doi.org/10.1007/s11055-013-9803-3 676 82 Arnold, J. et al. Imbalance between sympathetic and sensory innervation in peritoneal endometriosis. Brain, Behavior, and Immunity 26, 132-141 (2012). 678 https://doi.org/https://doi.org/10.1016/j.bbi.2011.08.004 83 Andrade, C. Z Scores, Standard Scores, and Composite Test Scores Explained. Indian J 680 Psychol Med 43, 555-557 (2021). https://doi.org/10.1177/02537176211046525 84 Yin, M. et al. Comprehensive Analysis of RNA-Seq in Endometriosis Reveals Competing 682 Endogenous RNA Network Composed of circRNA, lncRNA and mRNA. Front Genet 13, 828238 (2022). https://doi.org/10.3389/fgene.2022.828238 684 85 Ma, J. et al. Single-cell transcriptomic analysis of endometriosis provides insights into fibroblast fates and immune cell heterogeneity. Cell & Bioscience 11, 125 (2021). 686 https://doi.org/10.1186/s13578-021-00637-x 86 Poli-Neto, O. B., Meola, J., Rosa-e-Silva, J. C. & Tiezzi, D. Transcriptome meta-analysis 688 reveals differences of immune profile between eutopic endometrium from stage I-II and III-IV endometriosis independently of hormonal milieu. Scientific Reports 10, 313 (2020). 690 https://doi.org/10.1038/s41598-019-57207-y 87 Cassim, S. & Montemagno, C. Neural Innervation of Tumors: Mechanisms, Hallmarks, 692 and Therapeutic Opportunities. Cancers 18, 1063 (2026). 88 Fielder, G. C. et al. The GDNF Family: A Role in Cancer? Neoplasia 20, 99-117 (2018). 694 https://doi.org/10.1016/j.neo.2017.10.010 89 Jafarabady, K. et al. Brain-derived neurotrophic factor (BDNF) as a potential marker of 696 endometriosis: a systematic review and meta-analysis. BMC Womens Health 24, 39 (2024). https://doi.org/10.1186/s12905-023-02877-0 698 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 26 90 Chang, L., Shan, J., Li, D. & Wang, X. Neuroendocrine-Immune Axis in Endometriosis: A Review on How the Nervous System Goes Beyond Pain Perception. Biomolecules 15 700 (2025). https://doi.org/10.3390/biom15111536 91 Guo, Z. et al. TRESK K(+) Channel Activity Regulates Trigeminal Nociception and 702 Headache. eNeuro 6 (2019). https://doi.org/10.1523/eneuro.0236-19.2019 92 Somersan-Karakaya, S. et al. Monoclonal antibodies against GFRα3 are efficacious 704 against evoked hyperalgesic and allodynic responses in mouse join pain models but, one of these, REGN5069, was not effective against pain in a randomized, placebo-706 controlled clinical trial in patients with osteoarthritis pain. Neurobiol Pain 14, 100136 (2023). https://doi.org/10.1016/j.ynpai.2023.100136 708 93 Lee, G. J., Porreca, F. & Navratilova, E. Prolactin and pain of endometriosis. Pharmacol Ther 247, 108435 (2023). https://doi.org/10.1016/j.pharmthera.2023.108435 710 94 S. Lampl, B., R. King, C., Attaran, M. & K. Feldman, M. Adolescent endometriosis: clinical insights and imaging considerations. Abdominal Radiology 50, 4844-4853 712 (2025). https://doi.org/10.1007/s00261-025-04870-7 95 Dowlut-McElroy, T. & Strickland, J. L. Endometriosis in adolescents. Current Opinion in 714 Obstetrics and Gynecology 29 (2017). 96 Xue, Q. et al. Promoter methylation regulates estrogen receptor 2 in human 716 endometrium and endometriosis. Biol Reprod 77, 681-687 (2007). https://doi.org/10.1095/biolreprod.107.061804 718 97 Chantalat, E. et al. Estrogen Receptors and Endometriosis. International journal of molecular sciences 21 (2020). https://doi.org/10.3390/ijms21082815 720 98 Greaves, E., Grieve, K., Horne, A. W. & Saunders, P. T. Elevated peritoneal expression and estrogen regulation of nociceptive ion channels in endometriosis. The Journal of 722 clinical endocrinology and metabolism 99, E1738-1743 (2014). https://doi.org/10.1210/jc.2014-2282 724 99 Smycz-Kubanska, M., Wendlocha, D., Witek, A. & Mielczarek-Palacz, A. The role of selected cytokines from the interleukin-1 family in the peritoneal fluid of women with 726 endometriosis. Ginekol Pol (2024). https://doi.org/10.5603/gpl.101419 100 Liu, Z. et al. Fractalkine/CX3CR1 Contributes to Endometriosis-Induced Neuropathic 728 Pain and Mechanical Hypersensitivity in Rats. Frontiers in cellular neuroscience 12 (2018). https://doi.org/10.3389/fncel.2018.00495 730 101 Becker, C. M., Gattrell, W. T., Gude, K. & Singh, S. S. Reevaluating response and failure of medical treatment of endometriosis: a systematic review. Fertil Steril 108, 125-136 732 (2017). https://doi.org/10.1016/j.fertnstert.2017.05.004 102 Ness, T. J. Models of Visceral Nociception. ILAR journal 40, 119-128 (1999). 734 https://doi.org/10.1093/ilar.40.3.119 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 27 103 Ding, W. et al. Highly synchronized cortical circuit dynamics mediate spontaneous pain 736 in mice. The Journal of clinical investigation 133 (2023). https://doi.org/10.1172/jci166408 738 104 Jones, J. et al. Selective Inhibition of NaV1.8 with VX-548 for Acute Pain. New England Journal of Medicine 389, 393-405 (2023). https://doi.org/10.1056/NEJMoa2209870 740 105 Jones, G., Jenkinson, C., Taylor, N., Mills, A. & Kennedy, S. Measuring quality of life in women with endometriosis: tests of data quality, score reliability, response rate and 742 scaling assumptions of the Endometriosis Health Profile Questionnaire. Human Reproduction 21, 2686-2693 (2006). https://doi.org/10.1093/humrep/del231 744 746 748 750 752 754 756 758 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 28

760 Animal model Adult female C57Bl/6J mice (Jackson Laboratory, Stock #000664) were utilized in all studies 762 due to the nature of the disease impacting persons with a uterus. Animals were housed on a 12- hour light/dark cycle with free access to food and water. To induce the endometriosis model, a 764 syngeneic donor-host model was utilized as previously described47. Briefly, five-week-old donor mice were primed with a subcutaneous injection of 3 µg estradiol benzoate (Cayman Chemical, 766 #10006487) dissolved in 100% ethanol and brought to 100 µL in sterile saline seven days prior to tissue harvest. Following euthanasia, the uterine horns of the donor mice were harvested and 768 longitudinally incised to expose the endometrium. The tissue was minced into fragments (<1 mm3) in ice-cold Hanks’ Balanced Salt Solution (HBSS). The resulting fragments were 770 subsequently resuspended in 500 µL of pre-warmed (37°C) HBSS. Littermate recipient animals were randomly assigned to either the endometriosis (Endo) or Sham group. Endo mice received 772 an intraperitoneal (i.p.) injection of the uterine tissue suspension (one donor horn per recipient) via a 1 mL syringe fitted with an 18-gauge needle. Sham animals received 500 µL of warm 774 HBSS without tissue. To mitigate cage-effect bias and ensure investigator blinding in subsequent assays, each cage housed both Sham and Endo animals as littermate controls. 776 Behavior All behavioral testing was performed by an investigator blinded to the experimental groups. At 778 five and eight weeks post-disease induction, animals were acclimated to a raised wire mesh flooring within individual transparent plexiglass chambers for 30 minutes prior to assessment. 780 Ambient room conditions were kept consistent throughout all studies, and all testing was conducted in the morning. First, animals were observed without disruption for 10 minutes to 782 quantify spontaneous pain-like behaviors. A behavioral bout was defined as the continuous (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 29 episode of one of the following specific behaviors from onset to cessation: “Abdominal 784 squashing” was defined as the animal pressing its abdomen against the mesh floor as previously described47; Writhing was defined as abnormal stretching or rotation of the abdomen, 786 consistent with models of visceral pain102; Abdominal licking was defined as grooming behaviors specifically directed to the abdomen that was not subsequent to other grooming behaviors. 788 Following the assessment of spontaneous behaviors, abdominal mechanical sensitivity was measured by percentage withdrawal response to von Frey filaments of increasing force (0.04, 790 0.08, 0.16, and 0.32 grams). This testing paradigm was chosen based on pilot data showing a dynamic range of responses in control animals across the forces, with ceiling or floor effects 792 observed outside this range (data not shown). Positive responses were defined as the animal retreating from the filament, kicking the hind legs, or jumping in response to the fiber. The 794 number of positive responses out of 10 total trials per filament was recorded to calculate the percent response. An inter-stimulus interval of 1-2 minutes was maintained between 796 applications, with at least five minutes between different filaments. At the end of behavioral assessments, the external genitalia were observed and each animal was assigned to an estrous 798 cycle phase (proestrus, estrus, metestrus, or diestrus) based on previous reports66. Finally, animals were weighed and returned to their home cages. 800 Tissue Collection and Histological Validation For gross anatomical characterization of the disease, animals were euthanized by cervical 802 dislocation under deep isoflurane anesthesia. The peritoneum was carefully opened, and the abdominal walls were systematically surveyed for the presence of endometriotic-like lesions. 804 Next, the reproductive organs, abdominal fat pads, and mesentery were inspected for additional ectopic lesions. The location and number of lesions were documented and biopsies were 806 embedded and snap-frozen in Optimal Cutting Temperature (OCT) compound on dry ice. To confirm the presence of endometrial glands and stroma, histological validation was performed 808 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 30 on 20 µm cryosections. Briefly, sections were stained with Mayer’s Hematoxylin (Sigma-Aldrich) and Eosin (Sigma-Aldrich) (H&E) according to standard protocols, followed by ethanol 810 dehydration. Retrograde dye injections 812 At eight weeks post-disease induction animals underwent in vivo retrograde labeling of DRG neurons that innervate endometriotic lesions or the peritoneal wall. Under isoflurane anesthesia 814 (2.5% and 2% oxygen), a midline laparotomy was performed to expose the peritoneal cavity and visualize the abdominal wall. Fluorophore-conjugated Wheat Germ Agglutinin (WGA; WGA-816 640R #29026-1 and WGA-532 #29064-1; Biotium) were reconstituted in sterile ultrapure water and stored at -20°C. For targeted injections, either an endometriotic lesion or control peritoneal 818 wall tissue was pierced with a 33-gauge needle attached to a Hamiliton syringe and allowed to equilibrate for 1 minute. 0.5-2 µL of WGA was injected into the tissue over 60 seconds, with 820 volume dependent on the capacity of the tissue. After injection, the needle remained in the tissue for at least one minute prior to slow retrieval to minimize tracer backflow. As the needle 822 was withdrawn, sterile gauze was immediately placed on the site to prevent leak of the dye. In a subset of animals, distinct WGA fluorophores were injected into different anatomical locations 824 (e.g. lesion vs. peritoneal wall) to distinguish specific innervation patterns. Following injections, the abdominal muscles were sutured with 6-0 silk, and the skin was approximated with surgical 826 staples. Animals were allowed to recover for 3-7 days to permit optimal retrograde transport prior to tissue harvest. 828 DRG Dissociation and Electrophysiology Following retrograde tracer transport, animals were euthanized and transcardially perfused with 830 ice-cold HBSS. Bilateral dorsal root ganglia (DRG) from spinal levels T10 through S2 were rapidly dissected and placed into ice-cold HBSS. Immediately following DRG collection, gross 832 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 31 disease pathology was confirmed as described above. DRG neurons were dissociated as previously described68. Briefly, ganglia underwent enzymatic digestion with papain (0.33 mg/mL; 834 Worthington) for 20 minutes at 37°C, followed by collagenase type II (1.5 mg/mL; Sigma- Aldrich) for 20 minutes at 37°C. After enzymatic treatment, cells were mechanically triturated, 836 washed, and passed through a 40-µm cell strainer. Cells were centrifuged at 1000 rpm for 3 minutes, resuspended, and plated onto glass coverslips pre-coated with poly-D-lysine and 838 collagen. Neurons were maintained in complete DRG medium consisting of Neurobasal A (Gibco) supplemented with 5% fetal bovine serum (Gibco), 1% penicillin/streptomycin (Corning), 840 GlutaMAX (Life Technologies), and B-27 supplement (Gibco). Whole-cell patch-clamp recordings were performed 16-48 hours post-plating, with recording 842 times matched across groups. Recording parameters were implemented as previously described68 at room temperature. Recordings were performed in an external solution containing 844 145 mM NaCl, 2 mM CaCl2, 1.2 mM MgCl2, 7 mM glucose, and 10 mM HEPES, pH 7.3 with NaOH and 300-310 mOsm. Cells were recorded within 1 hour of removal from DRG media. 846 Neurons were required to have a stable resting membrane potential (RMP) <-35 mV and stable access resistance. To visualize WGA fluorophores, a 625 or 530 nm LED light sources 848 (ThorLabs) were used. Once a neuron was identified to be labeled by a single dye, a thick- walled borosilicate glass recording pipette (Sutter Instrument) with an average resistance of 4-6 850 MΩ (pulled with a P-97 horizontal puller; Sutter Instrument) containing intracellular solution (120 mM potassium gluconate, 5 mM NaCl, 2 mM MgCl2, 0.1 mM CaCl2, 10 mM HEPES, 1.1 mM 852 EGTA, 4 mM Na2ATP , 0.4 mM Na2GTP , 15 mM sodium phosphocreatine, adjusted to pH = 7.3 with KOH, and 292 mOsm with sucrose) was used to create a giga-ohm seal. Data were 854 acquired using a MultiClamp 700B amplifier and a Digidata 1550B digitizer (Axon Instruments) controlled by Clampex software (v11.1; Molecular Devices). Signals were sampled at 20 kHz 856 and analyzed offline. (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 32 Upon achieving the whole-cell configuration, intrinsic and evoked electrophysiological properties 858 were recorded in current-clamp mode. Intrinsic properties included cell diameter, membrane capacitance, spontaneous activity and RMP. To assess evoked excitability, cells were held at -860 60 mV and stimulated with a series of 1 second depolarizing step stimuli (square pulses) at 10 pA increments. Rheobase was defined as the minimum amount of current necessary to evoke at 862 least one action potential (AP). To evaluate repetitive firing capacity, current was injected at multiples (1-4x) of the calculated rheobase. Neurons were classified as repetitive firing if they 864 fired more than one AP during any of these current injections. AP kinetics were analyzed from the first AP following each cell’s rheobase. These parameters include the AP68: threshold 866 (voltage when the first derivative of the potential exceeded 20 mV/ms), half-width (time at 50% AP amplitude), amplitude (voltage difference from threshold to peak), and AP peak (maximum 868 depolarized membrane potential reached during the AP). All electrophysiological data were analyzed offline using Easy Electrophysiology software (v2.6.1). 870 Immunohistochemistry All tissue was collected fresh and snap frozen on dry ice. DRG were sectioned at 10 µm and 872 non-neuronal tissue was sectioned at 20 µm using a cryostat. Sections were mounted onto slides to be used immediately or stored -20°C until use36. A perimeter around sections was 874 drawn using a hydrophobic pen (Vector Laboratories) prior to fixing with 4% paraformaldehyde (PFA) for 10 minutes at room temperature. After washing with PBS, tissue was blocked and 876 permeabilized with a buffer containing 1% BSA in PBS, 0.1% Triton-X 100 and 0.1% Sodium Azide for one hour at room temperature before incubating with the target antibody/s overnight at 878 4°C in blocking buffer. Primary antibodies included: Chicken anti-Peripherin (1:500 or 1:1000; Part#: A21449); goat anti-GDNF (1:100; R&D Systems; Part#: AF-212-NA); and rabbit anti-880 GFRα1 (1:100; Abcam; Part#: Ab8026). The next day, the slides were washed in PBS prior to appropriate fluorescent-conjugated secondary antibody incubation (1:500) and/or Isolectin B4 882 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 33 (1:300; Invitrogen; Part#:121412; in the presence of 3mM CaCl2 and MgCl2) for one hour at room temperature. Slides were washed prior to counterstaining with DAPI (1:10,000; Invitrogen; 884 #D1306) and coverslipping with Prolong Gold Antifade Mountant (Invitrogen). Images were captured using a confocal microscope (Leica Stellaris 5). Acquisition parameters were kept 886 consistent across groups by an investigator blind to the groups. All histological analyses were completed by an experimenter blind to the condition and pain scores. For lesion analysis, 888 regions of interest (ROIs) were manually drawn around all glands in tissue sections identified by characteristic epithelial and stromal layers. The mean fluorescence intensity (MFI) was then 890 calculated for each ROI across at least three nonconsecutive sections per sample, given that glands were identified. To account for variations in stromal abundance within lesions, the 892 average GDNF MFI was normalized to the average DAPI signal per participant as an approximation of cell density. Correlative analysis between Peripherin and GFRα1 was 894 performed by relating the average MFI per participant. To assess the overall pathway activity, the MFI for Peripherin, GFRα1 and GDNF were converted to standardized Z-scores by the 896 following equation: ZFactor = (MFIFactor – MFIFactor Mean) / MFIFactor Standard Deviation. The Z-score for each factor was summed to create a single score for each participant (Composite Z-Score)83,103. 898 Proteomics Total protein was isolated from surgically identified lesions (5-30 mg). Protein was isolated in 900 350 µL RIPA buffer (Sigma-Aldrich, #R0278) containing a protease inhibitor cocktail (Sigma- Aldrich, #11836153001) using a motorized probed micro-tissue homogenizer on ice prior to 902 centrifugation at 14,000 g for 10 minutes at 4°C. The supernatant total protein was calculated using a BCA assay according to manufacturer’s direction (Thermo Scientific, #23225). Samples 904 were diluted to 0.5 mg/mL total protein and analyzed using the OLink Proteomic Services (Explore HT) technologies at the High-Throughput Biomarker Core at Vanderbilt University 906 Medical Center using next generation sequencing (NGS) and including quality control, NGS (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 34 read counts, and Normalized Protein Expression (NPX) values. NPX is a relative quantification 908 unit on a log2 scale. This data was exported into R (v4.3.3) for downstream analysis with the OlinkAnalyze (v4.3.1), limma (v3.56.2), and ggplot2 (v3.5.2) packages. Quality control was 910 assessed through principal component analysis (PCA), and differential expression between high/severe and mild/moderate groups was tested with the empirical-Bayes moderated t-statistic 912 from limma. Growth factors were defined a priori as a curated list of 52 genes covering the GDNF/neurotrophin family (GDNF, BDNF, NGF, NTF3/4, CNTF, NRTN, ARTN, PSPN), VEGF, 914 PDGF, FGF, EGF, IGF/IGFBP , TGF-β, BMP, NOTCH1 and colony-stimulating-factor families. Per-protein effect sizes were summarized as forest plots with family-wide Benjamini-Hochberg 916 correction. Pathway enrichment was performed on the differential-expression (DE) result from the limma high/severe vs. mild/moderate contrast, using the R package clusterProfiler with 918 org.Hs.eg.db for human gene annotation, ReactomePA for Reactome pathway enrichment, and enrichplot for visualization. 920 Patient data Participants were recruited through the Obstetrics and Gynecology (OBGYN) clinics at 922 Washington University and Barnes-Jewish Hospital (Table 1). Prior to scheduled minimally invasive gynecological surgery for the diagnosis and resection of endometriosis lesions, 924 participants completed a series of surveys during a pre-surgical study visit. The clinical pain metrics were used for all subsequent correlations with proteomics and histological data. The 926 pain surveys included ratings on the average intensity of endometriosis-related pain over the preceding 30 days on a scale ranging from no pain to the worst pain imaginable. In addition, 928 participants completed the Endometriosis Health Profile 30 (EHP-30) questionnaire104 which assesses self-reported quality of life of women with endometriosis within the past 4 weeks. The 930 survey includes 30 multiple choice questions, and the sum of the questions was calculated with higher scores indicating a worse quality of life. In the present study, the pain subscale was used 932 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 35 with higher scores indicating worse pain105. After the surgery, patients were clinically assigned endometriosis based on the standards of the American Association of Gynecologic 934 Laparoscopists (AAGL) scoring system which evaluates surgical complexity based on the number, location and infiltration depth of lesions31. 936 Statistics Statistical analyses and data visualization were performed using R v4.3.3 or GraphPad Prism 938 v10.3.1. Normality of the data was assessed using the Shapiro-Wilk test. Measurements of two groups over time from the same animal were analyzed using a two-way repeated measures 940 ANOVA followed by Tukey’s post hoc analysis. For data involving two groups over time from different animals, a two-way ANOVA with Tukey’s post hoc analysis was utilized. Measurements 942 between three or more groups at a single timepoint were tested via one-way ANOVA followed by Tukey’s post hoc test, or the Kruskal-Wallis test with Dunn’s post hoc analysis for non-normally 944 distributed data. Categorical data were analyzed using Fisher’s exact test. Linear regressions were performed to measure associations between multiple proteins or between a single protein 946 and clinical pain scores. The goodness of fit (R2) and the result of the F-test result are presented for each association. Appropriate tests of variation are denoted in individual figure legends. The 948 critical significance value was set at α<0.05, and exact p values and associations are marked on figure panels or in individual figure legends. 950 Study Approval Animals: All experimental procedures were approved by the Institutional Animal Care and Use 952 Committee (IACUC) of Washington University in St. Louis and conducted in accordance with the US National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals. 954 Patients: All study procedures were approved by the Washington University Institutional Review Board and written informed consent or assent was obtained from all participants prior to 956 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 36 enrollment. The study was preregistered in ClinicalTrail.gov (NCT06101303). Inclusion criteria included patients aged 12-45 years old with suspected or known endometriosis. Exclusion 958 criteria included pregnancy and lactation. 960 Data Availability Data is available in the Supporting Data Values file. 962 Author Contributions A.J.D. conceived of and performed most of the experiments and analyses and wrote the 964 manuscript. M.F. helped to establish the mouse model including behavioral and confirmation/cycle related analyses including histology. A.J.K. helped perform and analyze 966 human immunohistochemistry levels. J.M.M. performed proteomic analyses. M.E.M. performed Ret/IB4 in vitro experiments. R.B. helped perform mouse immunohistochemistry analyses. 968 J.G.P. advised and helped with the Ret/IB4 experiments. E.B. and W.T.R. are gynecologic surgeons who performed the endometriosis resection surgeries. W.T.R. and H.N.A. oversaw 970 recruitment of patients to the study. W.T.R, H.N.A. and R.W.G. helped conceive of the project and made primary edits to the manuscript. 972 Funding Support National Institute of General Medical Sciences (NIGMS), Washington University School of 974 Medicine, Department of Anesthesiology training grant T32GM108539 (AJD). National Institute of Health, National Institute of Child Health and Human Development, 1R21HD115568-01 976 (HNA/WTR). National Institute of Health, National Institute of Child Health and Human Development, 1K23HD110710-01 (WTR). 978 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 37 Figures and Table 980 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 38 Figure 1. Induction and validation of a mouse model of endometriosis. A. Schematic of the 982 experimental design used to induce and test the model of endometriosis (Endo) or control (Sham) in mice. B. Gross anatomy representation and histological analyses of human vesicular 984 endometriosis and mouse lesions appear similar; both present with a gland (marked by a “G”) surrounded by endometrial stroma (marked by “Str.”) and tissue inflammation (marked by 986 “Infl.”). C. In Endo animals, lesions are detected at an average rate of 1.6/animal. D. Average weight of animals between groups is not different at measured times. E. Endo and littermate 988 Sham animals were tested for mechanical sensitivity of their abdomen five weeks following the induction of the model. Percent withdrawal to calibrated von Frey fibers were recorded in 990 response to 10 abdominal applications. Endo animals responded more frequently with robust withdrawal behaviors to 0.16 and 0.32 gram-force (Two way RM ANOVA, Tukey’s; *p=0.012, 992 **p=0.003, #p=0.056; n=24-25 animals/group). F. Repeated behavior at eight weeks following model induction replicates these results (Two way RM ANOVA, Tukey’s; *p<0.05; n=22-23 994 animals/group). G. Spontaneous pain-like behaviors were also observed for 10 minutes. Abdominal dragging, defined by the animal pressing its abdomen against the grate it stands on, 996 was increased in Endo animals relative to Shams at five and eight weeks post induction (Two way RM ANOVA, Tukey’s; *p=0.017, ***p=0.0001; n=22-23 animals/group). H. Writhing 998 behavior was also increased in animals at five weeks, although not statistically different than Shams at eight weeks (Two way RM ANOVA, Tukey’s; *p=0.032, #p=0.06; n=22-23 1000 animals/group). I. No differences between groups was noted in abdominal-directed licking behaviors at any time point (Two way RM ANOVA, Tukey’s; n=22-23 animals/group). Scale=50 1002 µm. Data represented as mean +/- SEM. (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 39 1004 Figure 2. Lesion innervating neurons are more excitable compared to size-matched controls. A. Schematic of the experimental design depicting the model generation and time of 1006 laparotomy and injection of retrograde WGA into either (B) the ventral peritoneal wall or a (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 40 peritoneal lesion. C. Representative image of neurons in vitro with one labeled by WGA 1008 retrograde dye (arrow). D. Neurons that were ≤30 µm were selected for analysis with no difference in size between groups (One way ANOVA, Tukey’s). E. The resting membrane 1010 potential (RMP) of lesion innervating neurons is significantly depolarized relative to Sham Wall innervating neurons. Endo Wall innervating neurons are trending toward more depolarized 1012 potentials compared to Sham wall (One way ANOVA, Tukey’s; *p=0.013, #p=0.09). F. 23.5% of lesion-innervating neurons fire APs at rest while Sham Wall- and Endo Wall-innervating neurons 1014 do not fire APs at rest (Fisher’s exact; #p=0.055; number of cells indicated within the bars). G. The proportion of neurons that fire multiple APs or single APs is not different between groups 1016 (Fisher’s exact; p=0.32; number of cells indicated within the bars). H and H′. An AP was evoked from lesion-innervating neurons on average at lower step current injected (rheobase) compared 1018 to both controls (Kruskal-Wallis, Dunn’s; *p=0.025, **p=0.0031). I and I′. The average peak of APs in lesion-innervating neurons was significantly lower compared to wall-innervating neurons 1020 (Kruskal-Wallis, Dunn’s; **p=0.0022, ****p<0.0001). n=11 Sham Wall-innervating neurons sampled from 3 animals, 13 Endo Wall-innervating neurons sampled from 4 animals, 17 Lesion-1022 innervating neurons from 7 animals. Data represented as mean +/- SEM or percentage bars. 1024 1026 1028 1030 1032 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 41 1034 Figure 3. Neurons which innervate lesions are small-diameter in vitro. A. Coverslips of DRG neurons innervating Sham Wall, Endo Wall, or Lesions were scanned for all WGA+ cells. 1036 Lesion-innervating neurons are significantly smaller than Sham Wall-innervating neurons, and trending smaller compared to Endo Wall-innervating neurons (One way ANOVA, Tukey’s; 1038 #p=0.06, *p=0.015). B. Small-diameter (≤30 µm) neurons are more abundant than large diameter (˃30 µm) lesion-innervating neurons, but other groups have equal WGA+ 1040 representation of small and large diameter neurons (Fisher’s exact; **p=0.0061). n=24 Sham Wall, 12 Endo Wall, 13 Lesion. Data represented as the median and quartiles. 1042 1044 1046 1048 1050 1052 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 42 Figure 4. DRG neurons that innervate peritoneal lesions are preferentially small-diameter and IB4+. A. Representative images of a Sham Wall-innervating neuron (peripherin; teal) 1054 labeled with retrograde dye (WGA; magenta) but not co-stained with IB4 (orange/yellow; open arrow). B. Representative image of a Lesion-innervating neuron that co-stains with IB4 (closed 1056 arrow). C. The average proportion of WGA-labeled neurons relative to Peripherin+ neurons in WGA-present sections is unchanged by innervation target (Student’s t test). D. The diameter of 1058 Lesion-innervating neurons is smaller than Sham Wall-innervating neurons (Mann-Whitney U test; ****p<0.0001). E. Size-Frequency histograms reveals a greater frequency of small-1060 diameter neurons in DRG retrogradely labeled from lesions compared with DRG labeled from (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 43 the peritoneal wall of Sham mice. F. Co-staining of WGA marked neurons with IB4 occurs more 1062 frequently in Lesion-innervating neurons relative to Sham Wall-innervating neurons (Mann- Whitney U test, **p=0.0026). Data acquired from DRG segments T11-L2; Sham Wall n=187 1064 neurons, 5 animals; Lesion n=28 neurons, 5 animals with no dye detected in one animal. Data represented as mean +/- SEM or mean and interquartile range of WGA+ neurons with averages 1066 per animal indicated by individual dots. Scale=50 µm. 1068 1070 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 44 1072 Figure 5. Identification, abundance and probable cellular compartments of GDNF expression in lesions from patients with pain-defined endometriosis. A. Forest plot of 1074 proteomics data demonstrates the magnitude of change (Log2 Fold Change) of growth factors in endometriosis patients with high/severe endometriosis-related pelvic pain (Intensity = 7-9; 1076 N=6) compared to patients with mild/moderate pain (Intensity = 3-6; N=6). A right-shift is (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 45 indicative of enrichment in the high/severe pain group. B. Binned proteomic counts by mild 1078 (Intensity = 3-4; N=2), moderate (Intensity = 5-6; N=4), high (Intensity = 7; N=3), and severe (Intensity = 8-9; N=3) reveals a trending positive association of GDNF with pain intensity. C. 1080 Normalization of GDNF counts to total stromal cell (COL4A1) counts or (D) specifically stroma associated with endometrial glands (CD10) counts, indicates strong associations with average 1082 reported pain intensity for each. E. GDNF (yellow) expression is restricted to the stromal cell layer of endometriosis glands and is absent in sections which were not exposed to the primary 1084 antibody (E′). Arrows indicate GDNF positive stroma. F-F′′. Immunohistochemical representative images and (G) quantification of Peripherin and GFRα1 demonstrate a strong 1086 correlation in their expression levels. Arrowheads indicate double positive axons. Simple linear regressions, goodness of fit and significance values indicated on individual panels. Scale=50 1088 µm. 1090 1092 1094 1096 1098 1100 1102 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 46 1104 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 47 Figure 6. Associations between GDNF and GFRα1 expression levels in lesions with patient-reported pain. A-C. Representative images of GDNF in low and high pain reporting 1106 patients and quantification of mean fluorescence intensity of GDNF/DAPI versus reported pain intensity. D-E. Visualization of axons by peripherin staining at endometrial glands and 1108 quantification of levels by mean fluorescence intensity versus patient reported pain. G-I. Images demonstrating GFRα1 levels at endometrial glands and quantification by mean fluorescence 1110 intensity plotted against patient reported pain. J-K. MFI was converted to z-scores to make a composite of GDNF/DAPI, Peripherin, and GFRα1 to associate GDNF-GFRα1 signaling 1112 potential at endometrial glands in patients. Composite z-scores are correlated with reported pain intensity (J) and demonstrate a trend with calculated EHP-30 pain score (K). Simple linear 1114 regressions, N=6 (averaged across all glands present in at least 3 non-consecutive sections), goodness of fit and significance values indicated on individual panels. Scale=50 µm. 1116 1118 1120 1122 1124 1126 1128 1130 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 48 Figure 7. Schematic summary of the results. Dorsal root ganglion sensory neurons which innervate lesions are smaller in diameter than those which innervate control peritoneal wall. 1132 They are also preferentially IB4/Ret+ and hyperexcitable. The same subtype of sensory neurons (GFRα1+) innervate patient lesions most densely in the stromal cell compartment where GDNF 1134 is present. In participants which self-reported their endometriosis-associated pain, the levels of GDNF, innervation and GFRα1 associate with their pain score. Graphic made using BioRender. 1136 1138 1140 1142 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint 49 Table 1. Patient demographics, clinical characteristics, and pain scores. Data include self-1144 reported endometriosis pain intensity, calculated EHP-30 pain scores, and AAGL surgical stage. Anatomical location and corresponding dataset for each sample (proteomics and/or IHC) are 1146 also noted. N/A staging indicates that stage was not assigned. (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted August 26, 2026. ; https://doi.org/10.64898/2026.08.21.744503doi: bioRxiv preprint