{"paper_id":"8c3245d5-4a3d-420c-a994-d1c8da23cb84","body_text":"Applying a computational transcriptomics-based drug repositioning pipeline \nto identify therapeutic candidates for endometriosis \n \nTomiko T Oskotsky1*, Arohee Bhoja1,3*, Daniel Bunis1*, Brian L Le1,2, Idit Kosti1,2, Christine \nLi1, Sahar Houshdaran4, Sushmita Sen4, Júlia Vallvé-Juanico4, Wanxin Wang4, Erin Arthurs5, \nLauren Mahoney5, Lindsey Lang5, Brice Gaudilliere6, David K Stevenson7, Juan C Irwin4, Linda \nC Giudice4, Stacy McAllister5*, Marina Sirota1,2* \n \n1Bakar Computational Health Sciences Institute, UCSF, San Francisco, CA,  \n2Department of Pediatrics, UCSF, San Francisco, CA,  \n3Carnegie Mellon University, Pittsburgh, PA,  \n4Department of Obstetrics, Gynecology and Reproductive Sciences, UCSF, San Francisco, CA,  \n5Gynecology and Obstetrics, Gynecology and Obstetrics, Emory University, Atlanta, GA \n6Department of Anesthesiology, Pain and Perioperative Medicine, Stanford University, Stanford, \nCA,  \n7Department of Pediatrics, Stanford University, Stanford, CA \n \n  \nAbstract \nEndometriosis is a common, inflammatory pain disorder comprised of disease in the pelvis and abnormal uterine \nlining and ovarian function that affects ~200 million women of reproductive age worldwide and up to 50% of those \nwith pelvic pain and/or infertility. Existing medical treatments for e ndometriosis-related pain are often ineffective, \nwith individuals experiencing minimal or transient pain relief or intolerable side effects limiting long-term use - thus \nunderscoring the pressing need for new drug treatment strategies. In this study, we applied a computational drug \nrepurposing pipeline to endometrial gene expression data in the setting of endometriosis and controls in an \nunstratified manner as well as stratified by disease stage and menstrual cycle phase in order to identify potential \ntherapeutics from existing drugs, based on expression reversal. Out of the 3,131 unique genes differentially \nexpressed by at least one of six endometriosis signatures, only 308, or 9.8%, were in common. Similarities were \nmore pronounced when looking at therapeutic predictions: 221 out of 299 drugs identified across the six signatures, \nor 73.9%, were shared, and the majority of predicted compounds were concordant across disease stage-stratified and \ncycle phase-stratified signatures. Our pipeline returned many known treatments as well as novel candidates. We \nselected the NSAID fenoprofen, the top therapeutic candidate for the unstratified signature and among the top-\nranked drugs for the stratified signatures, for further investigation. Our drug target network analysis shows that \nfenoprofen targets PPARG and PPARA which affect the growth of endometrial tissue, as well as PTGS2 (i.e., \nCOX2), an enzyme induced by inflammation with significantly increased gene expression demonstrated in patients \nwith endometriosis who experience severe dysmenorrhea. NSAIDs are widely prescribed for endometriosis-related \ndysmenorrhea and nonmenstrual pelvic pain. Our analysis of clinical records across University of California \nhealthcare systems revealed that while NSAIDs have been commonly prescribed to the 61,306 patients identified \nwith diagnoses of endometriosis, dysmenorrhea, or chronic pelvic pain (36,543, 59.61%), fenoprofen was \ninfrequently prescribed to those with these conditions (5, 0.008%). We tested the effect of fenoprofen in an \nestablished rat model of endometriosis and determined that it successfully alleviated endometriosis-associated \nvaginal hyperalgesia, a surrogate marker for endometriosis-related pain. These findings validate fenoprofen as a \npotential endometriosis therapeutic and suggest the utility of future investigation into additional drug targets \nidentified. \n \n \nIntroduction \nEndometriosis is an estrogen-dependent inflammatory condition characterized by the presence of endometrial-like \ntissue, refluxed during menses into the pelvis or, less commonly, by hematogenous or lymphatic spread to other \nAll rights reserved. No reuse allowed without permission. \npreprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for thisthis version posted December 21, 2022. ; https://doi.org/10.1101/2022.12.20.22283736doi: medRxiv preprint \nNOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice.\n\n \nparts of the body. It affects over 200 million people of reproductive age worldwide and up to 50% of those with \ninfertility1. The most common symptom of endometriosis is pain, and ~ 50% of women with chronic pelvic pain \nhave endometriosis2. First-line treatment for endometriosis-associated pain involves non-steroidal anti-inflammatory \ndrugs (NSAIDs) and hormonal suppressive therapy with progestins, combined oral contraceptives, or GnRH \nagonists or antagonists\n3. However, these treatments are often ineffective, with nearly 19% of patients experiencing \nno reduction in pain and up to 59% having remaining pain4, thus making it essential to identify effective therapeutic \ncandidates for endometriosis-related pain.  \n \nNew drug development has been limited for endometriosis-related pain, likely due to numerous factors including the \nheterogeneity of the disease subtypes and presenting symptoms, and less investment globally in women’s health \nrelated disorders, including those associated specifically with menstruation and pain\n5. Also, traditional drug \ndevelopment is time consuming and expensive; it can take over 15 years and $1 billion to bring a new drug to \nmarket\n6. This is especially true of endometriosis due to the complexity of the disease, its etiology, and \npathophysiology. It becomes essential, then, that alternate paths are pursued. Computational drug repurposing is the \nprocess of identifying novel therapeutic applications for existing compounds via computational methods. In recent \nyears, large public datasets have been made possible by high throughput profiling technology, and efficient \ncomputation and analysis of big data have become more accessible. As a result, computational drug repurposing has \ngained traction as a modern innovation on traditional methodologies. Narrowing down candidates for experimental \nvalidation to existing drugs with transcriptomic profiles that suggest therapeutic effectiveness when applied to a \ndisease mitigates the risk of failure in early stages of drug development. In addition, since every candidate is FDA \napproved, identified drugs have already been subjected to clinical trials and have established safety and side-effect \nprofiles. This combination of factors vastly decreases time and cost, and shortens the path from initial development \nto clinical use. \n \nOne method of computational drug repurposing, pioneered by our group, uses a pattern-matching strategy to identify \ndrugs and diseases with reversed differential gene expression profiles — where genes downregulated in a disease are \nupregulated by the drug treatment and vice versa. This approach relies on transcriptomics data, which can be \nleveraged to generate profiles of gene changes for both drugs and diseases. These profiles measure genome-wide \nchanges in gene expression between an experimental state and a control state (in this case, a disease sample vs. a \nhealthy control, or a drug-exposed sample vs. unperturbed cells). The hypothesis behind this method is that a drug \nmay have a therapeutic effect on a disease if their differential gene expression profiles are opposite\n7. In the past, this \nmethod has been successfully applied to identify both known and novel treatments for inflammatory bowel disease8, \ndermatomyositis9, and liver cancer 10. In addition, it has been used to identify novel therapeutics for preterm birth 11 \nand COVID-1912, indicating the potential applications of drug repurposing to reproductive health.  \n \nIn the past, transcriptomics work in endometriosis has allowed us to characterize the unique environment of \nendometrial lesions, which includes distinctive perivascular mural cells that promote angiogenesis and immune cell \nmigration\n13; analyze patterns in gene expression between healthy controls and endometriosis patients, taking into \naccount age, disease stage, menstrual cycle phase, and other clinical factors 14; apply computational approaches to \nidentify the individual contributions of cell subtypes to the overall endometriosis phenotype 15; and identify specific \nsubtypes of cells that are only enriched in control or disease tissue — proliferating uterine natural killer cells are \nuniquely enriched in healthy samples, and endometrial stromal cells are enriched in disease samples\n16. \nTranscriptomic profiling and analysis have allowed us to better understand the mechanism underlying \nendometriosis, and the greater availability of public datasets creates opportunities for drug repurposing. \n \nIn this study, a computational drug repurposing pipeline was applied to endometrial gene expression data in the \nsetting of endometriosis and controls in order to identify potential therapeutics from existing drugs based on \nexpression reversal. Moreover, we established a rat model to validate the NSAID fenoprofen, our top drug \ncandidate, as a potential endometriosis therapeutic.  \n  \n \nMethods \n \nStudy Design \nThe overall study overview is shown in Figure 1.  \n \nAll rights reserved. No reuse allowed without permission. \npreprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for thisthis version posted December 21, 2022. ; https://doi.org/10.1101/2022.12.20.22283736doi: medRxiv preprint \n\n \nGene Expression Signature of Endometriosis \nMicroarray-based transcriptional profiling data from eutopic endometrial tissues of women, either with \nendometriosis or with no uterine or pelvic pathology (NUPP), were obtained from the National Center for \nBiotechnology Information (NCBI) Gene Expression Omnibus (GEO) Database (series accession number \nGSE51981) and cleaned and batch corrected as described in Bunis et al. (2021). Sample metadata were used to \nclassify samples by lab of origin, by disease stages I-II, disease stages III-IV\n17 or NUPP control, and by cycle-phases \nproliferative endometrium (PE), early secretory endometrium (ESE), or mid secretory endometrium (MSE), with \nany samples that could not be unambiguously mapped (n=2) discarded 18. Remaining data (n=105) were then \nnormalized with the R package justRMA 19 and batch corrected using the package ComBat 20 to reduce signal \nassociated with lab of origin, while protecting signals associated with the disease stage and the cycle phase. Using \nthe package limma\n21), unstratified and, for sensitivity analysis, stratified differential gene expression analyses were \nperformed on subsets of the samples: unstratified = all samples; stage-stratified = all control samples and either \nstage I-II or stage III-IV samples; phase-stratified = control and disease samples of the given cycle-phase (PE, ESE, \nor MSE). Genes passing cutoffs FDR-adjusted P-value < 0.05 and logFC > 1.1 and that were represented in the \nConnectivity Map (CMap) dataset from the Broad Institute\n22 were considered the significant genes for each \nsignature. \n  \nComputational Drug Repurposing Based on Gene Expression Profiling \nTo identify potential drug candidates for endometriosis, we used a nonparametric rank-based method based on \ndifferential gene expression profiles using the Kolmogorov-Smirnov statistic\n10. The hypothesis is that drugs with \nopposite transcriptional effects to those observed in endometriosis could potentially have a therapeutic effect in \ntreating endometriosis. On the drug side, CMap was used to obtain gene expression profiles from various cell lines \ntreated with small-molecule drugs. The CMap dataset has gene expression profiles (~22,000 genes) for 1,309 small-\nmolecule drug compounds cultured in up to 5 different cell lines\n22. For endometriosis, differential gene expression \nsignatures were generated as described using six different stratifications: unstratified (all samples), stratified by \nstage (stages I-II or stages III-IV), and stratified by phase (PE, ESE, or MSE). \n \nFor each endometriosis disease signature, reversal scores were calculated for each drug in the CMap dataset based \non the relationship between the gene expression profiles for the disease relative to the drug. A negative score \nindicated reverse profiles between the drug and disease signature, where upregulated genes in the disease signature \nwere downregulated in the drug signature and vice versa. A positive score indicated the opposite — similar profiles \nbetween the drug and disease signature. For drugs with multiple gene expression profiles from different cell lines or \nconcentrations, we kept the profile with the largest reversal effect, or most negative score. Permutation analysis was \ncarried out to assess significance, and drug hits with q-values < 0.0001 or reversal scores < 0 (indicating signature \nreversal) were examined further. \n \nElectronic Health Record Analysis \nThe study was approved by the University of California, San Francisco, institutional review board and considered \nsecondary research for which consent is not required. Patients with endometriosis, chronic pelvic pain, or \ndysmenorrhea who were prescribed (a) any NSAID and (b) fenoprofen were identified from the UC Data Discovery \nPortal’s UC-wide OMOP-based EMR database, which includes clinical data from over 8 million patients from \nJanuary 1, 2012 to July 30, 2022 at UC San Francisco, UC Davis, UC Irvine, UC Los Angeles, and UC San Diego. \nDuring August 2022, patients from these five UC institutions with endometriosis, chronic pelvic pain, or \ndysmenorrhea were identified by inclusion criteria of having a self- or provider- identified sex of female with at \nleast one OMOP concept id for endometriosis (OMOP concept ids 4211992, 37117191, 4072148, 4146995, \n4264439, 4182703, 36713393, 4288543, 4307585, 4176409, 4051345, 4058381, 4200841, 4260818, 194420, \n4272614, 4132140, 37209400, 197033, 4222798, 4317964, 4127413, 4019817, 139882, 37209399, 199881, \n4189364, 36713394, 4276944, 37119080, 194421, 4230333, 46273242, 42536674, 36717630, 433527, 37110261, \n37110262, 4224161, 4195507, 37209188, 4034016, 44806162, 37396113, 44806981, 4167725, 42737048, 2109446, \n42737049, 2109445, 2109444, 4306918, 4202522, and 4270918), for chronic pelvic pain (OMOP concept ids \n4133035, 4034006, and 42534971), or for dysmenorrhea (OMOP concept ids 4137754, 4159586, 4117874, and \n194696). Among these patients, we identified individuals who were ever (a) prescribed at least one of the following \nNSAIDs: ibuprofen, naproxen, celecoxib, diclofenac, etodolac, indomethacin, piroxicam, sulindac, oxaprozin, \nmeloxicam, diflunisal, ketorolac, meclofenamate, nabumetone, salsalate, and fenoprofen, and (b) prescribed \nfenoprofen.  \n \nAll rights reserved. No reuse allowed without permission. \npreprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for thisthis version posted December 21, 2022. ; https://doi.org/10.1101/2022.12.20.22283736doi: medRxiv preprint \n\n \nAnimal Model Validation  \nSubjects and vaginal cytology. Animal subjects were 24 virgin female Sprague Dawley rats (175-200 g at arrival; \nCharles River, Raleigh, NC). All rats were single housed in plastic cages lined with chip bedding and ad libitum \naccess to rodent chow and water. Housing conditions were environmentally controlled (room temperature ~22˚C, \n12-hour light/dark cycle, lights on at 07:00). Reproductive status was determined by vaginal lavage performed ~2 \nhours after lights on using traditional nomenclature for the 4 estrous stages of proestrus, estrus, metestrus, and \ndiestrus23. The study and procedures were approved by the Emory University Institutional Animal Care and Use \nCommittee (IACUC) as protocol #2021000201. All laboratory animal experimentation adhered to the NIH Guide for \nthe Care and Use of Laboratory Animals. \n  \nEndometriosis model (ENDO). The ENDO model was performed as previously described 24. At 18-20 weeks of \nage, in diestrus, rats were anesthetized intraperitoneally  with a mixture of ketamine hydrochloride (73 mg/kg) and \nxylazine (8.8 mg/kg) and placed on a heating pad to maintain body temperature (~37°C). An off-midline (left side) \nincision was made through the skin and muscle layer to expose the pelvic and abdominal organs. An /i11-cm \nsegment of mid-left uterine horn was excised and placed in warm saline. Four, 2 × 2-mm pieces of excised uterus \nwere sewn onto alternate mesenteric arteries that supply the caudal small intestine starting from the caecum using \n4.0 nylon sutures. After it was confirmed that there was no bleeding in the abdominal cavity, the muscle layer and \nskin incision were closed with chromic gut and non-absorbable suture, respectively. Rats were monitored closely \nduring recovery; the postoperative recovery period was uneventful.  \n \nBehavioral assessment of vaginal nociception. The behavioral training and testing procedures were performed as \npreviously described25. Rats were trained to perform an escape response to terminate vaginal distention produced by \nan inflatable latex balloon. During each testing session, eight different distention volumes were delivered three times \neach in random order at intervals of ~60 seconds, and percent escape response to each volume was assessed.  \n \nBehavioral apparatus and stimulator . The training and testing apparatus was a grill-floored Plexiglas® chamber \nallowing movement but preventing the rat from turning around. In the front of the chamber, a hollow tube is \nextended containing a light-emitting diode and photo sensor. When a rat extends her nose into this tube, the light \nbeam is broken, and the stimulus is terminated constituti ng an “escape response”. An opening in the rear of the \nchamber allows the catheter (attached to the vaginal stimulator) to connect to a computer-controlled stimulus-\ndelivery device.  \nThe vaginal stimulator is a small latex balloon (~ 10mm × 1.5 mm uninflated) tied to a catheter with silk suture. \nPrior to testing, the uninflated balloon is lubricated with K-Y® jelly and inserted into the mid-vaginal canal. The \nvaginal canal is then distended by delivery of different volumes of water to the balloon.  \n \nBehavioral training. Rats were first allowed to acclimate to the testing chamber 10 minutes daily for 3-4 days. Then, \nover a time period of 4-6 weeks, rats were trained to: face forward in the testing chamber without turning around \n(box training: 3-4 sessions, consecutive days), perform an “escape response” by extending their head into a hollow \ntube to interrupt a light beam (tail pinch training: 4-8 sessions, non-consecutive days), and perform an identical \n“escape response” to terminate vaginal distention stimuli (balloon training: 3-5 sessions, non-consecutive days).  \n \nBehavioral testing. Once trained, 1-hour long testing sessions consisted of 24 computer-controlled escape trials run \nat ~1-minute intervals (range 50-70 seconds). Each trial consisted of rapid inflation of the balloon (1 mL/s) to a \nfixed volume until the rat made an “escape response” or 15 seconds had elapsed, when the balloon rapidly deflated \n(0.5 mL/s). Eight different distention volumes (0.15, 0.30, 0.40, 0.55, 0.70, 0.80, 0.90 mL), including a control \nvolume (0.01mL), were delivered to the balloon three times each in random order. The computer recorded the \nstimulus and escape response for each trial. The maximum latency of 15 seconds was considered no response. \nTesting sessions were run 3 times/week, on non-consecutive days. \n \nExperimental groups. There were four groups of rats analyzed: Group 1: ENDO, fenoprofe n (30 mg/kg/day, p.o); \nGroup 2: ENDO, i buprofen (30 mg/kg/day, p.o.) (positive control); Group 3: ENDO, no treatment (negative \ncontrol); and Group 4: no ENDO, no treatment (negative c ontrol). Ibuprofen was selected as our positive control as \nit is a commonly used analgesic agent for rodents and has been effectively used in pain studies in rodents (e.g. \ninflammatory pain models) 26–28. In all groups, vaginal nociception was behaviorally assessed over three \nchronological testing periods as follows: (i) testing period 1: an initial baseline period of 8 weeks, (ii) testing period \n2: a post-ENDO or middle-testing period of 10 weeks, and (iii) testing period 3: a post-treatment or late-testing \nAll rights reserved. No reuse allowed without permission. \npreprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for thisthis version posted December 21, 2022. ; https://doi.org/10.1101/2022.12.20.22283736doi: medRxiv preprint \n\n \nperiod of 4 weeks. Data for rats in control Groups 3 and 4 (n=3/group) were retrieved and reanalyzed from an earlier \nstudy29.  \n \n \nResults \nComputational identification of drug repurposing candidates.  \nThe gene expression data were derived from samples collected from women with minimal to mild (stage I–II) or \nmore advanced (stage III-IV) endometriosis, and those without uterine or pelvic pathology (control) in proliferative, \nearly secretory, and mid-secretory phases of the menstrual cycle. General cohort characteristics are shown in \nSupplementary Table 1 and provided previously in greater detail 18. The numbers of significant differentially \nexpressed genes from unstratified, stage-stratified (i.e., stage I–II or stage III-IV), and phase-stratified (i.e., PE, \nMSE, or ESE) comparisons of patients with endometriosis to patients in the control cohort are represented in Table \n1, with the specific differentially expressed genes represented in Supplementary Tables 2A-F. Out of the 3,131 \nunique genes differentially expressed by at least one endometriosis signature and represented in the CMap database, \nonly 308, or 9.8%, were common across all six signatures (Figure 2). \n \nBy analyzing via our drug repositioning pipeline, the unstratified and stratified differential gene expression \nsignatures with the drug signatures from the CMap dataset, 236-289 drug candidates were determined per signature \nand 299 unique drugs that significantly reversed the expression profiles of the disease. Si milarities across the six \nsignatures were more pronounced for drug candidates than for differentially expressed genes: 221 out of 299 drugs, \nor 73.9%, were common to every signature (Figure 3A). Of the 221 drugs common across all six signatures, many \nreturned high reversal scores across the board, suggesting that these compounds could be used to treat different \nstages of endometriosis and across every cycle phase. As there was consistency in the majority of the drug \ncandidates from the unstratified and stratified signatures, we moved forward based solely on the unstratified \nendometriosis signature. \n \nSeveral of the drug candidates identified are from classes of medications used to treat endometriosis. \nLevonorgestrel, among the top 10 drug candidates, is a progestin recommended for treating endometriosis\n30. Also \namong the identified drug candidates are non-steroidal anti-inflammatory drugs (NSAIDs) such as acetylsalicylic \nacid (commonly known as aspirin), mefenamic acid, indomethacin, naproxen, and diclofenac, which are frequently \nrecommended to alleviate pain and inflammation in dysmenorrhea patients .\n The NSAID ibuprofen, its isomer \ndexibuprofen, and the COX-2 selective inhibitor NSAIDs celecoxib, rofecoxib, and valdecoxib are not among our \npredicted therapeutic candidates since ibuprofen is not represented in CMAP, and dexibuprofen as well as the COX-\n2 inhibitors are represented in CMAP but are filtered out during pre-processing due to profile inconsistencies\n10. A \nheatmap of the top 20 drug candidates and their reversal scores for the six endometriosis signatures is shown in \nFigure 3B and demonstrates consistency of predictions across the signatures.  \n \nUsing the DrugBank database\n31, we were able to identify several proteins targeted by our top 20 drug candidates, \nincluding 13 that were targeted by two or more of identified drugs. These interactions were used to construct a \nnetwork, which can be used to visualize the unique and shared interactions between the top 20 drug candidates and \ntheir protein targets (Figure 3C). Out of the proteins displayed in the network, several were found to have a link to \nendometriosis. Peroxisome proliferator activated receptors gamma and alpha (PPARG and PPARA), which are \ncommonly targeted by NSAID drugs including fenoprofen, can impede the growth of endometrial tissue when \nactivated\n32,33. Prostaglandin-endoperoxidase synthase 2 (PTGS2) gene expression has been found to be significantly \nincreased in ectopic and eutopic endometrium of women with endometriosis compared to women without this \ncondition\n34,35. Moreover, among endometriosis patients, PTGS2 expression in eutopic endometrium has been shown \nto be significantly greater in women with higher pain scores for dysmenorrhea36. Dopamine receptor type-2 (DRD2) \npolymorphisms have been identified in patients with endometriosis, and treatment with DRD2 agonists has been \nassociated with the disappearance or decrease in size of peritoneal endometriotic lesions\n37,38. Increased gene \nexpression of steroid 5 alpha-reductase 1 (SRD5A1) has been found in ovarian endometriosis compared to normal \nendometrium\n39. In addition, the nuclear receptor proteins NR3C1 (nuclear receptor subfamily 3 group C member 1), \nAR (androgen receptor), PR (progesterone receptor), and ESR1 (estrogen receptor 1) are expressed in endometrial \ncells\n40,41. \n \nAll rights reserved. No reuse allowed without permission. \npreprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for thisthis version posted December 21, 2022. ; https://doi.org/10.1101/2022.12.20.22283736doi: medRxiv preprint \n\n \nThrough leveraging this approach, fenoprofen, an NSAID commonly used to treat pain and arthritis, was identified \nas the top drug candidate for the unstratified signature and among the top seven drug candidates for the stratified \nsignatures. When visualizing the gene regulation of the six input endometriosis signatures and fenoprofen, the \noverall reversal pattern can be observed (Figure 3D). As fenoprofen had the highest reversal score of our drug \ncandidates and belongs to a gold standard treatment category of drugs for endometriosis, our validation efforts \nherein were focused on this medication. \n \nElectronic Health Record Analysis \nFrom the analysis of electronic medical records (EMR) across five University of California (UC) healthcare \ninstitutions (UC San Francisco (UCSF), UC Davis, UC Irvine, UC Los Angeles, and UC San Diego), there were a \ntotal of 61,306 patients with endometriosis, chronic pelvic pain, or dysmenorrhea, am ong whom 36,543 (59.61%) \nhad a prescription for an NSAID and 5 (0.008%) had a prescription for fenoprofen (Table 2). For the individual UC \nhealthcare institutions, among the 12,476 patients at UCSF with endometriosis, chronic pelvic pain, or \ndysmenorrhea, 7,752 (62.14%) had a prescription for an NSAID and 1 (0.008%) had a prescription for fenoprofen; \namong the 10,040 patients at UC Davis with endometriosis, chronic pelvic pain, or dysmenorrhea, 6163 (61.38%) \nhad a prescription for an NSAID and 0 (0.000%) had a prescription for fenoprofen; among the 5,694 patients at UC \nIrvine with endometriosis, chronic pelvic pain, or dysmenorrhea, 3,545 (62.26%) had a prescription for an NSAID \nand 0 (0.000%) had a prescription for fenoprofen; among the 23017 patients at UC Los Angeles with endometriosis, \nchronic pelvic pain, or dysmenorrhea, 13,041 (56.66%) had a prescription for an NSAID and 3 (0.013%) had a \nprescription for fenoprofen; and among the 10,079 patients at UC San Diego with endometriosis, chronic pelvic \npain, or dysmenorrhea, 6,042 (59.95%) had a prescription for an NSAID and 1 (0.010%) had a prescription for \nfenoprofen (Table 2).  \n \nAnimal Model Validation \nTo validate the in-silico findings, we used an established animal model of e ndometriosis that produces vaginal \nhyperalgesia, a surrogate marker for endometriosis-related pain. In this model, uterine pieces are autotransplanted \nonto mesenteric abdominal arteries which form vascularized cyst-like structures. Vaginal hyperalgesia, or an \nincrease in vaginal nociception, develops and stab ilizes by ten weeks in this model\n29. Vaginal nociception was \nassessed as an escape response to a noxious stimulus, a water filled ball oon. Escape response was measured as a \nfunction of vaginal balloon distention volume. Fenoprofen was dosed orally for four weeks. As controls, ibuprofen \nwas orally dosed or no treatment was delivered. As an additional control, rats with no endometriosis received no \ntreatment. In all four groups, vaginal nociception was assessed and compared over three testing periods (i) an initial \nbaseline period of eight weeks (ii) a post-endo or middle-testing period of ten weeks, and (iii) a post-treatment or \nlate-testing period of four weeks. \n \nResponses among fenoprofen (30 mg/kg/day, p.o.) were significantly increased during the post-endo surgery period \ncompared to the baseline period, when volumes of 0.15, 0.3, 0.4, 0.55, 0.7, and 0.8 mL of water were delivered \n(Mann Whitney U test, Bonferroni-corrected p-value threshold of 0.05. Figure 5A, Table 3). During the post-\ntreatment period, escape responses were significantly decreased compared to the post-endo surgery period when \nvolumes of 0.15, 0.3, 0.4, 0.55, 0.7, and 0.8 mL of water were delivered (Mann Whitney U test, Bonferroni-\ncorrected p-value threshold of 0.05. Figure 5A, Table 3). No statistically significant difference was found in the \nescape responses between the baseline period and the post-treatment period for any volume of water delivered to the \nfenoprofen treated subjects (Mann Whitney U test, Bonferroni-corrected p-value threshold of 0.05. Figure 5A, Table \n3). \n \nSimilarly, among ibuprofen (30 mg/kg/day, p.o.) treated animals, escape responses were significantly increased \nduring the post-endo surgery period compared to the baseline period, when volumes of 0.15, 0.3, 0.4, 0.55, 0.7, and \n0.8 mL of water were delivered (Mann Whitney U test, Bonferroni-corrected p-value threshold of 0.05. Figure 5B, \nTable 4). During the post-treatment period, escape responses were significantly decreased compared to the post-\nendo surgery period, when volumes of 0.15, 0.3, 0.4, 0.55, 0.7, and 0.8 mL of water were delivered (Mann Whitney \nU test, Bonferroni-corrected p-value threshold of 0.05. Figure 5B, Table 4). No statistically significant difference \nwas found in the escape responses between the baseline period and the post-treatment period for any volume of \nwater delivered to the Ibuprofen treated subjects (Mann Whitney U test, Bonferroni-corrected p-value threshold of \n0.05. Figure 5B, Table 4). \n \nAll rights reserved. No reuse allowed without permission. \npreprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for thisthis version posted December 21, 2022. ; https://doi.org/10.1101/2022.12.20.22283736doi: medRxiv preprint \n\n \nAmong animals that received neither endo surgery nor treatment, no statistically significant difference was found in \nthe escape responses between the baseline and post-endo surgery periods, the post-endo surgery and post-treatment \nperiods, or the baseline and post-treatment periods when any volume of water was delivered (Mann Whitney U test, \nBonferroni-corrected p-value threshold of 0.05. Figure 5C, Table 5). \n \nAmong animals that received endo surgery but no treatment, escape responses were significantly increased during \nthe post-endo surgery period compared to the baseline period, when volumes of 0.15, 0.3, 0.4, 0.55, and 0.7 mL of \nwater were delivered (Mann Whitney U test, Bonferroni-corrected p-value threshold of 0.05. Figure 5D, Table 6). \nDuring the post-treatment period, escape responses were also significantly increased compared to the baseline period \nwhen volumes of 0.3, 0.4, and 0.55 mL of water were delivered (Mann Whitney U test, Bonferroni-corrected p-\nvalue threshold of 0.05. Figure 5D, Table 6). No statistically significant difference was found in the escape \nresponses between the post-endo surgery period and the post-treatment period for any volume of water delivered to \nthese subjects (Mann Whitney U test, Bonferroni-corrected p-value threshold of 0.05. Figure 5D, Table 6). \n \n \nDiscussion \nEndometriosis is an estrogen-dependent inflammatory disorder, with both local (pelvic) and systemic components \nthat commonly contribute to pelvic pain and infertility\n42–44. Therapies for pain include surgical resection of disease \nand/or medical approaches mostly aimed at reducing ovarian estrogen action or production. Unfortunately, ~50% of \npatients need repeat surgery within 5 years or recurrent symptoms, and medical therapies are either ineffective or \npromote intolerable side effects that limit their long-term use 1,4. Recent FDA approval of new drugs for \nendometriosis (e.g., GnRH antagonists Elagolix and Myfembree) brings hope for those who suffer from this \ndisease45,46; however, off-target effects (e.g., bone density) await long-term, post-marketing studies. Thus, there is a \npressing need for novel drug discovery to improve patient symptoms and quality of life. Our study identified several \nexisting drugs with potential therapeutic applications to endometriosis using a transcriptomics-based computational \ndrug repurposing approach. The differential gene expression profiles for endometriosis that were unstratified as well \nas, for sensitivity analysis, stratified by disease stage and menstrual cycle phase were compared with the profiles of \nseveral small molecule compounds tested on human cell lines, yielding 299 unique drug hits with significant (q-\nvalue < 0.0001) reversal effects. We found that therapeutic predictions were relatively consistent across stage and \ncycle phase with 221 shared predictions. \nWhen categorized by drug class/ATC code, two prominent categories for the predicted therapeutics were anti-\ninflammatory drugs and sex hormones; drugs from both categories have extensively been used to treat \nendometriosis\n3. Several drugs that the pipeline returned are current gold standard treatments, such as levonorgestrel, \nmefenamic acid, acetylsalicylic acid (aspirin), and naproxen; others were novel candidates. \nFrom the drugs identified by the computational drug repurposing approach, we chose fenoprofen for further \nvalidation since it returned the highest reversal score and belongs to a class of drugs (NSAIDs) that is a current first-\nline treatment for endometriosis. Fenoprofen is a medication available by prescription only and has been in clinical \nuse since this drug was approved by the FDA in 1976\n47 and is indicated for the relief of mild to moderate pain in \nadults and, in particular, relief of signs and symptoms of rheumatoid arthritis and osteoarthritis 48. In our analysis of \nthe electronic medical records across five University of California healthcare institutions, we found that while \nNSAIDs have been commonly prescribed (56.66% to 62.26%) for patients with endometriosis, chronic pelvic pain, \nor dysmenorrhea diagnosis, the NSAID fenoprofen was prescribed for the minority (0% to 0.013%) of patients with \nthese conditions.  \nWe tested the NSAID fenoprofen in a rat model of endometriosis that displays vaginal hyperalgesia, a surrogate \nmarker for endometriosis-related pain. We determined that oral treatment with fenoprofen significantly alleviated \nendometriosis-associated vaginal hyperalgesia, s imilar to oral ibuprofen treatment. In endometriosis rats with no \ntreatment, vaginal hyperalgesia was maintained, which confirmed that the alleviation of hyperalgesia observed in the \ntreatment groups was not due to additional vaginal nociceptive testing post-endometriosis. In rats with no \nendometriosis and no treatment, no significant changes in vaginal nociception occurred, which suggests that any \nobserved changes in the other groups were not due to lengthy vaginal nociceptive testing alone. Overall, these \nfindings support fenoprofen as a potential therapeutic for endometriosis-associated pain. \nAll rights reserved. No reuse allowed without permission. \npreprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for thisthis version posted December 21, 2022. ; https://doi.org/10.1101/2022.12.20.22283736doi: medRxiv preprint \n\n \nNSAIDs prevent or reduce production of prostaglandins, which in turn can help relieve pain from endometriosis. \nWhile NSAIDs are a commonly prescribed, first-line treatment of endometriosis, there is currently limited evidence \nto support the effectiveness of any NSAID over another for endometriosis pain relief49. Our findings suggest that the \nNSAID fenoprofen, currently infrequently prescribed for endometriosis, may be an effective treatment for \nindividuals with this condition, although this would need further validation in patient cohorts. Our drug target \nnetwork analysis for the top drug candidates shows that PPARG and PPARA, which impede the growth of \nendometrial tissue when activated, are both targeted by fenoprofen\n32,33. In addition, our network analysis showed \nthat fenoprofen targets the enzymes PTGS1 and PTGS2 (i.e., COX1 and COX2, respectively). PTGS1 and PTGS2 \nhave been shown to be inhibited to varying degrees in blood and gastric tissue by the available NSAIDs\n50, but any \ndifferences in the degree of inhibition of PTGS1 and PTGS2 in endometrial tissue by fenoprofen and other NSAIDs \nhave not yet been demonstrated. PTGS1 is a constitutively expressed enzyme and involved in maintaining cell \nhomeostasis\n51. In contrast, PTGS2 is an enzyme uncommonly expressed normally, but is induced during \ninflammation as well as cell proliferation and differentiation 34. Significantly increased gene expression of PTGS2 \nhas been found in the ectopic and eutopic endometrium of women with endometriosis compared to women without \nthis condition\n34,35. Furthermore, PTGS2 expression in the eutopic endometrium of women with endometriosis has \nbeen shown to be significantly higher in those with more severe dysmenorrhea 36. Targeting key factors involved in \nendometriosis pathogenesis and symptomatology may contribute to the effectiveness of fenoprofen in treating \nendometriosis.  \nOur study has several limitations. The endometriosis signatures were generated from bulk gene data that included \n105 microarray samples; they could be made more robust by incorporating additional public expression datasets. \nSingle-cell data could also be used to investigate the potential effects of drug hits on specific types of endometrial \ncells and for combination therapy predictions. Drugs like ibuprofen and GnRH antagonists that lack representation \nin the CMAP data used by the drug repurposing pipeline would not be identified by the pipeline, nor would drugs \nthat are represented in CMAP but filtered out during pre-processing due to profile inconsistencies\n10, such as the \nCOX-2 selective inhibitor NSAIDs. The nature of the drug repurposing pipeline prioritizes drugs that have a high \nreversal effect on the disease signature; it does not take into account whether transcriptional effects are limited \nsolely to genes that the disease also affects. A drug that causes wide-ranging gene changes—including reversal to \nthe gene changes caused by endometriosis—may present in the list of identified drugs, and may be therapeutic. \nHowever, unrelated gene changes could cause undesirable side effects, and depending on specificity and severity of \nside effects, these drugs may have limited applicability clinically. Moreover, in our study, we did not assess the \neffects of fenoprofen on disease burden. Out of all the drug candidates across every endometriosis signature, many \nwere antipsychotics and other drugs that affect a wide range of genes. In addition, the compounds from the CMap \ndataset were tested on cancer cell lines; the drug’s effects on endometrial tissue would be far better determiners of \nits potential applications to endometriosis. A further limitation of our study is that we validated fenoprofen as a \npotential endometriosis therapeutic in a rodent model. Although our model mimics many disease features of women \nwith endometriosis, rats do not menstruate or develop endometriosis spontaneously. Therefore, menstruating non-\nhuman primates could be considered a more appropriate model as they develop endometriosis spontaneously; \nhowever, because of their close phylogenetic relationship to humans, these models come with unique ethical \nconsiderations\n52,53 as well as limiting financial cost.  \nTo summarize, we applied a computational drug repurposing pipeline to identify potential therapeutics for \nendometriosis-related pain. The pipeline returned many known treatments as well as novel candidates. We tested the \nidentified therapeutic candidate fenoprofen in an established rat model of endometriosis. We determined that \nfenoprofen successfully alleviated endometriosis-associated vaginal hyperalgesia, a surrogate marker for \nendometriosis-related pain. These findings validate fenoprofen as a potential endometriosis therapeutic and suggest \nthe utility of future investigation into additional drug candidates identified. \n \n \n \n \nContributions \nD.B, A.B, L.C.G, S.M., and M.S. designed the study, experiments, and analytic plan. T.T.O, A.B., D.B, C.L., and \nS.M. carried out data acquisition, processing, and analysis. T.T.O, A.B., D.B, B.L.L, I.K., B.G., D.K.S., J.C.I., \nL.C.G., S.M., and M.S. carried out computational and statistical analysis. T.T.O, A.B., D.B, B.L.L, I.K., B.G., \nAll rights reserved. No reuse allowed without permission. \npreprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for thisthis version posted December 21, 2022. ; https://doi.org/10.1101/2022.12.20.22283736doi: medRxiv preprint \n\n \nD.K.S., J.C.I., L.C.G., S.M., and M.S. interpreted results. E.A., L.M., L.L., and S.M. carried out the validation \nexperiments and analyzed the relevant data. T.T.O, A.B., and S.M. wrote the manuscript. All the authors \nparticipated in relevant discussions, edited and reviewed the manuscript. \n \nAcknowledgements \nThe work was in part supported by NIH P01HD106414 (T.T.O., A.B., B.G., D.K.S., J.C.I., L.C.G., S.M., M.S.), \nNIH P50 HD055764 (A.B., S.H., S.S., W.W., J.C.I., L.C.G., M.S.), and NIH R00HD093858 (E.A., L.M., L.L., \nS.M.), as well as by the March of Dimes Prematurity Research Center at UCSF (T.T.O., B.L.L., I.K., M.S.), the \nMarch of Dimes Prematurity Research Center at Stanford University (B.G., D.K.S.), and the Stanford Maternal and \nChild Health Research Institute (B.G., D.K.S.). The authors acknowledge the use of resources developed and \nsupported by the UCSF Bakar Computational Health Sciences Institute Information Commons team, and thank \nmembers of this team for technical support. The authors also thank the Center for Data-driven Insights and \nInnovation at UC Health (CDI2; https://www.ucop.edu/uc-health/functions/center-for-data-driven-insights-and-\ninnovationscdi2.html), for its analytical and technical support related to use of the UC Health Data Warehouse and \nrelated data assets, including the UC COVID Research Data Set (CORDS). \n \nCompeting Interests: \nM.S. is an advisor to Aria Pharmaceuticals. The other authors declare no competing financial interests. \n \nData and Code Availability \nData were obtained from the National Center for Biotechnology Information (NCBI) Gene Expression Omnibus \n(GEO) Database (series accession number GSE51981) \nhttps://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE51981 \nThe UCSF EHR database is available to UCSF-affiliated individuals who can contact UCSF’s Clinical and \nTranslational Science Institute (CTSI) (ctsi@ucsf.edu) or the UCSF’s Information Commons team for more \ninformation (Info.Commons@ucsf.edu). UCDDP is only available to UC researchers who have completed analyses \nin their respective UC first and have provided justification for scaling their analyses across UC health centers. \nCode for transcriptomic data processing associated with the current submission is available at \nhttps://doi.org/10.3389/fimmu.2021.788315, and code for computational drug repurposing pipeline associated with \nthe current submission is available at https://doi.org/10.1053%2Fj.gastro.2017.02.039.  \n \nAll rights reserved. No reuse allowed without permission. \npreprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for thisthis version posted December 21, 2022. ; https://doi.org/10.1101/2022.12.20.22283736doi: medRxiv preprint \n\n \nReferences \n1. Giudice, L. C. Endometriosis. N. Engl. J. Med. 362, 2389–2398 (2010). \n2. Gambone, J. C., Mittman, B. S., Munro, M. G., Scialli, A. R. & Winkel, C. A. Consensus statement for \nthe management of chronic pelvic pain and endometriosis: proceedings of an expert-panel consensus \nprocess. Fertil. Steril. 78, 961–972 (2002). \n3. Becker, C. M. et al. ESHRE guideline: endometriosis†. Hum. Reprod. Open 2022, hoac009 (2022). \n4. Becker, C. M., Gattrell, W. T., Gude, K. & Singh, S. S. Reevaluating response and failure of medical \ntreatment of endometriosis: a systematic review. Fertil. Steril. 108, 125–136 (2017). \n5. As-Sanie, S. et al. Assessing research gaps and unmet needs in endometriosis. Am. J. Obstet. Gynecol. \n221, 86–94 (2019). \n6. Wouters, O. J., McKee, M. & Luyten, J. Estima ted Research and Development Investment Needed to \nBring a New Medicine to Market, 2009-2018. JAMA 323, 844–853 (2020). \n7. Sirota, M. et al. Discovery and preclinical validation of drug indications using compendia of public gene \nexpression data. Sci. Transl. Med. 3, 96ra77 (2011). \n8. Dudley, J. T. et al. Computational repositioning of the anticonvulsant topiramate for inflammatory bowel \ndisease. Sci. Transl. Med. 3, 96ra76 (2011). \n9. Cho, H. G., Fiorentino, D., Lewis, M., Sirota, M. & Sarin, K. Y. Identification of alpha-adrenergic \nagonists as potential therapeutic agents for dermatomyositis through drug-repurposing using public \nexpression datasets. J. Invest. Dermatol. 136, 1517–1520 (2016).\n \n10. Chen, B. et al. Computational Discovery of Niclosamide Ethanolamine, a Repurposed Drug Candidate \nThat Reduces Growth of Hepatocellular Carcinoma Cells In Vitro and in Mice by Inhibiting Cell Division \nCycle 37 Signaling. Gastroenterology 152, 2022–2036 (2017).\n \n11. Le, B. L., Iwatani, S., Wong, R. J., Stevenson, D. K. & Sirota, M. Computational discovery of therapeutic \ncandidates for preventing preterm birth. JCI Insight 5, e133761 (2020). \n12. Le, B. L. et al.  Transcriptomics-based drug repositioning pipeline identifies therapeutic candidates for \nCOVID-19. bioRxiv 2020.10.23.352666 (2020) doi:10.1101/2020.10.23.352666.  \n13. Tan, Y. et al. Single cell analysis of endometriosis reveals a coordinated transcriptional program driving \nAll rights reserved. No reuse allowed without permission. \npreprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for thisthis version posted December 21, 2022. ; https://doi.org/10.1101/2022.12.20.22283736doi: medRxiv preprint \n\n \nimmunotolerance and angiogenesis across eutopic and ectopic tissues. 2021.07.28.453839 Preprint at \nhttps://www.biorxiv.org/content/10.1101/2021.07.28.453839v1 (2021).  \n14. Gabriel, M. et al. A relational database to identify differentially expressed genes in the endometrium and \nendometriosis lesions. Sci. Data 7, 284 (2020). \n15. Bunis, D. G. et al.  Whole-Tissue Deconvolution and scRNAseq Analysis Identify Altered Endometrial \nCellular Compositions and Functionality Associated With Endometriosis. Front. Immunol. 12, 788315 \n(2022). \n16. Shih, A. J. et al.  Single cell analysis of menstrual endometrial tissues defines phenotypes associated with \nendometriosis. 2022.02.10.22270810 Preprint at \nhttps://www.medrxiv.org/content/10.1101/2022.02.10.22270810v1 (2022).  \n17. American Society for Reproductive Medicine. Revised American Society for Reproductive Medicine \nclassification of endometriosis: 1996. Fertil. Steril. 67, 817–821 (1997). \n18. Tamaresis, J. S. et al. Molecular Classification of Endometriosis and Disease Stage Using High-\nDimensional Genomic Data. Endocrinology 155, 4986–4999 (2014). \n19. Gautier, L., Cope, L., Bolstad, B. M. & Irizarry, R. A. affy--analysis of Affymetrix GeneChip data at the \nprobe level. Bioinforma. Oxf. Engl. 20, 307–315 (2004). \n20. Leek, J. et al. sva: Surrogate Variable Analysis. (2022) doi:10.18129/B9.bioc.sva.  \n21. Ritchie, M. E. et al.  limma powers differential expression analyses for RNA-sequencing and microarray \nstudies. Nucleic Acids Res. 43, e47 (2015). \n22. Lamb, J. et al. The Connectivity Map: using gene-expression signatures to connect small molecules, \ngenes, and disease. Science 313, 1929–1935 (2006). \n23. Becker, J. B. et al. Strategies and methods for research on sex differences in brain and behavior. \nEndocrinology 146, 1650–1673 (2005). \n24. Berkley, K. J., Dmitrieva, N., Curtis, K. S. & Papka, R. E. Innervation of ectopic endometrium in a rat \nmodel of endometriosis. Proc. Natl. Acad. Sci. U. S. A. 101, 11094–11098 (2004). \n25. Berkley, K. J., McAllister, S. L., Accius, B. E. & Winnard, K. P. Endometriosis-induced vaginal \nhyperalgesia in the rat: effect of estropause, ovariectomy, and estradiol replacement. Pain 132, S150–S159 \nAll rights reserved. No reuse allowed without permission. \npreprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for thisthis version posted December 21, 2022. ; https://doi.org/10.1101/2022.12.20.22283736doi: medRxiv preprint \n\n \n(2007). \n26. Foley, P. L., Kendall, L. V. & Turner, P. V. Clinical Management of Pain in Rodents. Comp. Med. 69, \n468–489 (2019). \n27. Ashraf, S., Bouhana, K. S., Pheneger, J., Andrews, S. W. & Walsh, D. A. Selective inhibition of \ntropomyosin-receptor-kinase A (TrkA) reduces pain and joint damage in two rat models of inflammatory \narthritis. Arthritis Res. Ther. 18, 97 (2016). \n28. Craft, R. M., Hewitt, K. A. & Britch, S. C. Antinociception produced by nonsteroidal anti-inflammatory \ndrugs in female vs male rats. Behav. Pharmacol. 32, 153–169 (2021). \n29. McAllister, S. L., Dmitrieva, N. & Berkley, K. J. Sprouted innervation into uterine transplants contributes \nto the development of hyperalgesia in a rat model of endometriosis. PloS One 7, e31758 (2012). \n30. Bahamondes, L., Petta, C. A., Fernandes, A. & Monteiro, I. Use of the levonorgestrel-releasing \nintrauterine system in women with endometriosis, chronic pelvic pain and dysmenorrhea. Contraception \n75, S134-139 (2007).\n \n31. Wishart, D. S. et al. DrugBank 5.0: a major update to the DrugBank database for 2018. Nucleic Acids Res. \n46, D1074–D1082 (2018). \n32. Lebovic, D. I., Kavoussi, S. K., Lee, J., Banu, S. K. & Arosh, J. A. PPAR γ  Activation Inhibits Growth and \nSurvival of Human Endometriotic Cells by Suppressing Estrogen Biosynthesis and PGE2 Signaling. \nEndocrinology 154, 4803–4813 (2013). \n33. Chen, Z. et al. Lipidomic Alterations and PPARα  Activation Induced by Resveratrol Lead to Reduction in \nLesion Size in Endometriosis Models. Oxid. Med. Cell. Longev. 2021, e9979953 (2021). \n34. Ota, H., Igarashi, S., Sasaki, M. & Tanaka, T. Distribution of cyclooxygenase-2 in eutopic and ectopic \nendometrium in endometriosis and adenomyosis. Hum. Reprod. 16, 561–566 (2001).  \n35. Santulli, P. et al. Hormonal Therapy Deregulates Prostaglandin-Endoperoxidase Synthase 2 (PTGS2) \nExpression in Endometriotic Tissues. J. Clin. Endocrinol. Metab. 99, 881–890 (2014). \n36. Matsuzaki, S. et al. Cyclooxygenase-2 expression in deep endometriosis and matched eutopic \nendometrium. Fertil. Steril. 82, 1309–1315 (2004). \n37. Bilibio, J. P. et al. Dopamine receptor D2 genotype (3438) is associated with moderate/severe \nAll rights reserved. No reuse allowed without permission. \npreprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for thisthis version posted December 21, 2022. ; https://doi.org/10.1101/2022.12.20.22283736doi: medRxiv preprint \n\n \nendometriosis in infertile women in Brazil. Fertil. Steril. 99, 1340–1345 (2013). \n38. Gómez, R. et al. Effects of hyperprolactinemia treatment with the dopamine agonist quinagolide on \nendometriotic lesions in patients with endometriosis-associated hyperprolactinemia. Fertil. Steril. 95, 882-\n888.e1 (2011). \n39. Hevir, N., Vouk, K., Šinkovec, J., Ribi č -Pucelj, M. & Lanišnik Rižner, T. Aldo-keto reductases AKR1C1, \nAKR1C2 and AKR1C3 may enhance progesterone metabolism in ovarian endometriosis. Chem. Biol. \nInteract. 191, 217–226 (2011). \n40. Simmons, R. M., Satterf ield, M. C., Welsh, T. H., Bazer, F. W. & Spencer, T. E. HSD11B1, HSD11B2, \nPTGS2, and NR3C1 Expression in the Peri-Implantation Ovine Uterus: Effects of Pregnancy, \nProgesterone, and Interferon Tau1. Biol. Reprod. 82, 35–43 (2010).\n \n41. Pluchino, N. et al. Estrogen receptor-α  immunoreactivity predicts symptom severity and pain recurrence \nin deep endometriosis. Fertil. Steril. 113, 1224-1231.e1 (2020). \n42. Burney, R. O. & Giudice, L. C. Pathogenesis and pathophysiology of endometriosis. Fertil. Steril. 98, \n511–519 (2012). \n43. Zondervan, K. T. et al. Endometriosis. Nat. Rev. Dis. Primer 4, 1–25 (2018). \n44. Taylor, H. et al. Pre-IVF treatment with a GnRH antagonist in women with endometriosis (PREGNANT): \nstudy protocol for a prospective, double-blind, placebo-controlled trial. BMJ Open 12, e052043 (2022). \n45. Taylor, H. S. et al. Treatment of Endometriosis-Associated Pain with Elagolix, an Oral GnRH Antagonist. \nN. Engl. J. Med. 377, 28–40 (2017). \n46. Giudice, L. C. et al.  Once daily oral relugolix combination therapy versus placebo in patients with \nendometriosis-associated pain: two replicate phase 3, randomised, double-blind, studies (SPIRIT 1 and 2). \nThe Lancet 399, 2267–2279 (2022).\n \n47. PEMD-90-15 FDA Drug Review: Postapproval Risks 1976-1985. 132.  \n48. Fenoprofen: MedlinePlus Drug Information. https://medlineplus.gov/druginfo/meds/a681026.html (2021).  \n49. Brown, J., Crawford, T. J., Allen, C., Hopewell, S. & Prentice, A. Nonsteroidal anti /i4 inflammatory drugs \nfor pain in women with endometriosis. Cochrane Database Syst. Rev. (2017) \ndoi:10.1002/14651858.CD004753.pub4. \nAll rights reserved. No reuse allowed without permission. \npreprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for thisthis version posted December 21, 2022. ; https://doi.org/10.1101/2022.12.20.22283736doi: medRxiv preprint \n\n \n50. Cryer, B. & Feldman, M. Cyclooxygenase-1 and Cyclooxygenase-2 Selectivity of Widely Used \nNonsteroidal Anti-Inflammatory Drugs. Am. J. Med. 104, 413–421 (1998). \n51. Masferrer, J. L. et al. The Role of Cyclooxygenase-2 in Inflammation. Am. J. Ther. 2, 607–610 (1995). \n52. PHS Policy on Humane Care and Use of Laboratory Animals. PHS Policy on Humane Care and Use of \nLaboratory Animals | OLAW https://olaw.nih.gov/policies-laws/phs-policy.htm (2015).  \n53. Tardif, S. D., Coleman, K., Hobbs, T. R. & Lutz, C. IACUC Review of Nonhuman Primate Research. \nILAR J. 54, 234–245 (2013). \n  \nAll rights reserved. No reuse allowed without permission. \npreprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for thisthis version posted December 21, 2022. ; https://doi.org/10.1101/2022.12.20.22283736doi: medRxiv preprint \n\n \nTables \nTable 1: Significant genes that were also represented in CMAP and total drug hits \nSignature Significant Genes \n(adj P-val < 0.05, log2FC > 1.1) \nTotal Drug Hits \n(q-val < 0.0001, reversal score > 0) \nUnstratified 620  282  \nStage 1/2 1094 275 \nStage 3/4 504 284 \nPE 1298  283  \nESE 197  236  \nMSE 650  289  \n \n \nTable 2. Prevalence of NSAID prescriptions: Patients with endometriosis, chronic pelvic pain, or dysmenorrhea \nand prescribed (a) any NSAID and (b) fenoprofen at five UC Health Care Institutions (UCSF, UCD, UCI, UCLA, \nUCSD) \nInstitution \nPatients with \nendometriosis, chronic \npelvic pain, or \ndysmenorrhea \nPatients with endometriosis, \nchronic pelvic pain, or \ndysmenorrhea and prescribed any \nNSAID (%) \nPatients with endometriosis, \nchronic pelvic pain, or \ndysmenorrhea and prescribed \nfenoprofen (%) \nUCSF 12476 7752 (62.14%) 1 (0.008%) \nUC Davis 10040 6163 (61.38%) 0 (0%) \nUC Irvine 5694 3545 (62.26%) 0 (0%) \nUC Los \nAngeles 23017 13041 (56.66%) 3 (0.013%) \nUC San \nDiego 10079 6042 (59.95%) 1 (0.010%) \nTotal 61306 36543 (59.61%) 5 (0.008%) \n \n \n \n \n \n \nAll rights reserved. No reuse allowed without permission. \npreprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for thisthis version posted December 21, 2022. ; https://doi.org/10.1101/2022.12.20.22283736doi: medRxiv preprint \n\n \n \n \nTable 3. Responses with Fenoprofen treatment. Median escape response with interquartile range (IQR) for each \ndelivered volume (0.01, 0.15,0.30, 0.40, 0.55, 0.70, 0.80, and 0.90 mL) during the baseline, post-endo surgery, and \npost-treatment periods, with Bonferroni-corrected p-values from Mann-Whitney U test for baseline period vs post-\nendo surgery period, post-endo surgery vs post-treatment period, and baseline period vs post-treatment period. * \ndenotes Bonferroni-corrected p-values below significance threshold of 0.05.\n \n  \nVolume \nBaseline \nperiod \n(BL) \nmedian \nBL \nIQR \nPost-\nEndo \nperiod \n(PE) \nmedian \nPE \nIQR \nPost-\nTreatment \nperiod (PT) \nmedian \nPT \nIQR \nBaseline \nperiod vs \nPost-Endo \nperiod \nMWU test \nBonferroni-\nadjusted \np-values \nPost-Endo \nperiod vs Post-\nTreatment \nperiod \nMWU test \nBonferroni-\nadjusted \np-values \nBaseline \nperiod vs Post-\nTreatment \nperiod \nMWU test \nBonferroni-\nadjusted \np-values \n0.01 0 (0-0) 0 (0-0) 0 (0-0) 1   1   1   \n0.15 0 (0-0) 33.3 \n(0-\n66.6) 0 (0-0) 2.7E-04 * 2.3E-03 * 0.61   \n0.3 0 (0-25) 33.3 \n(33.3-\n66.6) 0 \n(0-\n33.3) 2.0E-05 * 3.0E-06 * 1   \n0.4 0 \n(0-\n33.3) 66.6 \n(33.3-\n91.7) 16.65 \n(0-\n33.3) 4.0E-06 * 1.8E-05 * 1   \n0.55 33.3 \n(8.3-\n33.3) 83.3 \n(66.6-\n100) 33.3 \n(33.3-\n66.6) 2.7E-07 * 7.9E-07 * 0.16   \n0.7 66.6 \n(33.3-\n66.6) 100 \n(100-\n100) 66.6 \n(66.6-\n100) 1.5E-03 * 4.4E-02 * 0.97   \n0.8 66.6 \n(66.6-\n100) 100 \n(100-\n100) 100 \n(66.6-\n100) 4.8E-05 * 5.8E-03 * 1   \n0.9 100 \n(100-\n100) 100 \n(100-\n100) 100 \n(100-\n100) 0.34   0.80   1   \n  \n \n  \nAll rights reserved. No reuse allowed without permission. \npreprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for thisthis version posted December 21, 2022. ; https://doi.org/10.1101/2022.12.20.22283736doi: medRxiv preprint \n\n \nTable 4. Responses with Ibuprofen treatment (positive control). Median escape response with interquartile range \n(IQR) for each delivered volume (0.01, 0.15,0.30, 0.40, 0.55, 0.70, 0.80, and 0.90 mL) during the baseline, post-\nendo surgery, and post-treatment periods, with Bonferroni-corrected p-values from Mann-Whitney U test for \nbaseline period vs post-endo surgery period, post-endo surgery period vs post-treatment period, and baseline period \nvs post-treatment period. * denotes Bonferroni-corrected p-values below significance threshold of 0.05.\n \n  \nVolume \nBaseline \nperiod \n(BL) \nmedian \nBL \nIQR \nPost-\nEndo \nperiod \n(PE) \nmedian \nPE \nIQR \nPost-\nTreatment \nperiod (PT) \nmedian \nPT \nIQR \nBaseline \nperiod vs Post-\nEndo \n period MWU \nperiod \nBonferroni-\nadjusted \np-values \nPost-Endo \nperiod vs Post-\nTreatment \nperiod \n MWU test \nBonferroni-\nadjusted \n p-values \nBaseline \nperiod vs Post-\nTreatment \nperiod \n MWU test \nBonferroni-\nadjusted \n p-values \n0.01 0 (0-0) 0 (0-0) 0 (0-0) 1   1   1   \n0.15 0 (0-0) 16.65 \n(0-\n33.3) 0 (0-0) 3.5E-03 * 2.3E-03 * 1   \n0.3 0 (0-0) 33.3 \n(0-\n66.6) 0 (0-0) 2.0E-03 * 9.7E-04 * 1   \n0.4 0 (0-25) 66.6 \n(33.3-\n66.6) 0 \n(0-\n33.3) 1.5E-05 * 2.9E-06 * 1   \n0.55 33.3 \n(8.3-\n33.3) 83.3 \n(66.6-\n100) 0 \n(0-\n33.3) 2.5E-07 * 3.4E-08 * 1   \n0.7 66.6 \n(66.6-\n66.6) 100 \n(100-\n100) 66.6 \n(33.3-\n66.6) 5.8E-05 * 5.1E-06 * 1   \n0.8 100 \n(66.6-\n100) 100 \n(100-\n100) 100 \n(66.6-\n100) 3.9E-03 * 1.8E-03 * 1   \n0.9 100 \n(74.9-\n100) 100 \n(100-\n100) 100 \n(100-\n100) 0.12   0.80   1   \n  \n  \n  \n  \nAll rights reserved. No reuse allowed without permission. \npreprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for thisthis version posted December 21, 2022. ; https://doi.org/10.1101/2022.12.20.22283736doi: medRxiv preprint \n\n \nTable 5. Responses with no endo surger y and no treatment (negative control). Median escape response with \ninterquartile range (IQR) for each delivered volume (0.01, 0.15,0.30, 0.40, 0.55, 0.70, 0.80, and 0.90 mL) during the \nbaseline, post-endo surgery, and post-treatment periods, with Bonferroni-corrected p-values from Mann-Whitney U \ntest for baseline period vs post-endo surgery period, post-endo surgery vs post-treatment period, and baseline period \nvs post-treatment period. * denotes Bonferroni-corrected p-values below significance threshold of 0.05.\n \n  \nVolume \nBaseline \nperiod \n(BL) \nmedian \nBL \nIQR \nPost-\nEndo \nperiod \n(PE) \nmedian \nPE \nIQR \nPost-\nTreatment \nperiod (PT) \nmedian \nPT \nIQR \nBaseline \nperiod vs \nPost-Endo \n period MWU \ntest \nBonferroni-\nadjusted \np-values \nPost-Endo \nperiod vs Post-\nTreatment \nperiod \n MWU test \nBonferroni-\nadjusted \n p-values \nBaseline \nperiod vs Post-\nTreatment \nperiod \n MWU test \nBonferroni-\nadjusted \n p-values \n0.01 0 (0-0) 0 (0-0) 0 (0-0) 1   1   1   \n0.15 0 (0-0) 0 (0-0) 0 (0-0) 1   1   1   \n0.3 0 (0-0) 0 \n(0-\n33.3) 0 \n(0-\n33.3) 1   1   1   \n0.4 33.3 \n(0-\n33.3) 33.3 \n(0-\n33.3) 33.3 \n(0-\n33.3) 1   1   1   \n0.55 33.3 \n(33.3-\n33.3) 33.3 \n(33.3-\n66.6) 33.3 \n(33.3-\n66.6) 1   1   1   \n0.7 66.6 \n(33.3-\n66.6) 66.6 \n(66.6-\n66.6) 66.6 \n(66.6-\n66.6) 1   1   1   \n0.8 100 \n(66.6-\n100) 100 \n(66.6-\n100) 100 \n(66.6-\n100) 1   1   1   \n0.9 100 \n(100-\n100) 100 \n(100-\n100) 100 \n(100-\n100) 1   1   1   \n  \n  \nAll rights reserved. No reuse allowed without permission. \npreprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for thisthis version posted December 21, 2022. ; https://doi.org/10.1101/2022.12.20.22283736doi: medRxiv preprint \n\n \nTable 6. Responses with no treatment (positive control). Median escape response with interquartile range (IQR) \nfor each delivered volume (0.01, 0.15,0.30, 0.40, 0.55, 0.70, 0.80, and 0.90 mL) during the baseline, post-endo \nsurgery, and post-treatment periods, with Bonferroni-corrected p-values from Mann-Whitney U test for baseline \nperiod vs post-endo surgery period, post-endo surgery period vs post-treatment period, and baseline period vs post-\ntreatment period. * denotes Bonferroni-corrected p-values below significance threshold of 0.05.\n \n  \nVolume \nBaseline \nperiod \n(BL) \nmedian \nBL \nIQR \nPost-\nEndo \nperiod \n(PE) \nmedian \nPE \nIQR \nPost-\nTreatment \nperiod (PT) \nmedian \nPT \nIQR \nBaseline \nperiod vs Post-\nEndo \n period MWU \nperiod \nBonferroni-\nadjusted \np-values \nPost-Endo \nperiod vs Post-\nTreatment \nperiod \n MWU period \nBonferroni-\nadjusted \n p-values \nBaseline \nperiod vs Post-\nTreatment \nperiod \n MWU period \nBonferroni-\nadjusted \n p-values \n0.01 0 (0-0) 0 (0-0) 0 (0-0) 1   1   1   \n0.15 0 (0-0) 0 \n(0-\n33.3) 0 \n(0-\n33.3) 0.02 * 1   0.06   \n0.3 0 (0-0) 33.3 \n(33.3-\n66.6) 33.33 \n(33.3-\n66.6) 4E-05 * 1   2E-06 * \n0.4 0 \n(0-\n33.3) 66.6 \n(33.3-\n66.7) 66.6 \n(58.3-\n75) 4E-06 * 1   2E-05 * \n0.55 33.3 \n(33.3-\n66.6) 66.66 \n(66.6-\n100) 66.63 \n(66.6-\n100) 6E-06 * 1   2E-05 * \n0.7 66.6 \n(66.6-\n100) 100 \n(100-\n100) 100 \n(100-\n100) 7E-03 * 1   0.07   \n0.8 100 \n(75-\n100) 100 \n(100-\n100) 100 \n(100-\n100) 0.43   1   0.60   \n0.9 100 \n(100-\n100) 100 \n(100-\n100) 100 \n(100-\n100) 1   1   1   \n \nAll rights reserved. No reuse allowed without permission. \npreprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for thisthis version posted December 21, 2022. ; https://doi.org/10.1101/2022.12.20.22283736doi: medRxiv preprint \n\nConnectivity Map:\ngene expression profiles\n~22,000 genes\n1309 \ncompounds\nBulk Gene Expression Data: \n105 microarray samples from GEO \nnormalized and batch corrected using \njustRMA and ComBat\nDifferential expression signatures (disease vs. control): \nLimma  FDR < 0.05, log2FC > 1.1\nUn-\nstratified\nAll Samples\nStage Stratified\n1-2   3-4\nPhase Stratified\nPE   ESE  MSE\nFor all signatures: Rank-based, non-parametric search algorithm used to identify \ndisease-drug pairs with opposite transcriptomic effects (q-value < 0.0001)\nDisease\nDrugs\n299 unique drug hits identified\n221 common across all six \nsignatures\nEMR characterization and experimental validation of \nfenoprofen in animal model\nDrug Expression \nSimilar to Disease\nDrug Expression \nOpposite to Disease\nAll rights reserved. No reuse allowed without permission. \npreprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for thisthis version posted December 21, 2022. ; https://doi.org/10.1101/2022.12.20.22283736doi: medRxiv preprint \n\n203\n162\n72 70 65 53 42 30 28\n0\n100\n200\n300\n400\nIntersection Size\nPE\nStages I−II\nMSE\nUnstratified\nStages III−IV\nESE\n0 200 400 600\nSet Size\n384\n324\n256\n238\n200\n165\n124 114\n90\n0\n100\n200\n300\n400\nIntersection Size\nPE\nStages I−II\nUnstratified\nMSE\nStages III−IV\nESE\n0 500 1000 1500\nSet Size Set Size\nUpregulated Genes\n2240 total upregulated genes,\n238 (10.63%) overlap across all six signatures \nDownregulated Genes\n891 total downregulated genes,\n70 (7.86%) overlap across all six signatures \nOverlap in differentially expressed genes\nacross six endometriosis signatures\nAll rights reserved. No reuse allowed without permission. \npreprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for thisthis version posted December 21, 2022. ; https://doi.org/10.1101/2022.12.20.22283736doi: medRxiv preprint \n\nunstratified InII IIInIV PE ESE MSE\nadipiodone\nsertaconazole\nsulfamethoxazole\nprocyclidine\niopanoic acid\nmetolazone\nmedrysone\ntrifluoperazine\nbepridil\ncycloserine\ncloperastine\nflunisolide\nlevonorgestrel\nzuclopenthixol\nscopolamine\nprimaquine\nirinotecan\npromazine\nflumetasone\nfenoprofen\nReversal Scores for Top 20 Drug Candidates\nAcross All Endometriosis Signatures\nNo Reversal\nEffect (reversal\nscore > 0, or\nq−value > 0.0001)\nMaximum Reversal\n Effect (reversal\nscore ≤ 0)\n221\n38\n5 4 4 3 3 3 3\n0\n50\n100\n150\n200\n250\nIntersection Size\nMSE\nStages III−IV\nPE\nUnstratified\nStages I−II\nESE\n0100200300\nSet Size\nOverlap in drugs across six endometriosis signatures\n(sets with ≥ drugs shown)\nProtein Targets for Top 20 Drug Candidates\n Fenoprofen Drug Signature vs Disease Signatures\nFenoprofen Disease Signatures\nShared\nGenes\nDownregulated\nUpregulated\nUnstratifiedStage I - IIStage III - IVPE ESE MSE\nA. B.\nC. D.\nAll rights reserved. No reuse allowed without permission. \npreprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for thisthis version posted December 21, 2022. ; https://doi.org/10.1101/2022.12.20.22283736doi: medRxiv preprint \n\nA. B. C.\nAll rights reserved. No reuse allowed without permission. \npreprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for thisthis version posted December 21, 2022. ; https://doi.org/10.1101/2022.12.20.22283736doi: medRxiv preprint \n\nA. B.\nC. D.\nFenoprofen Ibuprofen\nControl: No Endo Surgery, No Treatment Control: No Treatment\nAll rights reserved. No reuse allowed without permission. \npreprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for thisthis version posted December 21, 2022. ; https://doi.org/10.1101/2022.12.20.22283736doi: medRxiv preprint","source_license":"CC0","license_restricted":false}