{"paper_id":"38ff511f-b7bb-45ee-a843-83cdfacf67e3","body_text":"OPEN\nEffect of hydroxychloroquine and characterization of\nautophagy in a mouse model of endometriosis\nA Ruiz1,7, S Rockfield 1,7, N Taran1, E Haller 2, RW Engelman 3, I Flores 4,5, P Panina-Bordignon 6 and M Nanjundan* ,1\nIn endometriosis, the increased survival potential of shed endometrial cells (which normally undergo anoikis) is suggested to\npromote lesion development. One mechanism that may alter anoikis is autophagy. Using an autophagic flux inhibitor\nhydroxychloroquine (HCQ), we identified that it reduces the in vitro survival capacity of human endometriotic and endometrial\nT -HESC cells. We also identified that HCQ could decrease lesion numbers and disrupt lesion histopathology, as well as increase\nthe levels of peritoneal macrophages and the IP-10 (10 kDa interferon- γ-induced protein) chemokine in a mouse model of\nendometriosis. We noted that RNA levels of a subset of autophagic markers were reduced in lesions relative to uterine horns from\nendometriosis-induced (untreated) mice. In addition, the RNA levels of autophagic markers were decreased in uterine horns of\nendometriosis-induced mice compared with those from controls. However, we noted that protein expression of LC3B (microtubule-\nassociated protein 1 light-chain 3 β; an autophagic marker) was increased in uterine horns of endometriosis-induced mice\ncompared with uterine horns of controls. By immunohistochemical staining of a human endometriosis-focused tissue microarray,\nwe observed LC3B expression predominantly in epithelial relative to stromal cells in both eutopic and ectopic endometria. Via\ntransmission electron microscopy, cells from eutopic endometria of endometriosis-induced mice contained more lipid droplets\n(rather than autophagosomes) compared with uterine horns from controls. Collectively, our findings indicate that the autophagic\npathway is dysregulated in both ectopic and eutopic endometrium in a murine model of endometriosis and that HCQ has potential\nas a therapeutic agent for women afflicted with endometriosis.\nCell Death and Disease (2016) 7, e2059; doi:10.1038/cddis.2015.361; published online 14 January 2016\nEndometriosis is a chronic, painful, and debilitating disease in\nwhich endometrium-like glandular and stromal cells grow out-\nside the uterine cavity.\n1,2 It is an inflammatory and estrogen-\ndependent disease that affects 6 –10% of women during their\nreproductive years and up to 50% of women receiving fertility\ntreatments.\n3 Sampson’s hypothesis (the most accepted\ntheory) states that shed endometrial tissue during menses\nreaches the peritoneal cavity by exiting the uterus through the\nfallopian tubes by retrograde menstruation.\n4–6 These shed\nendometrial cells survive, implant, and grow at ectopic\nlocations, developing into endometriotic lesions. 5,7\nEpithelial cells normally undergo anoikis, a mechanism of\nprogrammed cell death, upon detachment from the extra-\ncellular matrix. 8 One mechanism that we propose could\npotentially alter the anoikis response in endometrial cells is\nautophagy. This cellular pathway needs to be carefully\nregulated to maintain cellular homeostasis. 9,10 Under condi-\ntions of stress, changes in autophagic flux can lead to altered\ncellular survival. 9,10 Autophagy is a complex process that\nbegins with the formation of double-membrane vesicles,\ntermed autophagosomes, which engulf cytoplasmic compo-\nnents. For a comprehensive review of the autophagic pathway ,\nrefer to Feng et al .\n10 Briefly, autophagosomes fuse with\nlysosomes to degrade and recycle their cargo comprised of\noxidized proteins, lipids, and damaged organelles. Presently,\nthere is limited evidence that autophagy contributes to the\ndevelopment and progression of endometriosis. In a surgical\ninduction model of murine endometriosis, increased expres-\nsion of A TG9A, an autophagic mediator that is involved in\nvesicle formation,\n11 was detected in the eutopic endometria\nfrom endometriosis-induced mice. 12 In human endometrio-\nmas (ovarian endometriosis), there was a reduction of LC3-II\n(the conjugated form of LC3) protein compared with control\neutopic endometrial tissue.\n13 In contrast, an independent\nstudy reported that the protein expression of LC3-II was\nelevated, while p62 (which binds ubiquitinated cargo for\n1Department of Cell Biology, Microbiology, and Molecular Biology, University of South Florida, T ampa, FL, USA; 2Department of Integrative Biology, University of South\nFlorida, T ampa, FL, USA; 3Department of Pathology and Cell Biology, University of South Florida, T ampa, FL, USA; 4Department of Basic Science-Microbiology, Ponce\nHealth Sciences University and School of Medicine, Ponce Research Institute, Ponce, Puerto Rico; 5Department of Clinical Sciences and Ob-Gyn, Ponce Health Sciences\nUniversity and School of Medicine, Ponce Research Institute, Ponce, Puerto Rico and 6Reproductive Sciences Laboratory , Division of Genetics and Cell Biology, IRCCS\nOspedale San Raffaele, Milan, Italy\n*Corresponding author: Dr M Nanjundan, Department of Cell Biology, Microbiology, and Molecular Biology, University of South Florida, 4202 East Fowler Avenue, ISA2015,\nT ampa, FL 33620, USA. T el: +8139748133; Fax: +8139741614; E-mail: mnanjund@usf.edu\n7These authors contributed equally to this work.\nReceived 14.6.15; revised 10.11.15; accepted 12.11.15; Edited by GM Fimia\nAbbreviations: A TG, autophagy-related gene; BNIP3, BCL2/adenovirus E1B 19 kDa interacting protein; CK8, cytokeratin 8; EIF2AK3, eukaryotic translation initiat ion\nfactor 2- α kinase 3; ER α, estrogen receptor- α; FBS, fetal bovine serum; GABARAPL1, GABA(A) receptor-associated protein like 1; G-CSF, granulocyte colony-\nstimulating factor; GI, gastrointestinal tract; HCQ, hydroxychloroquine; IGF1, insulin-like growth factor 1; IP-10, 10 kDa interferon- γ-induced protein; IRGM1, immunity-\nrelated GTPase family M1; ITS, insulin transferrin selenium; LC3B, microtubule-associated protein 1 light-chain 3 β; PE, phosphatidylethanolamine; PBS, phosphate-\nbuffered saline; PRKAA1, protein kinase AMP-activated, α1 catalytic subunit; PR, progesterone receptor; PTEN, phosphatase and tensin homolog; SQSTM1,\nSequestosome 1; TEM, transmission electron microscopy; TMA, tissue microarray\nCitation: Cell Death and Disease (2016) 7, e2059; doi:10.1038/cddis.2015.361\n& 2016 Macmillan Publishers Limited All rights reserved 2041-4889/16\nwww.nature.com/cddis\n\nDC\nControlHCQ\nRelative Light Units\nC D\n****\n****\nLC3B-I\nLC3B-II\nLC3B-I\nLC3B-II\nShort Exposure\nLong Exposure\nControl\nHCQ\nControl\nHCQ\nC D\nPan-Actin\n17 kDa\n12 kDa\n17 kDa\n12 kDa\n52 kDa\n38 kDa\n***\nNumber of Endometriotic\nLesions per Mouse \nLesion Area (mm2)\nLesion volume (mm3)\nNS\nNS\nEstradiol Injection\nEndometriosis Induction\nand\nPBS or HCQ Treatment  \n420\nDonor Mice\n0\nPBS or HCQ\nTreatment\nSample\ncollection \nWeeks\nWeeks\n1\n3\nRecipient Mice\nAutophagy in a mouse model of endometriosis\nA Ruiz et al\n2\nCell Death and Disease\n\ndegradation) was decreased in ovarian endometriomas com-\npared with eutopic endometria of disease-free participants. 14\nHerein, our main aim was to provide further evidence for a\nrole of autophagy in endometriosis development. Specifically,\nwe sought to determine the therapeutic effects of a lysosomo-\ntropic agent and known autophagic flux inhibitor, hydroxy-\nchloroquine (HCQ),15–17 on human endometriotic cells and in\nan established mouse model of endometriosis. The results\npresented herein are of high clinical translational value, as we\nidentify a potential new non-hormonal treatment for this still\nincurable and common disease.\nResults\nHCQ alters human endometrial and endometriotic cell\nsurvival as well as lesion number and histopathology in a\nmouse model of endometriosis. T o assess whether an\nautophagic flux inhibitor could alter the survival capacity of\ncells isolated from human endometriotic lesions, we treated\nlife-extended human endometriotic cells (cells were derived\nfrom peritoneal ( ‘C’) and ovarian ( ‘D’) lesions obtained from\ntwo independent patients and were thus tested separately)\nwith 25 μM HCQ. This dose was selected based on our\nprevious studies. 18 As shown in Figures 1a and b, we\nobserved a marked reduction in cell survival of human\nendometriotic cells from two different types of lesions\n(Po0.0001) following 5 days of HCQ treatment. T o validate\nthe activity of HCQ, we performed western blot analysis for\nLC3B, which showed that LC3B-II increased with HCQ\ntreatment in these human endometriotic cells (Figure 1c). A\nsimilar reduction in cell survival and increase in LC3B-II\nprotein was noted in the T -HESC human endometrial stromal\ncells (derived from myoma, Supplementary Figure 1). T o\nconfirm the effect of autophagy inhibition, we performed\nsiRNA knockdown for A TG5, beclin-1, A TG7, PIK3C3\n(phosphatidylinositol 3-kinase, catalytic subunit type 3), and\nLC3B in our human endometrial and endometriotic cells.\nAlthough we failed to obtain suitable numbers of viable cells\nHCQPBS\nUterine HornLesion\nFigure 1 Continued.\nFigure 1 HCQ reduces endometriotic cell survival as well as lesion number and histopathology in a mouse model of endometriosis. ( a) Representative images of life-\nextended endometriotic cells using human endometriotic cells derived from two different lesion types:‘C’ and ‘D’ treated for 18 h with 25 μM HCQ. (b) Cell survival of life-extended\nendometriotic cells treated with 25 μM HCQ for 5 days was assessed by CellTiter-glo and measuring luminescence. (c) Cell lysate from life-extended endometriotic cells treated\nwith 25 μM HCQ for 18 h were analyzed by western blotting using the indicated antibodies. Three independent experiments were performed. (d) Schematic representation of the\nexperimental design. Mice were intraperitoneally injected with β-estradiol at 6 weeks of age. After 1 week, these mice were killed and their uterine horns were removed, minced,\nand injected into the peritoneal cavity of the same-age mice (recipients at 7 weeks of age). The same day of endometriosis induction, mice received an intraperitoneal injection of\nHCQ or PBS. A second dose was administered 1 week later. T wo weeks after induction, mice were killed and samples were collected. (e) The lesion numbers, area, and volume\nper mouse are shown for HCQ- and PBS-treated endometriosis-induced mice. A subset of lesions (PBS- (n = 12) and HCQ- (n = 10) treated mice) was measured lengthwise and\nwidthwise to determine the area and volume. (f) Uterine horns and lesions from PBS- and HCQ-treated mice were subjected to H&E staining. Black arrowheads indicate glandular\ncompartments (top panels). Black arrows indicate epithelial cells (bottom panels). All images were captured at 10 × magnification\nAutophagy in a mouse model of endometriosis\nA Ruiz et al\n3\nCell Death and Disease\n\nupon transfection of human endometriotic cells for further\nanalysis, we successfully obtained 490% knockdown\nefficiency of the above-described autophagic mediators in\nT -HESC cells (Supplementary Figure 1d). Interestingly,\nprotein levels of p21 (a cyclin-dependent kinase inhibitor\ninvolved in cell cycle arrest) markedly increased with siRNA\ntargeting A TG7 and to a lesser degree with LC3B and beclin-\n1. Therefore, we selected these autophagic mediators to\ninvestigate their effects on cell viability using the CellTiter-glo\nassay in these cells. As shown in Supplementary Figure 1e,\nthe cell viability of T -HESC was particularly reduced with\nA TG7 knockdown (and to a lesser degree with LC3B and only\nslightly with beclin-1). These observations suggest that the\nuse of HCQ (or targeting specific autophagic mediators) may\nbe detrimental to both human endometrial and endometriotic\ncell survival.\nT o determine whether treatment with HCQ alters the\nformation of endometriotic lesions, we used an induced model\nof murine endometriosis in which mice receive injections of\nuterine horn fragments that develop into lesions within 2\nweeks.19,20 We treated recipient (endometriosis-induced)\nmice with 60 mg/kg HCQ 21 or phosphate-buffered saline\n(PBS; Figure 1d). This treatment was repeated once every\n7 days post-induction. Mice that were neither injected with\nuterine horn fragments nor treated were used as controls\n(Supplementary Figure 2a). All of the mice were killed at\nthe same time (14 days after endometriosis induction for both\nthe PBS and HCQ treatment groups). Ectopic lesions that\ndeveloped in the recipient mice (white arrow; Supplementary\nFigure 2b) were counted, measured, and collected for RNA\nand protein analysis, as well as for histological staining. No\nlesions were observed in the control group (labeled as N;\nSupplementary Figure 2b). The majority (87.5%) of\nendometriosis-induced mice developed lesions. At the time\nof collection, we noted that the endometriotic lesions varied in\nsize, color, and location across the treatment groups. As\nshown in Figure 1e, there was a significant reduction in the\nnumber of lesions that developed in mice treated with HCQ\ncompared with those treated with PBS ( P = 0.0007; PBS-\ntreated mice, n = 24 (with a total of 46 lesions) and HCQ-\ntreated mice, n = 25 (with a total of 18 lesions)). However,\nthere was no significant difference in lesion size or volume\nbetween these two treatment groups (Figure 1e).\nA randomly selected subset of the collected lesions and\nuterine horns were processed for staining with hematoxylin\nand eosin (H&E) (pathologically confirmed endometriotic\nlesions from PBS- and HCQ-treated mice, n = 15 each; uterine\nhorns derived from PBS-treated mice, n = 10; and uterine\nG-CSF Eotaxin IP-10\n***\npg/ml\nG-CSF Eotaxin IP-10\npg/ml\nNormal\nRecipient\n% of Total Cells\nPBS\nHCQ\n% of Total Cells\n**\nFigure 2 HCQ treatment increases the numbers of peritoneal macrophages and chemokine levels of IP-10. ( a) Peritoneal fluid collected for control and recipient mice were\nanalyzed for 32 cytokines/chemokines. The data are presented as a dot plot (showing individual sample values), and the line indicates the average ± S.E.M. (b) Peritoneal fluid\nwas collected from HCQ- and PBS-treated mice for cytokine/chemokine analysis. The data are presented as a dot plot (showing individual sample values) and the line represents\naverage ± S.E.M. (c) Macrophages were stained with CD11b and F4/80 antibodies and then analyzed by flow cytometry . Representative images of the raw flow cytometry data\nare shown. The data are presented as a dot plot (showing individual sample values) and the line represents average ± S.E.M. (d) Macrophages were collected from the same\nspecimens analyzed in (b). Macrophages were stained as described in (c). Representative images of the raw flow cytometry data are shown. The data are presented as a dot plot\n(showing individual sample values) and the line represents average ± S.E.M\nAutophagy in a mouse model of endometriosis\nA Ruiz et al\n4\nCell Death and Disease\n\nhorns derived from HCQ-treated mice, n = 10). Interestingly,\nas shown in Figure 1f, we observed an irregular epithelium\npattern in 5 out of 10 uterine horns derived from HCQ-treated\nmice compared with those derived from PBS-treated mice. In\naddition, we noted that the ectopic growths from HCQ-treated\nmice did not histologically resemble endometriotic lesions\n(i.e., did not contain the expected glandular components),\nwhereas those treated with PBS did (Figure 1f, black arrow-\nheads indicate glandular compartments, while black arrows\nindicate epithelial cells within the lesions) ( P = 0.03, per\nFisher’s exact test). T aken together, these results reveal that\nHCQ reduces the number of endometriotic lesions and alters\nthe cellular organization within these tissues.\nAltered levels of peritoneal macrophages and IP-10\ncytokine from HCQ-treated mice. T o investigate changes\nin the inflammatory response to endometriosis, we quantified\n32 cytokines/chemokines in the peritoneal fluid collected from\ncontrol ( n = 4) and recipient ( n = 3) mice using a mouse\ncytokine and chemokine magnetic bead panel assay . Of the\n32 analyzed cytokines/chemokines, we identified that G-CSF ,\neotaxin, and IP-10 (10 kDa interferon- γ-induced protein; also\nknown as CXCL10) were within the sensitivity and detection\nlimits of the assay; however, there were no significant\ndifferences in these proteins between control and recipient\nmice (Figure 2a). In contrast, we identified that IP-10 was\nsignificantly increased ( P = 0.0079) in peritoneal fluid\nUterine Horns: PBS Uterine Horns: HCQ\nH&ECK8VimentinER αPRLC3B\nOvaries: PBS Ovaries: HCQ\nFigure 3 Immunohistochemical analyses of murine endometria, ovaries, and lesions. ( a) Representative immunohistochemical images of uterine horns and ovaries from\nPBS- and HCQ-treated mice are shown. The cores were processed for H&E staining as well as epithelial and stromal markers (CK8 and vimentin, respectively), ovarian hormone\nreceptors (ER α and PR), and the autophagy marker LC3B. (b) Representative immunohistochemical images are shown from lesions collected from PBS- and HCQ-treated mice.\nThe sections were stained as described in (a). (c) Representative images of antibody immunohistochemical staining controls (both positive and negative staining) are shown. For\nLC3B, mouse brain was used as a positive staining control tissue. For PR and ER α, mouse mammary glands were used as positive staining control tissues. For vimentin and\nCK8, mouse uterine horns were used as positive staining control tissues. Negative staining controls were performed in the absence of primary antibod y . The images of PR-\npositive and -negative staining controls are shown at 20 × magnification; all other images are shown at 10 × magnification\nAutophagy in a mouse model of endometriosis\nA Ruiz et al\n5\nCell Death and Disease\n\nobtained from HCQ-treated mice ( n = 5) compared with PBS-\ntreated mice ( n = 5), whereas G-CSF and eotaxin remained\nunchanged (Figure 2b).\nAs previous research identified a significant increase in\nmacrophage numbers in endometriosis-induced mice 22 and\nthat HCQ can inhibit cytokine production in human macro-\nphages,23 we therefore assessed macrophage numbers in\ncontrol, recipient (untreated), PBS-treated, and HCQ-treated\nrecipient mice. There was no significant change in the\nmacrophage numbers present in the peritoneal cavity of\ncontrol and endometriosis-induced mice at the time of\nsample collection (2 weeks post-induction), using the cano-\nnical macrophage markers CD11b and F4/80 (Figure 2c).\nHowever, we did find a significant increase ( P = 0.0079) in\nmacrophage numbers in HCQ-treated mice compared with\nPBS-treated mice (Figure 2d). These data indicate that HCQ\nalters the inflammatory response of endometriosis-\ninduced mice.\nHCQ induces cellular disorganization in murine endome-\ntriotic lesions and eutopic endometrial. T o determine\nwhether HCQ treatment alters the histopathology (tissue\norganization) of the recipient ’s uterine horns and other\ntissues, we developed a murine tissue microarray (TMA)\ncomprised of 113 cores and performed H&E as well as\nimmunohistochemical staining. The TMA contained uterine\nhorns and ovaries from 10 PBS- and 10 HCQ-treated mice,\nas well as a mammary gland, a kidney , a lymph node, and a\nsmall intestine from a PBS-treated mouse for use as antibody\ncontrols. Based on H&E staining, we observed that the\nluminal epithelium of the uterine horn endometrium from\nHCQ-treated mice had an irregular pattern (Figure 3a).\nHowever, vimentin and cytokeratin 8 (CK8) appeared to be\nappropriately localized to the stromal and epithelial compart-\nments, respectively, independently of HCQ treatment. As\nexpected, estrogen receptor α (ER α) was primarily localized\nto the epithelial cell layer of the endometrial glands, whereas\nprogesterone receptor (PR) appeared to be evenly distributed\nbetween the stromal and epithelial cell compartments; 24\nhowever, the PR staining was comparatively much weaker to\nthat for ER α. Again, no differences were noted in the tissues\nfrom PBS- and HCQ-treated mice for ER α and PR staining\npattern or intensity. LC3B expression appeared more intense\nin HCQ-treated mice relative to PBS-treated mice in both the\nstromal and epithelial compartments (Figure 3a). We also\nstained for the same immunohistochemical markers in\novaries, but we did not observe any marked differences\nin the intensity or localization pattern of any of these proteins\nLesion: PBS Lesion: HCQ\nLC3B CK8VimentinER αPR\nNegativePositive\nLC3B PR ER α Vimentin CK8\nFigure 3 Continued.\nAutophagy in a mouse model of endometriosis\nA Ruiz et al\n6\nCell Death and Disease\n\nLC3B\nPBS HCQ\nNS\nNS\nNS\nNS\nBeclin-1\n****\nNS NS\nNS\nATG7\nNS\nNS\nNS\nNS ULK1\nNS\nNS NS\nNS\nPIK3C3\nNS\nNS\nNS\nNS\nATG5\n*\nNS\nNS\n*\nATG4B\n***\nNS\nNS\nNS\nATG9A\nNS\nNS\nNS\nNS\nPBS HCQ\nATG3\nNS\nNS\nNS\n*\nATG2B\n*\nNS\nNS\nNS\nPBS HCQ\nPBS HCQ\nRNA-Fold Change RNA-Fold Change RNA-Fold Change\nUterine Horns: PBS\nLesions: PBS\nUterine Horns: HCQ\nLesions: HCQ\nPan-Actin\n52 kDa\n38 kDa\nUH-1\nUH-2\nUH-3\nUH-4\nLL-3\nSL-4\nUHL-4\nUterine Horns Lesions\nPBS-treated\nUH-1\nUH-2\nUH-3\nUH-4\nUH-5\nFRL-1\nML-2\nUterine Horns Lesions\nHCQ-treated\n17 kDa\n12 kDa\nLC3B-I\nLC3B-II\nFBFL-5\nFM-3\nUH-5\np6252 kDa\n36 kDa\np6252 kDa\n36 kDa\nFOXO176 kDa\n52 kDa\nFOXO176 kDa\n52 kDa\n102 kDa\n102 kDa\n17 kDa\n12 kDa\nLC3A-I\nLC3A-II\nAMPK α\n76 kDa\n52 kDa\nGABARAPL117 kDa\n12 kDa\nShort Exposure\nLong Exposure\nShort Exposure\nLong Exposure\nGABARAPL1\n17 kDa\n12 kDa\nShort Exposure\nLong Exposure\nFigure 4 Autophagy gene expression and protein levels are decreased in uterine horns and lesions from endometriosis-induced mice, independently of HCQ trea tment.\n(a) A subset of samples was analyzed to quantify transcript levels of autophagic markers by real-time PCR. The line indicates average ± S.E.M. (b) Protein expression was\nassessed by western blot analyses across the indicated groups using the indicated antibodies. Pan-actin was used as a loading control. The selected w estern blot presented in\nthe panel is representative of the data obtained across all of these specimens analyzed and includes: (1) PBS-treated mice: uterine horns ( n = 5); (2) HCQ-treated mice: uterine\nhorns (n = 5); (3) PBS-treated mice: lesions (n = 5); and (4) HCQ-treated mice: lesions (n = 2). UH, uterine horns; FM, mass located near fat; LL, lesion located on the liver; SL,\nlesion located near the surface of the peritoneal cavity; UHL, lesion located near the uterine horn; FBFL, blood-filled lesion located near the fat; F RL, red lesion located near the\nfat; ML, lesion located near the mesentery . ( c) Densitometric analyses of the presented western blots (as presented in ( b)) are shown (presented as average ± S.E.M.). For\nGABARAPL1 and p62, analysis of the short exposure is shown. For FOXO1, analysis of the long exposure is shown\nAutophagy in a mouse model of endometriosis\nA Ruiz et al\n7\nCell Death and Disease\n\nin these tissues from HCQ-treated mice relative to those from\nPBS-treated mice.\nA certified pathologist confirmed epithelial and stromal\ncomponents in the lesions analyzed. Lesions (independent\nblocks and not on the above-described TMA (see Materials\nand Methods)) were also immunostained for CK8, vimentin,\nER α, and PR, and LC3B (Figure 3b). The epithelial cells of the\nglands were positive for CK8, ER α, and PR expression, which\nprovides supporting data that the collected lesions originated\nfrom endometrial tissue (Figure 3b). Interestingly, there was an\nabsence of glandular components in four out of the seven\nstained lesions from HCQ-treated mice as demonstrated by\nCK8, ER α, and PR immunohistochemical staining. These\nresults (with the H&E data presented in Figure 1f) suggest that\nHCQ alters the organization of ectopic growths in the murine\nmodel of endometriosis. Figure 3c displays representative\nimages of positive and negative staining controls for the\nantibodies used.\nInduction of endometriosis downregulates mRNA and\nprotein expression of autophagic markers in ectopic\ncompared with eutopic murine endometrium. T o deter-\nmine whether the expression of autophagic mediators in\nuterine horns and lesions differs between PBS- and HCQ-\ntreated mice, we used real-time PCR to quantify the mRNA\ntranscript levels of 10 major molecules involved in the\nautophagic pathway . In PBS-treated animals, we determined\nthat the mRNA levels of A TG5 ( P = 0.0294), A TG4B\n(P = 0.0004), A TG2B (P = 0.0440), and beclin-1 ( Po0.0001)\nwere significantly decreased in the analyzed endometriotic\nlesions compared with uterine horns (uterine horns from\nPBS-treated mice, n = 14; lesions from PBS-treated mice,\nn = 28; uterine horns from HCQ-treated mice, n = 15; and\nlesions from HCQ-treated mice, n = 7) (Figure 4a). Owing to\nlimited sample availability, LC3B and A TG2B were analyzed\nusing a smaller number of samples (i.e., for LC3B: uterine\nhorns from PBS-treated mice, n = 9; lesions from PBS-\ntreated mice, n = 18; uterine horns from HCQ-treated mice,\nn = 10; and lesions from HCQ-treated mice, n = 4; For\nA TG2B: uterine horns from PBS-treated mice, n = 5; lesions\nfrom PBS-treated mice, n = 10; uterine horns from HCQ-\ntreated mice, n = 5; and lesions from HCQ-treated mice,\nn = 2). No significant differences were noted between lesions\nand uterine horns from HCQ-treated C57BL/6 mice (likely\ndue to smaller lesion numbers available in the HCQ group,\nalthough similar trends were apparent). However, lesions\nobtained from these HCQ-treated mice had a significant\nincrease in A TG5 ( P = 0.0499) and A TG3 ( P = 0.0248)\ncompared with lesions from PBS-treated mice (Figure 4a).\nWe also used the Balb/c mouse strain for the induction\nmodel to demonstrate that the changes observed in autop-\nhagy gene expression are independent of the mouse genetic\nstrain used.20 In this model, we identified that the mRNA levels\nof A TG7 ( P = 0.0174), A TG4B ( P = 0.0020), and beclin-1\n(Po0.0001) were significantly reduced in endometriotic\nlesions ( n = 8) compared with uterine horns ( n = 8) derived\nfrom the same recipient mice (Supplementary Figure 3).\nWe next analyzed the protein levels of autophagic markers\nin both lesions and uterine horns from PBS- and HCQ-treated\nmice (Figure 4b) from the following groups: (1) PBS-treated\nmice: uterine horns ( n = 15); (2) HCQ-treated mice: uterine\nhorns ( n = 15); (3) PBS-treated mice: lesions ( n = 10); and\n(4) HCQ-treated mice: lesions ( n = 7). LC3B-I, LC3B-II, LC3A-\nI, and LC3A-II were decreased in lesions compared with\nuterine horns from both PBS- and HCQ-treated mice.\nExpression of GABARAPL1-I (GABA(A) receptor-associated\nprotein like 1) was detected in uterine horns collected from\nboth groups of treated mice and was decreased in the lesions;\nhowever, the conjugated form, GABARAPL1-II, was not\nobserved in any of the murine specimens. We also observed\na decrease in p62 in endometriotic lesions relative to uterine\nhorns that was independent of HCQ treatment. FOXO1 and\nAMPKα protein levels in the uterine horns were variable\namong the samples analyzed, although they were both\nreduced within the lesions (Figures 4b and c). T o determine\nwhether HCQ treatment altered the expression of autophagic\nLC3B-IILC3B-I\nNormalized Protein\nExpression \n*\nLC3A-IILC3A-I\nGABARAP p62\nNormalized Protein\nExpression \nFOXO1\nPBS HCQ\nNS\nNS\nNS\nNS\nNS\nNS\nNS\n**\nNS\nNS\nNS\n*\nNS\nNS\nNS\nNS\nNS\nNS\nNS\nNS\nNS\nNS\nNS\nNS\nNS\nNS\nNS\nAMPK α\nNS\nNS\nNS\nNS\nPBS HCQ PBS HCQ PBS HCQ\n3\n2\n1\n0\n3\n2\n1\n0\n3\n2\n1\n0\n3\n2\n1\n0\n3\n2\n1\n0\n3\n2\n1\n0\n3\n2\n1\n0\n3\n2\n1\n0\n-1\nFigure 4 Continued.\nAutophagy in a mouse model of endometriosis\nA Ruiz et al\n8\nCell Death and Disease\n\nmediators in other organs, we harvested various tissue\nspecimens (kidneys, thymus, spleen, lung, pancreas, heart,\nand liver) from each treatment group (five PBS-treated mice\nand five HCQ-treated mice) and assessed LC3B levels\n(Supplementary Figure 4). Among these tissues, only the\nlung and heart showed differences in LC3B-II expression\nfollowing HCQ treatment. Overall, these results suggest\nthat the protein expression of autophagic mediators is\ndysregulated in endometriotic lesions and is not affected by\ntreatment with HCQ in the majority of tissues analyzed,\nincluding uterine horns.\nRNA expression of autophagic markers is dysregulated\nin eutopic endometria upon induction of endometriosis.\nEvidence is accumulating that the eutopic endometria from\npatients with endometriosis differs markedly from the eutopic\nendometria from endometriosis-free subjects. 25,26 T o identify\nchanges in the expression of key autophagic markers in this\ncontext, we used an RT 2-PCR autophagy focused profiler\narray to analyze RNA isolated from uterine horns from control\n(non-induced) and recipient (untreated). In addition, we\ncompared the uterine horns from recipient mice with those\nfrom PBS-treated recipient mice to verify that there was no\nsignificant change that occurred upon intraperitoneal injection\nwith PBS. Three representative samples were selected from\neach group based on RNA quality. A heat map comparing\ngene expression in RNA isolated from uterine horns from\ncontrol mice to recipient mice is shown in Figure 5a; the\nresults indicate that there is a subset of autophagy genes that\nis differentially expressed. A volcano plot is shown in\nFigure 5b that displays the fold changes in autophagy genes\nin eutopic endometria between recipient and control mice.\nWe identified 13 dysregulated genes (with statistical sig-\nnificance) between these two groups of samples. Insulin-like\ngrowth factor 1 (IGF1) was the only autophagic marker that\nwas significantly increased ( P = 0.044); the remaining 12\nBNIP3\nATG9B\nIncreased ExpressionDecreased Expression\nLC3A\nIGF1\nLC3B\nPRKAA1\nATG4C\nFAS\nIRGM1\nEIF2AK3\nPTEN\nGABARAPL1\nSQSTM1\nDRAM2\nB2M\nFigure 5 Decreased RNA expression of autophagic markers in eutopic endometria from mice with endometriosis relative to controls. ( a) The selected RNA samples from\nnon-induced (control) and endometriosis-induced mice were analyzed using an RT2-PCR profiler array specific for 84 autophagy-related genes. A heat map shown depicting the\nmeasured CT values across these specimens is presented. ( b) A volcano plot is shown from the analyzed data presented in ( a). The horizontal axis indicates significance\n(P = 0.05) if targets are above the line; the dotted vertical bars denote at least a two-fold change in gene expression if targets are ⩾ − 1 or 1 Log2 units. Red arrows indicate the\nautophagic markers that were decreased 42-fold with a P-value of o0.05. (c) Validation using T aqMan real-time PCR of specimens used in ( a)\nAutophagy in a mouse model of endometriosis\nA Ruiz et al\n9\nCell Death and Disease\n\nmarkers were all significantly decreased (BNIP3 (BCL2/\nadenovirus E1B 19 kDa interacting protein), P = 0.015;\nA TG9B, P = 0.015; LC3A, P = 0.007; LC3B, P = 0.0012;\nprotein kinase AMP-activated, α1 catalytic subunit\n(PRKAA1), P = 0.023; A TG4C, P = 0.031; FAS, P = 0.003;\nIRGM1 (immunity-related GTPase family M1), P = 0.025;\nGABARAPL1, P = 0.045; PTEN (phosphatase and tensin\nhomolog), P = 0.048; EIF2AK3 (eukaryotic translation initia-\ntion factor 2- α kinase 3), P = 0.043; and SQSTM1 (sequesto-\nsome 1), P = 0.054). As shown in Supplementary Figure 5,\nwe did not observe any significant changes upon PBS\ntreatment in the RT 2-PCR array.\nT o validate these‘top hits’ (i.e., increased by at least two-fold\nwith Po0.05) identified from the autophagic pathway\nRT2-PCR profiler array, we performed real-time PCR using\nT aqMan FAM-labeled probes/primers (Figure 5c and\nSupplementary T able 1). Using this approach, we validated\n10 of the 13 ‘top hits’ (Figure 5c): A TG4C (P = 0.0167), A TG9B\n(0.0113), EIF2AK3 ( P = 0.0068), FAS ( P = 0.0034), LC3A\n(P = 0.0306), LC3B ( P = 0.0040), GABARAPL1 ( P = 0.0360),\nPTEN ( P = 0.0295), SQSTM1 ( P = 0.0008), and PRKAA1\n(P = 0.0065) were significantly reduced. Although the majority\nof the tested autophagic markers were not significantly\nchanged upon PBS treatment relative to recipient (untreated),\nwe did identify that the expression of EIF2AK3 ( P = 0.0014)\nwas increased (Figure 5c). T aken together, these data suggest\nthat autophagy is dysregulated in the eutopic endometria of\nendometriosis-induced mice.\nIncreased LC3 protein and lipid droplets in eutopic\nendometria of endometriosis-induced mice compared\nwith eutopic endometria of controls. T o determine\nwhether the RNA level changes of key autophagic markers\nobserved between the eutopic endometria of endometriosis-\ninduced mice (n = 10) and non-induced (control) mice ( n = 10)\ntranslated to protein level changes, we assessed their protein\nlevels via western blot analyses. As shown in Figure 6A\nand Supplementary T able 2, beclin-1 (2.20-fold change,\nP = 0.0330), LC3B-I (4.00-fold change, P = 0.0185), LC3B-II\n(6.76-fold change, P = 0.0364), LC3A-II (1.97-fold change,\nP = 0.0135), and GABARAPL1 (1.95-fold change,\nP = 0.0334) were significantly increased in uterine horns\nfrom endometriosis-induced mice relative to those from\ncontrol mice. LC3A-I and LC3B-I have an expected molecular\nweight of ~ 16 kDa, whereas LC3A-II and LC3B-II have an\nexpected molecular weight of ~ 14 kDa. 27 When we assessed\nGABARAPL1 expression, we did not detect the conjugated\nform, suggesting that the primary form expressed in these\nATG4C\n*\n*\nATG9B\n*\n*\nBNIP3\nNS\nNS\nEIF2AK3\n**\n*\nFAS\n**\n**\nLC3A\n*\n*\nLC3B\n**\n**\nGABARAPL1\n*\n**\nIGF1 NS\n**\nIRGM1\nNS\n**\nSQSTM1\n***\n***\nPRKAA1\n**\n**\nPTEN\n*\nNS\n**NS NS\nNS NS\nNS\nNS\nNS NS NS NS\nNS\nNS\nRNA-Fold Change RNA-Fold Change RNA-Fold Change RNA-Fold Change\nFigure 5 Continued.\nAutophagy in a mouse model of endometriosis\nA Ruiz et al\n10\nCell Death and Disease\n\ntissues is the cytosolic form (GABARAPL1-I). T o assess if the\nincreased levels of LC3B were specific to the uterine horns in\nthe endometriosis-induced mice, we analyzed LC3B protein\nlevels in homogenates prepared from kidneys, thymus,\nspleen, lung, pancreas, heart, liver, and ovaries from both\nrecipient ( n = 3) and control ( n = 4) mice. Out of the nine\ntissues analyzed, only the left kidney , spleen, and liver\nappeared to show differences in LC3B-II levels\n(Supplementary Figure 6).\nT o test whether the observed increases in LC3A and LC3B\ncorrelated with an increase in autophagosome formation in\nthe eutopic endometria of endometriosis-induced mice, we\nperformed transmission electron microscope (TEM)\n(Figure 6B). Although no autophagosomes were identified in\neutopic endometria from control mice (Figure 6Ba –d) and\neutopic endometria from endometriosis-induced mice (Figure\n6Ce–g), we observed an increase in lipid droplet numbers in\nthe epithelial cells of eutopic endometria from endometriosis-\ninduced mice. In addition, we also observed more ‘unhealthy’\nelectron-dense epithelial cells in uterine horns from\nendometriosis-induced mice (Figure 6Ce) compared with\ncontrol mice (Figure 6Ba). T aken together, these results\nsuggest that expression of autophagic mediators (i.e., LC3) is\ndysregulated in the eutopic endometria of endometriosis-\ninduced mice, which is associated with an accumulation of\nlipid droplets in the epithelial cells.\nImmunohistochemical staining of LC3B in the epithelium\nand stromal components of eutopic and ectopic endo-\nmetrium in patients with endometriosis. We next\naddressed the cellular localization of LC3B within human\neutopic and ectopic endometrium by applying an immuno-\nhistochemical approach using a human endometriosis and\nendometrium TMA. 28 Representative immunohistochemical\nimages for endometrium (controls and patients) and lesions\n(fallopian tubes, ovaries, peritoneal, gastrointestinal, and\nskin) are shown (Figure 7a). We noted that LC3B was\nlocalized primarily to the epithelium, although staining was\nalso noted in the stroma. T o quantify the intensity of LC3B\nexpression at these specific cellular locations, we segmented\nthe sections using the H-score system into strong, moderate,\nweak, or no expression (Figures 7b and c). The proportion of\nstrong expression was elevated in the epithelial cells of the\nproliferative endometrium from cases (40.6%) and those from\nBeclin-1\nBeclin-1\nPan-Actin\nLC3B-I\nLC3B-II\nUterine Horn-1\nUterine Horn-2\nUterine Horn-3\nUterine Horn-4\nUterine Horn-5\nUterine Horn-6\nUterine Horn-9\nUterine Horn-10\nUterine Horn-1\nUterine Horn-2\nUterine Horn-3\nUterine Horn-4\nUterine Horn-5\nUterine Horn-6\nUterine Horn-9\nUterine Horn-10\nLC3B-I\nLC3B-II\nLC3B-I\nLC3B-II\nControl Mice Recipient Mice\n76 kDa\n52 kDa\n76 kDa\n52 kDa\n38 kDa\n52 kDa\n17 kDa\n12 kDa\n17 kDa\n12 kDa\n17 kDa\n12 kDa\nGABARAPL1\n17 kDa\n12 kDa\nGABARAPL1\n17 kDa\n12 kDa\nShort Exposure\nLong Exposure\nLC3A-I\n17 kDa\n12 kDa LC3A-II\n54 kDa AMPK α\nLong Exposure\nMedium Exposure\nShort Exposure\n52 kDa\n38 kDa\n52 kDa\n38 kDa\np62\np62\nLong Exposure\nShort Exposure\nLong Exposure\nShort Exposure\n*\nLC3B-I\nBeclin-1\n*\np62\nNS\n*\nLC3A-II\nGABARAPL1\n*\nAMPKαα\nNS\nNormalized Protein\nExpression \nNormalized Protein\nExpression \nNormalized Protein\nExpression \n*\nLC3B-II\nNormalized Protein\nExpression \nLC3A-I\nFigure 6 Increased protein expression of autophagy markers in the eutopic endometria from endometriosis-induced mice relative to controls. ( A) Tissue homogenates from\nuterine horns from non-induced and endometriosis-induced mice were analyzed for protein expression by western blotting (left panels). Pan-actin was used as the loading control.\nA representation for 8 out of 10 total samples for both control and recipient groups is shown. Densitometric analyses of the presented western blots (r ight panels) are shown\n(presented as average ± S.E.M.). (B) Uterine horns from control and endometriosis-induced mice were analyzed by TEM and representative images of epithelial cells are shown.\n(a–d) Eutopic endometria from control mice; (e–g) eutopic endometria from endometriosis-induced mice. The images on the right are magnifications of the indicated boxed region\nin the respective left image\nAutophagy in a mouse model of endometriosis\nA Ruiz et al\n11\nCell Death and Disease\n\novarian and fallopian tube lesions (38.8% and 38.0%,\nrespectively). The endometriotic tissue with the highest\nproportion of strong stromal expression was the gastrointest-\ninal tract (GI) (17.4%), followed by proliferative endometrium\nfrom controls (14.1%), proliferative endometrium from endo-\nmetriosis patients (13.6%), and secretory endometrium from\ncontrols (12.0%) (Figures 7b and c). We noted a significant\ndifference in LC3B expression in the epithelium of secretory\nendometrium compared with proliferative endometrium\n(P = 0.0193) (Supplementary Figure 7, lower panel). We also\nfound significantly increased expression in the epithelium of\nfallopian tube and ovarian endometriotic lesions compared\nwith epithelium from the secretory endometrium of controls\n(P = 0.0220 and P = 0.0097, respectively). In the stroma of\nperitoneal endometriotic lesions, LC3B was decreased\ncompared with the stroma of proliferative endometrium from\ncontrols ( P = 0.0101). In addition, relative to the stroma,\npositive LC3B immunostaining was significantly more\nFigure 6 Continued.\nAutophagy in a mouse model of endometriosis\nA Ruiz et al\n12\nCell Death and Disease\n\nelevated in the epithelial component of the lesions in the\nfallopian tube, ovarian, and peritoneum but not in lesions\nderived from the gastrointestinal tract and the skin\n(Supplementary Figure 7, upper panel). Thus, collectively,\nLC3B expression and localization was predominant in the\nepithelium relative to the stromal components in all tissue\ntypes assessed.\nDiscussion\nT o our knowledge, our work presented herein is the first to\ninvestigate the effect of an autophagic flux inhibitor, HCQ, on\nendometriotic lesion development and histopathology. Our\nmain findings are summarized in Figure 8. We found that HCQ\ntreatment affected the survival of human endometrial and\nendometriotic cells in vitro , as well as decreased lesion\nnumbers using an in vivo mouse model of endometriosis. The\ndrug also appeared to have an effect on lesion histopathology\n(the absence of glandular components), but not on lesion size.\nWe also identified that HCQ increases the number of\nmacrophages and the IP-10 chemokine within the peritoneal\ncavity of a mouse model for endometriosis. Furthermore, we\nhave identified that autophagic markers are differentially\nexpressed in uterine horns from endometriosis-induced mice\ncompared with those from control mice. Although we noted\nthat LC3B protein level was increased in eutopic endometria of\nendometriosis-induced mice (compared with controls), we did\nnot identify increased autophagosomes by TEM in these\ntissues. However, TEM showed that endometria from experi-\nmental mice are less healthy and contained an increased\nnumber of lipid droplets compared with endometria from\ncontrol mice. Finally, we noted that LC3B was expressed and\nProliferative: Controls\nSecretory: Controls\nProliferative: Patients\nOvary\nEndometriosis\nFallopian tube\nPeritoneal\nSkin\nGastrointestinalAntibody: Positive control\nAntibody: Negative control\nEndometrium \nControls Patients \nControls Patients \n% Tissue Intensity in Each\nCategory \n% Tissue Intensity in Each\nCategory \nNone Weak Moderate Strong\nStroma\nEpithelium\nFigure 7 LC3B is predominantly localized in the epithelium of ectopic and eutopic endometrium. (a) LC3B immunostaining was performed on a TMA containing biopsy cores:\neutopic endometrium (from controls and endometriosis patients) and endometriotic lesions (from ovaries, fallopian tubes, peritoneal, gastrointestinal, and skin). Mammary glands\nwere used as a positive control tissue and the negative antibody staining control was performed in the absence of primary antibody . All images were cap tured at × 20\nmagnification. LC3B immunostaining was analyzed using the H-score system. The average immunostaining in the ( b) stromal and (c) epithelial compartments were categorized\nas strong, moderate, weak, and no expression\nAutophagy in a mouse model of endometriosis\nA Ruiz et al\n13\nCell Death and Disease\n\nlocalized predominantly to the epithelia in all tissue types\n(relative to the stroma) of human endometrium and endome-\ntriotic lesions from diverse locations using a TMA approach.\nChloroquine, and derivatives of this agent (including HCQ),\nhave been used to treat malaria, as well as inflam-\nmatory and autoimmune diseases. 29 Endometriosis\nshares some characteristics with autoimmune disorders,\nsuch as increased inflammation, altered tissue-remodeling\ncomponents, dysregulated cell death pathways, increased\nlocal and systemic cytokine levels.30 Inflammatory changes in\nthe peritoneal cavity may be associated with lesion\ndevelopment.31 Other co-morbid autoimmune disorders (i.e.,\nsystemic lupus erythematosus, rheumatoid arthritis, multiple\nsclerosis) can coincide with the presence of end-\nometriosis.32,33 These autoimmune disorders can be treated\nwith HCQ, which can antagonize communication among cells\nin the immune system that are inappropriately activated. 29,34\nTherefore, HCQ could potentially be used as an effective\ntherapeutic agent for other autoimmune related disorders,\nsuch as endometriosis. As described earlier, HCQ is\nconsidered a lysosomotropic agent where it increases the\npH of acidic compartments and also inhibits the fusion of the\nautophagosome with the lysosome. 15–17 We identified a\nmarked reduction in cell survival with this agent in human\nendometriotic and T -HESC cells. The involvement of lyso-\nsomes in endometriosis has yet to be investigated in detail; to\nour knowledge, only one study describes changes in numbers\nof lysosomes in the endometrium of women with\nendometriosis.35 We also successfully knocked down A TG7\n(in addition to other autophagic mediators) in T -HESC; we\nnoted a marked increase in p21 protein expression, which was\nassociated with attenuated cell viability with A TG7 siRNA\n(Supplementary Figure 1e); changes were observed with\nLC3B siRNA but only a subtle effect was noted with beclin-1\nsiRNA. However, with the exception of A TG5 and LC3B\nknockdown, we failed to detect the expected reduction in LC3-\nII to LC3-I ratio with PIK3C3 or beclin-1 siRNA (implicating a\nnoncanonical autophagic pathway), as well as with A TG7\nsiRNA (implicating autophagy independent role). In support, it\nhas been previously reported that A TG7 can lead to increased\np21 via a DNA damage pathway , 36 which appears to be\nindependent of its canonical role in the autophagic pathway .\nTherefore, it is possible that the reduction in cell viability could\nhave occurred via a similar noncanonical autophagy mechan-\nism or independently of autophagy. Further work is necessary\nto clarify the nature of the contribution of these autophagic\nmediators to endometriotic cell survival and lesion\ndevelopment.\nAlthough we identified increases in A TG5 and A TG3 in\nlesions from HCQ-treated mice relative to PBS-treated mice,\nthe majority of the remaining autophagic markers that we\nstudied were unchanged; however, we recognize that this may\nbe because of a limited lesion number sample size obtained\nfrom HCQ-treated mice. As we observed fewer glandular\nstructures and epithelial cells in lesions from endometriosis-\ninduced mice treated with HCQ, it is possible that this drug is\naffecting the epithelial cells of endometriotic lesions more than\nthe stromal cells, thus leading to a decrease in lesion number\nand quality. Our analyses of autophagic markers (by real-time\nPCR or western blotting) did not distinguish between stromal\nand epithelial cells; therefore, further tests would be required\nto dissect the response between the two cell types that may be\nleading to the observed decrease in lesion number.\nHCQ treatment increased the levels of IP-10 and the\nnumbers of macrophages in the peritoneal cavity in our mouse\nmodel of endometriosis. In humans, IP-10 has been reported\nto attract monocytes and T lymphocytes, as well as decrease\nangiogenesis.37,38 A decrease in IP-10 peritoneal levels has\nbeen reported in advanced stages of endometriosis in\npatients, and thus decreased levels of IP-10 may contribute\nFigure 8 Schematic of overall results. Injected fragmented uterine horns implanted and developed in endometriotic lesions. LC3B and lipid droplets were elevated in recipient\nuterine horns compared with uterine horns from recipients, as indicated by protein analyses and TEM, respectively . HCQ reduced lesion numbers relat ive to PBS treatment.\nMoreover, IP-10 levels and macrophages increased in the peritoneal cavity of HCQ-treated mice compared with those treated with PBS\nAutophagy in a mouse model of endometriosis\nA Ruiz et al\n14\nCell Death and Disease\n\nto lesion development, permitting angiogenesis, and decreas-\ning recruitment of natural killer cells and their cytolytic\neffects.39 Therefore, an increase in peritoneal IP-10 levels in\nHCQ-treated mice may have created an unfavorable environ-\nment for lesion development. Whether IP-10 modulates the\nautophagic pathway requires further investigation. It is\ninteresting that another chemokine CXCL12 has recently\nbeen implicated in modulating autophagy in human secretory\nphase endometrial stromal cells.\n40 Additional experiments will\nbe required to determine whether the HCQ effects leading to\nthe increased IP-10 levels corresponds with the observed\ndecrease in lesion number and the mechanisms whereby\nIP-10 could be exerting its anti-endometriotic growth effects.\nA previous study reported altered gene expression in\neutopic endometria from patients with endometriosis com-\npared with those of controls,\n41 suggesting that endometriosis\nmay be altering the endometria of patients. In our study, we\nnoted that the mRNA expression of autophagic markers was\ngenerally decreased in eutopic endometria from endo-\nmetriosis-induced mice when compared with eutopic endo-\nmetria of control mice. Specifically, we found that transcript\nlevels of beclin-1, A TG5, A TG4B, and A TG2B were signifi-\ncantly decreased in endometriotic lesions compared with\nuterine horns (from endometriosis-induced mice treated with\nPBS). Beclin-1 is required for vesicle nucleation during\nautophagosome formation and forms a complex with UVRAG\nto regulate PIK3C3.\n10 A TG5 and A TG4B are required to\ngenerate the cytosolic form of LC3, LC3-I. 42–44 In addition, we\nfound that LC3B protein levels were decreased in lesions\ncompared with uterine horns. T aken together, these changes\nsuggest that dysregulation of autophagic markers may alter\nthe autophagic response in endometriosis. Further research is\nneeded to determine whether autophagosome numbers are\naltered within ectopic lesions. In contrast to the observations in\nthe lesions, we identified increased protein levels of beclin-1,\nLC3B, LC3A, and GABARAPL1 in uterine horns from\nendometriosis-induced mice compared with uterine horns\nfrom controls, with a concurrent increase in lipid droplets. Lipid\ndroplets have been previously described in bovine endome-\ntrial epithelial cells,\n45 and recent findings have implicated LC3\nin intracellular lipid droplet formation. 46 It is possible that the\nobserved increases in LC3B protein are associated with the\naccumulated lipid droplets, but further work is needed to\nunderstand the clinical implications of this finding.\nThe observed changes in the expression of autophagic\nmarkers in eutopic endometria from induced mice could be\nexplained by an activated local immune/inflammatory\nresponse caused by transferring endometrial fragments into\nthe peritoneal cavity.\n3 Although we did not observe an\nincrease in inflammatory markers or macrophage levels\n2 weeks post-induction, we propose that any inflammatory\nresponse that occurred may have been diminished by the time\nof sample collection. T aken together, these observations may\nreveal why endometriosis is able to recur in women after\nsurgical removal of endometriotic lesions.\n47 Whether the\neutopic endometrium of women with endometriosis is altered\nbefore lesion establishment or whether it becomes more\ncompromised after the initiation and establishment of the\ndisease remains unclear and requires further study.\nIn contrast to the reduced expression of LC3B protein in\nlesions obtained from the mouse model of endometriosis\n(relative to uterine horns isolated from the same diseased\nmice), we identified an increase in LC3B levels in epithelial\ncells of endometriotic lesions and in eutopic endometria of\nendometriosis patients compared with endometrium from\ncontrols using immunohistochemistry . We propose that this\ndifference may be due to the possibility that the murine\nendometriosis lesions are equivalent to an early stage of\nendometriosis in patients. As the human disease is not\ndiagnosed until a delay period of 7–10 years,\n48,49 the nature of\nthe isolated lesions from human may be more advanced. In\naddition, the endometrium controls on the TMA are from\nwomen with other gynecological problems, which is a\nlimitation of our study. Despite these limitations, the results\nobtained with the human TMA are more close to the real-life\nscenario of patient heterogeneity in clinical presentations and\nhistory of disease (previous treatments, years with disease)\nand still the changes in LC3B remained significant.\nWhether the aforementioned observed expression changes\nimplicate autophagy dysregulation in a pro-survival or pro-\ndeath manner is currently unknown with regard to endome-\ntriotic lesion establishment. Additional studies will be required\nto determine how autophagy is regulated to allow development\nof endometriosis. Our results also show potential for a novel\ntherapeutic approach for treating patients with this disease.\nPatients with autoimmune diseases (i.e., lupus and arthritis)\nappear to tolerate HCQ treatment well; side effects include\nretinal damage with only rare reports of systemic reactions.\n50\nThus far, there seems to be no adverse effects of this drug on\nfertility,51 which makes HCQ superior to the current use of\nestroprogestinic treatments.\nMaterials and Methods\nEthics and TMA. All protocols in this study were approved by the Institutional\nReview Board at the Ponce Research Institute (Ponce, Puerto Rico). Samples in the\nTMA were obtained in a de-identified manner from archived samples at a private\npathology laboratory (Southern Pathology Laboratories in Ponce, Puerto Rico).\nDetails regarding the human TMA used in this study have been described\npreviously .\n28 Briefly, the TMA contains 164 cores, which is comprised of lesions\n(from the ovaries ( n = 29), fallopian tubes ( n = 16), peritoneum ( n = 34), skin\n(n = 4), and gastrointestinal tract ( n = 7)), eutopic endometrium from endometriosis\npatients ( n = 22), as well as secretory ( n = 38) and proliferative ( n = 14)\nendometrium from endometriosis-free patients. The patients and controls recruited\ninto this biobank were not currently or have been for at least 3 months before\nsurgery on any hormonal medication.\nAnimal handling\nC57BL/6 mouse model : Five-week-old C57BL/6 female mice were purchased\nfrom Jackson Laboratories (Bar Harbor, ME, USA). All animals were maintained\nunder standard 12-h photoperiod; food and water were available ad libitum\nthroughout the study . All experimental procedures and animal care were approved\nby the Animal Care and Use Committee (IACUC) of the University of South Florida\n(R IS00000101), in accordance with the principles described in the Guide for the\nCare and Use of Laboratory Animals of the National Institutes of Health. All surgical\nprocedures were performed under aseptic conditions using anesthesia. The mouse\nmodel of endometriosis was performed as described previously .\n19,20 Donor animals\nreceived a peritoneal injection of 3 μg per mouse of β-estradiol-17-valerate (Sigma,\nSt. Louis, MO, USA); the dose used was based on previously reported data. 20 One\nweek after estrogen injection, donor animals were killed and each uterine horn\nwas collected and minced using a Kirkland Tissue Mincer (Kirkland Products,\nPortland, OR, USA) with sterile normal saline. The minced material was centrifuged\nat 1500 r.p.m. for 1 min. Endometriosis was induced by injecting the uterine horn\nAutophagy in a mouse model of endometriosis\nA Ruiz et al\n15\nCell Death and Disease\n\nfragments intraperitoneally into the recipient animal. Mice were then randomly\ndivided into two groups: HCQ treatment animals were intraperitoneally injected with\n100 μl of 60 mg/kg of HCQ (No. AC26301; Fisher Scientific, Pittsburgh, P A, USA),\nwhereas control treatment animals received an intraperitoneal injection of 100 μl\nsterile PBS. The dose for HCQ used was based on previously published data and\nwas comparable to doses used in treating patients with autoimmune diseases.\n21 A\nsecond HCQ treatment was administered 1 week after endometriosis induction,\nusing the same dose. T wo weeks after induction, mice were killed and tissues\n(including lesions) were snap frozen into liquid nitrogen. Lesions were measured\nusing a caliper. Volume of the lesions was calculated according to the formula:\n4/3πr\n2R.52\nBalb/c mouse model : Eight-week-old Balb/c female mice were obtained from\nthe Charles River Laboratories (Calco, Como, Italy) and handled as described\npreviously19,20 and in accordance with the European Union guidelines, as well as\nwith the approval of the Institutional Animal Care and Use Committee of San\nRaffaele Scientific Institute (Protocol No. 484) (Milan, Italy). Briefly, donor mice were\ninjected with 17 β-estradiol (AMSA, Rome, Italy; 3 μg per mouse) and killed 1 week\nlater. The uterus was removed and fragmented, after scraping to remove the\nmyometrium, using scissors. The endometrial tissues were weighed and\nresuspended in saline with ampicillin (1 mg/ml). T wo recipient mice received an\nintraperitoneal injection, using a syringe containing half of the resuspension (day 0).\nMice were killed by administering a lethal dose of anesthetic on day 12. The\nabdomen was opened and lesions were isolated and collected by an operator\nblinded to the experiment.\nCell culture of life-extended human endometriotic and T -HESC\ncells, HCQ treatment, siRNA transfection, and survival assay.\nPrimary human endometriotic cells culture conditions and life extension have been\ndescribed previously .\n18 These cells were derived from peritoneal ( ‘C') and ovarian\n(‘D’) lesions obtained from two independent patients. These were assessed\nseparately as described below. Briefly , cells were maintained in MCDB 131:Medium\n199 (1 : 1 ratio) supplemented with 8% fetal bovine serum (FBS), penicillin/\nstreptomycin, and insulin/transferrin/selenium (ITS). Cells were life extended using\nsimian virus 40 large T antigen. Retroviral particles generated in HEK293T were\nused to infect the primary cells. Media containing puromycin (2.5 μg/ml) was used to\nselect primary cells resistant colonies. In addition, we obtained the T -HESC cell line,\nwhich are human endometrial stromal cells derived from a uterine myoma (A TCC,\nManassas, VA, USA). This cell line was maintained in phenol red-free DMEM/F12\n(1 : 1) containing 8% charcoal-dextran-treated FBS, 500 ng/ml puromycin, 1% ITS+\nPremix (BD Bioscience, San Jose, CA, USA), and 15 mM HEPES. The cell lines\nused in the present study were tested to be mycoplasma negative and short tandem\nrepeat profiled (Genetica DNA Laboratories, Cincinnati, OH, USA). Endometriotic\ncells were seeded at 50 000 cells per well in a 24-well plate, whereas T -HESC cells\nwere seeded at 250 000 cells per well in a 6-well plate. A 50 mM HCQ (no.\nAC26301; Fisher Scientific, Pittsburgh, P A, USA) stock was prepared in PBS (and\n0.22 μm filter sterilized); it was used at a final concentration of 25 μM in complete\nmedia.\n18,53 Cells were treated for 18 h with HCQ before protein harvest and western\nblotting analyses. For survival studies, cells were seeded at a density of 5 000 cells\nper well in a 96-well opaque plate and treated with 25 μM HCQ during 5 days. Cell\nviability was then assessed using CellTiter-glo reagent (Promega, Madison, WI,\nUSA).\n18\nFor siRNA transfection studies, T -HESC cells were seeded at 350 000 cells per\nwell in a 6-well plate. After overnight adherence, cells were then transfected with\neither non-targeting control siRNA, A TG5, beclin-1, A TG7, PIK3C3, or LC3B siRNA\naccording to previously described methods.\n18,54 The day after the second round of\nsiRNA transfection, cells were reseeded at 5 000 cells per well in opaque 96-well\nplates. Three days after reseeding, cell viability was assessed using CellTiter-glo\nreagent as described above.\nImmunohistochemistry of LC3B. Samples in the TMA were collected in a\nde-identified manner from archived samples in a Pathology Lab as described in\nHuman Subjects above. Briefly, slides were deparaffinized and stained using the\nautomated system Ventana Discovery XT (Ventana Medical Systems, T ucson, AZ,\nUSA) with EZ Prep solution. The heat antigen retrieval method was performed at a\npH of 8.0. The primary antibody , LC3 (AP1802a), that detects LC3B was obtained\nfrom Abgent (San Diego, CA, USA) and diluted at a ratio of 1 : 25 in Dako antibody\ndiluent (Dako, Carpenteria, CA, USA), followed by a 32 min incubation at room\ntemperature. Human breast cancer tissue was used as a positive control and the\nprimary antibody was omitted for the negative control. Ventana OmniMap anti-rabbit\nsecondary antibody and the Ventana Medical Systems (T ucson, AZ, USA) as the\ndetection system were used. Hematoxylin was used as the counterstain.\nThe LC3-stained TMA was then scanned using the Aperio ScanScope XT (Aperio,\nVista, CA, USA) with a × 200 magnification and a 0.8 numerical aperture objective\nlens via the Basler tri-linear array detection. Each core was then segmented using the\nTMA block software associated with the Spectrum program (version 10.2.5.2352),\nfollowed by manual segmentation into epithelial and stromal regions under the\nsupervision of a pathologist. Image analysis was performed using an Aperio Positive\nPixel Count v.9.0 algorithm with the following thresholds: Hue Value = 0.1; Hue\nWidth = 0.5; Color Saturation Threshold = 0.04; IWP(High) = 220; IWP(Low) = IP\n(High) = 175; IP(low) = ISP(High) = 100; ISP(Low) = 0 to segment positive staining\nof various intensities. The data were then compiled for each core in the separate\nepithelium and stromal regions, which was represented by percent positivity , and then\ndirectly correlated with protein expression.\nRNA isolation, real-time PCR, and RT\n2-PCR. T otal RNA was isolated\nusing the RNeasy Kit following the manufacturer ’s instructions (Qiagen, Valencia,\nCA, USA). RNA concentration and purity was determined using a 1000 NanoDrop\n(Thermo Scientific, Pittsburgh, P A, USA). Lesion mass varied by samples, and this\nwas reflected in the RNA amounts obtained (range of mass: 0.9 –25 mg).\nThree RNA samples from uterine horns, having a 260/280 ratio 41.8 and a\n260/230 ratio 41.7, were selected from recipient, donor, HCQ-treated, and PBS-\ntreated animals (12 samples in total) for RT\n2-PCR analyses. Synthesis of cDNA was\nperformed using 0.5 μg of total RNA, after DNA elimination step using the RT 2 First\nStrand Kit as per the manufacturer's instruction (Qiagen). After DNA elimination, the\nreaction mix was incubated at 42 °C for 15 min, followed by 95 °C for 5 min using a\nDNA Engine Peltier Thermal cycler (Bio-Rad, Hercules, CA, USA). A total of 102 μlo f\nthe cDNA reaction mix was added to the master mix containing 1248 μlo f\nRNAse-free water and 1350 μlo f2 xR T\n2 SYBR green master mix. T wenty-five\nmicroliters of the master mix were carefully added to each well of the RT2 profiler PCR\nautophagy array. Quantification was performed using the Applied Biosystems cycler\n(Life T echnologies, Grand Island, NY , USA). The PCR cycling program included\nactivation for 10 min at 95 °C, followed by 40 cycles for 15 s at 95 °C with 1 min\nat 60 °C. The PCR cycling program finalized with a melt-curve analysis and data was\nanalyzed using the Qiagen web-based software (http: //www.SABiosciences.com/\npcrarraydataanalysis.php).\nFor real-time PCR studies, the One-step Master Mix (Applied Biosystems, Foster\nCity , CA, USA) was used with the following probes and primers as described\npreviously:\n18 LC3B, Mm00782868_sH; A TG4B, Mm01701111_m1; A TG9A,\nMm01264420_m1; A TG5, Mm00504340_m1; A TG7, Mm00512209_m1; A TG3,\nMm00471287_m1; PIK3C3, Mm00619489_m1; ULK1 (unc-51-like autophagy-\nactivating kinase 1), Mm00437238_m1; A TG9B, Mm01157883_g1; Beclin-1,\nMm01265461_m1; A TG2B, Mm00512973_m1; A TG4C, Mm01259886_m1; BNIP3,\nMm01275600_g1; EIF2AK3, Mm00438700_m1; FAS, Mm01204974_m1; LC3A,\nMm00458725_g1; GABARAPL1, Mm00457880_m1; IGF1, Mm00439560_m1;\nIRGM1, Mm00492596_m1; SQSTM1 (p62), Mm00448091_m1; PRKAA1,\nMm01296700_m1; PTEN, Mm00477208_m1. C\nT values were normalized to β-actin\n(Mm00607939_s1) and RNA fold changes were determined using the 2 − ΔΔ C\nT\nequation.\nProtein isolation, SDS-P AGE, and western blot analyses. Tissues\nused for protein analyses included uterine horns, ectopic lesions, ovaries, thymus,\nkidneys, heart, pancreas, spleen, and liver. Samples were flash frozen in liquid\nnitrogen and stored at − 80 °C until use. Tissues were homogenized in ice-cold lysis\nbuffer containing 1% T riton X-100, 50 mM HEPES, 150 mM NaCl, 1 mM MgCl\n2,\n1 mM EGT A, 10% glycerol, and protease inhibitor cocktail (Roche, Indianapolis, IN,\nUSA) using a PowerGen 125 homogenizer (Fisher Scientific, Pittsburgh, P A, USA).\nSamples were centrifuged at 14 000 r.p.m. for 10 min at 4 °C. The supernatants\nwere collected and total protein concentration was determined using the BCA assay\n(Thermo Scientific, Rockford, IL, USA), and a BioT ek synergy 2 microplate reader\n(BioT ek, Winooski, VT , USA). Samples were normalized and then run onto 10 or\n12% SDS-polyacrylamide gels prepared in a Criterion Cassette system (Bio-Rad,\nHercules, CA, USA) as described previously .\n54 The following antibodies and\ndilutions were used: LC3B rabbit polyclonal (no. 2775, 1 : 1 000), LC3A rabbit\nmonoclonal (no. 4599 (D50G8), 1 : 1 000), Beclin-1 rabbit polyclonal (no. 3738,\n1 : 1,000), GABARAPL1 rabbit monoclonal (no. 13733 (E1J4E), 1 : 1000), AMPK α\nrabbit monoclonal (no. 2603 (23A3), 1 : 500), FOXO1 rabbit monoclonal (no. 2880\n(C29H4), 1 : 1000), and pan-actin rabbit polyclonal (no. 4968, 1 : 500) were all\nAutophagy in a mouse model of endometriosis\nA Ruiz et al\n16\nCell Death and Disease\n\nobtained from Cell Signaling T echnology (Danvers, MA, USA). The p62 mouse\nmonoclonal antibody (no. 610832, 1 : 1 000) was obtained from BD Biosciences\n(San Jose, CA, USA).\nHematoxylin/eosin staining, TMA construction, and immunohis-\ntochemistry. Collected samples were immediately preserved in 10% neutral-\nbuffered formalin at the animal facility. Samples were embedded in paraffin,\nsectioned, and transferred to slides for hematoxylin/eosin and immunohistochemical\nstaining. A pathologist reviewed each case and delimited the region of interest,\ncontaining epithelial and stromal cells, for each specimen. A mouse TMA was\nprepared at the Tissue Core Facility at the Moffitt Cancer Center. The mouse TMA\ncontained a total of 113 core samples, which included 10 uterine horns and 10\novaries from both PBS- and HCQ-treated mice. As control specimens for the\nantibodies used, the TMA included mouse mammary tissue, liver, small intestine,\nand lymph nodes from a PBS-treated mouse. Lesions were analyzed from\nindependent blocks. Slides were stained using a Leica Bond RX automated system\n(Leica Biosytems, Buffalo Grove, IL, USA) following the manufacturer's instructions\nwith proprietary reagents. Slides were deparaffinized on an automated system with\nDewax Solution (Leica Biosystems). The antigen retrieval method used for PR was\nenzymatic with Enzyme Solution 1 at 15 min (Leica), for vimentin and ER was\nheat induced with Epitope Retrieval Solution 1 at 20 min (Leica), for CK8 was heat\ninduced with Epitope Retrieval Solution 2 at 10 min (Leica), and for LC3B was\nheat induced with Epitope Retrieval Solution 1 at 10 min (Leica Biosystems).\nAll antibodies were diluted in Dako antibody diluent (Dako): PR (no. ab131486,\n1 : 500; Abcam, Cambridge, MA, USA), vimentin (no. 5741 (D21H3), 1 : 100; Cell\nSignaling), ERα (no. ab32063 (E115), 1 : 200; Abcam), CK8 (no. ab53280 (EP1628Y),\n1 : 200; Abcam), and LC3B (no. ab51520, 1 : 1500; Abcam) and incubated for 30 min.\nThe Leica Bond Polymer Refine Detection System Leica Biosystems was used with a\npolymer incubation for 8 min. Hematoxylin was used as a counterstain, and slides were\ndehydrated and covered with a coverslip, following standard histological protocol.\nAnalysis of murine peritoneal inflammatory molecules. After\nanimals were killed, 1 ml of sterile PBS was injected into the peritoneal cavity ,\nthe abdominal area was gently massaged, and the fluid collected. The collected\nfluid was centrifuged at 1390 r.p.m. for 5 min at 4 °C and the resulting supernatant\nwas then stored at − 80 °C. Levels of chemokines and cytokines were analyzed\nusing a MCYTOMAG-70K-PX32 (Millipore, Billerica, MA, USA) following the\nmanufacturer’s instructions. Briefly, 200 μl of wash buffer was added to each well\nand incubated for 10 min at room temperature in a plate shaker. After incubation,\nthe wash buffer was decanted and the plate was inverted and tapped on absorbent\ntowel several times. Then, 25 μl of assay buffer was added to each well, followed by\n25 μl of concentration standards, assay controls, or samples. The premixed bottle\nwas vortexed and 25 μl of the beads were added to each well. The plate was\nincubated overnight at 4 °C, and then protected from light. Then, the plate was\nincubated for 1 min on the hand-held magnet and the well content was gently\ndecanted and tapped on absorbent pads. Each well was washed two times using\n200 μl of wash buffer, followed by the incubation on the hand-held magnet. Antibody\ndetection solution was allowed to warm to room temperature, and then 25 μlw a s\nadded to each well and incubated for 1 h at room temperature on a plate shaker,\nprotected from light. Next, 25 μl of streptavidin-phycoerythrin was added to each\nwell containing the detection antibodies and incubated for 30 min at room\ntemperature protected from light on a plate shaker. After the incubation, the plate\nwas washed two times as previously described and 150 μl of sheath fluid was\nadded to each well. The plate was analyzed using MAGPIX instrument and\nxPONENT software solutions, version 4.2 Luminex Corporation (Austin, TX, USA).\nFlow cytometry. The pellet obtained after centrifugation of the peritoneal fluid\nwash (see above) was used for macrophage staining. When necessary, red blood\ncell lysis was performed according to the manufacturer ’s protocol (eBioscience, San\nDiego, CA, USA). The cell pellets were resuspended in 1 ml cold PBS and\ntransferred to flow cytometry tubes. Samples were centrifuged for 1 min at 1390 r.p.\nm. The supernatant was decanted and cells were resuspended in 100 μl of PBS.\nCells were blocked using 0.5 μg of Mouse BD Fc, Block (no. 553141; BD\nPharmingen, San Jose, CA, USA) for 5 min on ice. The cells were incubated in\n0.4 μg of APC rat anti-mouse CD11b clone M1/70 (no. 553312; BD Pharmingen)\nand anti- mPE-F4/80/EMR1 (no. FAB5580C; R&D Systems, Minneapolis, MN, USA)\nat room temperature for 30 min, protected from light. After incubation, 700 μl of PBS\nwas added to each tube and centrifuged for 1 min at 4 °C. The supernatant was\ndecanted, and the cells were resuspended in 300 μl in PBS and analyzed by flow\ncytometry.\nTransmission electron microscopy. Following induction of anesthesia,\nthe abdominal cavity of the mice was opened to expose the uterine horns. Both\nuterine horns were removed and cut in cross-sections of 2 –3 mm long pieces, which\nwere then rinsed in 0.1 M phosphate buffer to remove excess blood, and placed in\n2.5% glutaraldehyde in 0.1 M sodium phosphate buffer, pH 7.2, at 4 °C. The tissue\nwas fixed in glutaraldehyde at 4 °C for 24 h. Following fixation, the tissue was rinsed\nin buffer, sliced into 1-mm-thick rings and postfixed in 1% osmium tetroxide at 4 °C\nfor 2 h. Following buffer and distilled water rinses, the tissue was dehydrated\nthrough a graded series of acetone dilutions, cleared with propylene oxide, infiltrated\novernight, embedded in LX 112 epoxy resin mix (Ladd Research, Williston, VT ,\nUSA), and polymerized at 70 °C. Entire cross-sections of the uterine horns were\nobtained at 0.25 –0.35 μm thickness and 70 –80 nm thickness, and stained with 1%\ntoluidine blue stain (for light microscopy) or 8% uranyl acetate and Reynold ’s lead\ncitrate (for electron microscopy), respectively. The endometrium of both control and\nexperimental animals was observed and photographed using an FEI Morgagni TEM\n(FEI Company, Hillsboro, OR, USA) with an AMT ActiveVu camera (AMT , Woburn,\nMA, USA).\nStatistical analyses. All analyses were performed using GraphPad Prism\nsoftware (version 6.04; GraphPad, La Jolla, CA, USA). T o calculate the significance\nof the observed disorganization of epithelial cells in eutopic endometria from\nendometriosis-induced mice treated with HCQ (compared with those treated with\nPBS, as a control), we used the Fisher ’s exact test. All other statistical analyses\nwere calculated using the nonparametric Student's t-test and error bars displayed\nrepresent standard errors of the mean (S.E.M.). Statistical significance was set at\nP ⩽ 0.05 (* P ⩽ 0.05, ** P ⩽ 0.01, *** P ⩽ 0.001, and **** P ⩽ 0.0001).\nConflict of Interest\nA provisional patent application on autophagy and endometriosis has been\nsubmitted” (MN, IF). All other co-authors declare no conflict of interest.\nAcknowledgements. The work presented herein was supported by NICHD\nR21HD075225-01 to MN. In addition, this work was supported by the Moffitt Cancer\nCenter Tissue Core and Analytic Microscopy Core Facilities (P30-CA076292). We\ngratefully acknowledge Dr Ana Sanchez (Ospedale San Raffaele, Milan, Italy) as well\nas Aimee Bode and Jennifer Morse (Moffitt Cancer Center) for their discussions and\nguidance on the studies described herein. We kindly acknowledge the assistance of\nDr Karoly Szekeres (Flow Cytometry, College of Medicine, University of South\nFlorida). In addition, we thank Dr Brant Burkhardt and Melanie Kuehl (Department of\nCell Biology , Microbiology , and Molecular Biology , University of South Florida) for their\nassistance with the chemokine and cytokine measurements. We also thank Arielle\nSharp for her assistance with manuscript revisions. We gratefully acknowledge\nDr Miosotis Garcia for the development of the human TMA, Dr Adalberto Mendoza,\nMedical Director of the Southern Pathology Inc., for access to the FFPE blocks\nincluded in the TMA, and Samir Bello for technical assistance in the development of\nthe endometriotic cell lines. Development of the endometriosis TMA was funded by\nGrant No. U56CA126379 and the endometriotic cells lines was funded by Grant No.\nHDR01050559.\nAuthor contributions\nMN conceived this project. MN, AR, SR, NT , EH, IF , and PP performed the\nexperiments described herein. PP and RE assisted with development of the mouse\nmodels used. IF directed efforts to develop the human endometriosis TMA and the\nprimary cell lines. MN, AR, and SR wrote the manuscript. All authors analyzed the\ndata and critically reviewed the final manuscript before submission. AR and SR\ncontributed equally to this project.\n1. Giudice LC, Kao LC. Endometriosis. Lancet 2004; 364:1 7 8 9–1799.\n2. Bulun SE. Endometriosis. N Engl J Med 2009; 360: 268–279.\n3. Giudice LC. Clinical practice. Endometriosis. N Engl J Med 2010; 362: 2389–2398.\n4. Sourial S, T empest N, Hapangama DK. Theories on the pathogenesis of endometriosis. Int J\nReprod Med 2014; 2014: 179515.\n5. Sampson JA. Peritoneal endometriosis due to the menstrual dissemination of endometrial\ntissue into the peritoneal cavity. Am J Obstet Gynecol 1927; 14:4 2 2–469.\nAutophagy in a mouse model of endometriosis\nA Ruiz et al\n17\nCell Death and Disease\n\n6. Baldi A, Campioni M, Signorile PG. Endometriosis: pathogenesis, diagnosis, therapy and\nassociation with cancer (review). Oncol Rep 2008; 19: 843–846.\n7. Vercellini P, Vigano P, Somigliana E, Fedele L. Endometriosis: pathogenesis and treatment.\nNat Rev Endocrinol 2014; 10: 261–275.\n8. Fung C, Lock R, Gao S, Salas E, Debnath J. Induction of autophagy during extracellular\nmatrix detachment promotes cell survival. Mol Biol Cell 2008; 19: 797–806.\n9. Lamb CA, Y oshimori T, T ooze SA. The autophagosome: origins unknown, biogenesis\ncomplex. Nat Rev Mol Cell Biol 2013; 14: 759–774.\n10. Feng Y, He D, Y ao Z, Klionsky DJ. The machinery of macroautophagy.Cell Res2014;24:2 4–41.\n11. Legakis JE, Y en WL, Klionsky DJ. A cycling protein complex required for selective\nautophagy. Autophagy 2007; 3: 422–432.\n12. Pelch KE, Schroder AL, Kimball P A, Sharpe-Timms KL, Davis JW, Nagel SC. Aberrant gene\nexpression profile in a mouse model of endometriosis mirrors that observed in women. Fertil\nSteril 2010; 93: 1615–1627, e1618.\n13. Choi J, Jo M, Lee E, Kim HJ, Choi D. Differential induction of autophagy by mTOR is\nassociated with abnormal apoptosis in ovarian endometriotic cysts. Mol Hum Reprod 2014;\n20: 309–317.\n14. Allavena G, Carrarelli P, Del Bello B, Luisi S, Petraglia F, Maellaro E. Autophagy is\nupregulated in ovarian endometriosis: a possible interplay with p53 and heme oxygenase-1.\nFertil Steril 2015; 103: 1244–1251, e1241.\n15. Al-Bari MA. Chloroquine analogues in drug discovery: new directions of uses, mechanisms\nof actions and toxic manifestations from malaria to multifarious diseases. J Antimicrob\nChemother 2015; 70: 1608–1621.\n16. Amaravadi RK, Lippincott-Schwartz J, Yin XM, Weiss WA, T akebe N, Timmer W et al.\nPrinciples and current strategies for targeting autophagy for cancer treatment. Clin Cancer\nRes 2011; 17: 654–666.\n17. Calabretta B, Salomoni P. Inhibition of autophagy: a new strategy to enhance sensitivity of chronic\nmyeloid leukemia stem cells to tyrosine kinase inhibitors.Leuk Lymphoma 2011; 52:5 4–59.\n18. Bauckman KA, Haller E, Flores I, Nanjundan M. Iron modulates cell survival in a Ras- and\nMAPK-dependent manner in ovarian cells. Cell Death Dis 2013; 4: e592.\n19. Somigliana E, Vigano P, Rossi G, Carinelli S, Vignali M, Panina-Bordignon P. Endometrial\nability to implant in ectopic sites can be prevented by interleukin-12 in a murine model of\nendometriosis. Hum Reprod 1999; 14: 2944–2950.\n20. Mariani M, Vigano P, Gentilini D, Camisa B, Caporizzo E, Di Lucia P et al. The selective\nvitamin D receptor agonist, elocalcitol, reduces endometriosis development in a mouse\nmodel by inhibiting peritoneal inflammation. Hum Reprod 2012; 27: 2010–2019.\n21. McAfee Q, Zhang Z, Samanta A, Levi SM, Ma XH, Piao S et al. Autophagy inhibitor Lys05\nhas single-agent antitumor activity and reproduces the phenotype of a genetic autophagy\ndeficiency. Proc Natl Acad Sci USA 2012; 109: 8253–8258.\n22. Bacci M, Capobianco A, Monno A, Cottone L, Di Puppo F, Camisa B et al. Macrophages are\nalternatively activated in patients with endometriosis and required for growth and\nvascularization of lesions in a mouse model of disease. Am J Pathol 2009; 175: 547–556.\n23. Jang CH, Choi JH, Byun MS, Jue DM. Chloroquine inhibits production of TNF-alpha,\nIL-1beta and IL-6 from lipopolysaccharide-stimulated human monocytes/macrophages by\ndifferent modes. Rheumatology (Oxford) 2006; 45: 703–710.\n24. Dinulescu DM, Ince T A, Quade BJ, Shafer SA, Crowley D, Jacks T. Role of K-ras and Pten in\nthe development of mouse models of endometriosis and endometrioid ovarian cancer. Nat\nMed 2005; 11:6 3–70.\n25. Matsuzaki S, Canis M, Pouly JL, Botchorishvili R, Dechelotte PJ, Mage G. Differential\nexpression of genes in eutopic and ectopic endometrium from patients with ovarian\nendometriosis. Fertil Steril 2006; 86: 548–553.\n26. Meola J, Rosa e Silva JC, Dentillo DB, da Silva WA Jr., Veiga-Castelli LC, Bernardes LA et al.\nDifferentially expressed genes in eutopic and ectopic endometrium of women with\nendometriosis. Fertil Steril 2010; 93: 1750–1773.\n27. Klionsky DJ, Abdalla FC, Abeliovich H, Abraham RT, Acevedo-Arozena A, Adeli K et al.\nGuidelines for the use and interpretation of assays for monitoring autophagy. Autophagy\n2012; 8: 445–544.\n28. Colon-Diaz M, Baez-Vega P, Garcia M, Ruiz A, Monteiro JB, Fourquet J et al. HDAC1 and\nHDAC2 are differentially expressed in endometriosis. Reprod Sci 2012; 19: 483–492.\n29. Lee SJ, Silverman E, Bargman JM. The role of antimalarial agents in the treatment of SLE\nand lupus nephritis.\nNat Rev Nephrol 2011; 7: 718–729.\n30. Nothnick WB. T reating endometriosis as an autoimmune disease. Fertil Steril 2001; 76:\n223–231.\n31. Nirgianakis K, Bersinger NA, McKinnon B, Kostov P, Imboden S, Mueller MD. Regression of\nthe inflammatory microenvironment of the peritoneal cavity in women with endometriosis by\nGnRHa treatment. Eur J Obstet Gynecol Reprod Biol 2013; 170: 550–554.\n32. Sinaii N, Cleary SD, Ballweg ML, Nieman LK, Stratton P. High rates of autoimmune and\nendocrine disorders, fibromyalgia, chronic fatigue syndrome and atopic diseases among\nwomen with endometriosis: a survey analysis. Hum Reprod 2002; 17: 2715–2724.\n33. Kvaskoff M, Mu F, T erry KL, Harris HR, Poole EM, Farland L et al. Endometriosis: a high-risk\npopulation for major chronic diseases? Hum Reprod Update 2015; 21: 500–516.\n34. Ben-Zvi I, Kivity S, Langevitz P, Shoenfeld Y. Hydroxychloroquine: from malaria to\nautoimmunity. Clin Rev Allergy Immunol 2012; 42: 145–153.\n35. von Adamek EV, Simoes MJ, Freitas V, Patriarca MT, Soares JM Jr., Baracat EC. Lysosomal\nevaluation of endometrioma capsule epithelium and endometrium of patients with or without\nendometriosis. Clin Exp Obstet Gynecol 2005; 32:2 7–30.\n36. Lee IH, Kawai Y, Fergusson MM, Rovira II, Bishop AJ, Motoyama N et al. Atg7 modulates\np53 activity to regulate cell cycle and survival during metabolic stress. Science 2012; 336:\n225–228.\n37. T aub DD, Lloyd AR, Conlon K, Wang JM, Ortaldo JR, Harada A et al. Recombinant\nhuman interferon-inducible protein 10 is a chemoattractant for human monocytes and T\nlymphocytes and promotes T cell adhesion to endothelial cells. J Exp Med 1993; 177:\n1809–1814.\n38. Strieter RM, Kunkel SL, Arenberg DA, Burdick MD, Polverini PJ. Interferon gamma-inducible\nprotein 10 (IP-10), a member of the C-X-C chemokine family, is an inhibitor of angiogenesis.\nBiochem Biophys Res Commun 1995; 210:5 1–57.\n39. Galleri L, Luisi S, Rotondi M, Romagnani P, Cobellis L, Serio M et al. Low serum and\nperitoneal fluid concentration of interferon-gamma-induced protein-10 (CXCL10) in women\nwith endometriosis. Fertil Steril 2009; 91: 331–334.\n40. Mei J, Zhu XY, Jin LP, Duan ZL, Li DJ, Li MQ. Estrogen promotes the survival of human\nsecretory phase endometrial stromal cells via CXCL12/CXCR4 up-regulation-mediated\nautophagy inhibition. Hum Reprod 2015; 30: 1677–1689.\n41. Burney RO, T albi S, Hamilton AE, Vo KC, Nyegaard M, Nezhat CR et al. Gene expression\nanalysis of endometrium reveals progesterone resistance and candidate susceptibility genes\nin women with endometriosis. Endocrinology 2007; 148: 3814–3826.\n42. Marino G, Uria JA, Puente XS, Quesada V, Bordallo J, Lopez-Otin C. Human autophagins, a\nfamily of cysteine proteinases potentially implicated in cell degradation by autophagy. J Biol\nChem 2003; 278: 3671–3678.\n43. Geng J, Klionsky DJ. The Atg8 and Atg12 ubiquitin-like conjugation systems in\nmacroautophagy. 'Protein modifications: beyond the usual suspects' review series. EMBO\nRep 2008; 9: 859–864.\n44. Li M, Hou Y, Wang J, Chen X, Shao ZM, Yin XM. Kinetics comparisons of mammalian Atg4\nhomologues indicate selective preferences toward diverse Atg8 substrates. J Biol Chem\n2011; 286: 7327–7338.\n45. Wordinger RJ, Dickey JF, Ellicott AR. Histochemical evaluation of the lipid droplet content of\nbovine oviductal and endometrial epithelial cells. J Reprod Fertil 1977; 49:1 1 3–114.\n46. Shibata M, Y oshimura K, T amura H, Ueno T, Nishimura T, Inoue Tet al. LC3, a microtubule-\nassociated protein1A/B light chain3, is involved in cytoplasmic lipid droplet formation.\nBiochem Biophys Res Commun 2010; 393: 274–279.\n47. T aylor E, Williams C. Surgical treatment of endometriosis: location and patterns of disease at\nreoperation. Fertil Steril 2010; 93:5 7–61.\n48. Laux-Biehlmann A, d'Hooghe T, Zollner TM. Menstruation pulls the trigger for inflammation\nand pain in endometriosis. T rends Pharmacol Sci 2015; 36: 270–276.\n49. Hadfield R, Mardon H, Barlow D, Kennedy S. Delay in the diagnosis of endometriosis: a\nsurvey of women from the USA and the UK. Hum Reprod 1996; 11: 878–880.\n50. Costedoat-Chalumeau N, Dunogue B, Leroux G, Morel N, Jallouli M, Le Guern V et al. A\ncritical review of the effects of hydroxychloroquine and chloroquine on the eye. Clin Rev\nAllergy Immunol 2015; 49: 317–326.\n51. Janssen NM, Genta MS. The effects of immunosuppressive and anti-inflammatory\nmedications on fertility , pregnancy, and lactation. Arch Intern Med 2000;\n160: 610–619.\n52. Bilotas MA, Olivares CN, Ricci AG, Baston JI, Bengochea TS, Meresman GF et al.\nInterplay between endometriosis and pregnancy in a mouse model. PLoS One 2015; 10:\ne0124900.\n53. Perecko T, Kassab RB, Vasicek O, Pekarova M, Jancinova V, Lojek A. The effects of\nchloroquine and hydroxychloroquine on nitric oxide production in RAW 264.7 and bone\nmarrow-derived macrophages. Folia Biol (Praha) 2014; 60:3 9–44.\n54. Smith DM, Patel S, Raffoul F, Haller E, Mills GB, Nanjundan M. Arsenic trioxide induces a\nbeclin-1-independent autophagic pathway via modulation of SnoN/SkiL expression in\novarian carcinoma cells. Cell Death Differ 2010; 17: 1867–1881.\nCell Death and Disease is an open-access journal\npublished by Nature Publishing Group . This work is\nlicensed under a Creative Commons Attribution 4.0 International\nLicense. The images or other third party material in this article are\nincluded in the article ’s Creative Commons license, unless indicated\notherwise in the credit line; if the material is not included under the\nCreative Commons license, users will need to obtain permission from\nthe license holder to reproduce the material. To view a copy of this\nlicense, visit http://creativecommons.org/licenses/by/4.0/\nSupplementary Information accompanies this paper on Cell Death and Disease website (http://www.nature.com/cddis)\nAutophagy in a mouse model of endometriosis\nA Ruiz et al\n18\nCell Death and Disease","source_license":"CC0","license_restricted":false}