{"paper_id":"e23059e5-7a73-4a62-8f56-fdecd426e88a","body_text":"The human endometrium or mucosal lining of\nthe uterus is a unique, special tissue which consists\nof surface epithelium, glands and stroma. The endometrium\nundergoes intense periods of proliferation,\ngrowth and angiogenesis under the effect of\nsexual hormones ( 1 ,  2 ). This tissue plays a pivotal\nrole in reproduction; its growth and thickness is\none of determining factors of fertility ( 1 ,  3 ,  4 ).\nEndometriosis is a benign lesion in the pelvis\nand other parts of peritoneum which is defined as\nthe existence of endometrial glands and stroma\noutside of the uterus. It is a hormone-dependent\ndisease and a cause of infertility ( 5 ). Different genetic,\nimmunological and environmental factors\nare considered to be causes of endometriosis, although inflammatory factors such as prostaglandins\n(PG) may have a role in this disease ( 6 ). Endometriosis\nhas been considered to be an angiogenic disease\n( 7 ). Medications such as letrozole, ( 8 ) raloxifene,\n( 9 ) celecoxib, ( 10 ) and statins ( 11 ) have been\nintroduced as treatments for endometriosis.\nCyclooxygenase enzymes convert arachidonic\nacid to PG and exist in two main isoforms, COX-\n1 and COX-2. COX-1 is a housekeeping enzyme\nwhich is found in most human tissue, whereas\nCOX-2 is mostly located in the kidneys, brain, endothelium\nand female reproductive system. COX-2\nis induced by pathologic stimuli such as inflammation\nand its upregulation has been observed in some\ndiseases ( 12 ). There is increased COX-2 expression\nin eutopic and ectopic tissue of endometriosis patients\n( 13 ), along with increased levels of PG in their\nserum and peritoneal fluid ( 14 ). Some studies have\nreported that inhibition of PG production resulted in\ndecreased growth of ectopic endometrial tissue ( 15 ).\nCelecoxib is a diaryl-substituted pyrazole\n(C 17 H 14 F 3 N 3 O 2 S). It is the newest class of nonsteriod\nanti-inflammatory drugs (NSAIDs) and a potent,\nselective inhibitor of COX-2. Celecoxib has new\napplications in cancer chemoprevention and gynecology\n( 16 ). It is considered to be an anti-angiogenic\nagent ( 17 ) and apoptosis inducer ( 18 ). Celecoxib\nhas been shown to inhibit IL6 production, colony\nformation, cell viability and cell migration ( 19 ).\nCelecoxib has been proposed for inhibition of\nendometriosis lesions and VEGF secretion in endometriosis\n( 15 ,  16 ). Thus a study of the role of the\nCOX-2 inhibitor as a novel therapeutic modality in\nthis disease has been proposed ( 18 ). Endometrial\ntissue culture in a three-dimensional (3D) fibrin\nmatrix was introduced for endometriosis research\n( 20 ) and applied for a study of the drug’s effect on\nthe endometrium ( 21 ). However no scientific report\non the effect of celecoxib on human endometria has\nbeen published. Therefore, the aim of the present\nwork is to investigate the celecoxib effect on normal\nhuman endometria in a 3D culture model.\n\nIn this experimental study, endometrial biopsies\n(n=10) were taken from reproductive age women\n(25-40 years) who underwent surgery for either benign\nmyoma or diagnostic laparoscopy. The Ethics\nCommittee of Kermanshah University of Medical\nSciences accepted the work on human tissue in this\nstudy and all patients signed informed consents. All\nchemicals and enzymes were purchased from Sigma\nCompany (Germany). Fetal bovine serum (FBS) was\npurchased from Gibco Company (Denmark). The\nculture method has been previously reported ( 20 ) and\nthoroughly discussed in our previous study ( 21 ).\nThe endometrial samples were in the proliferative\nphase. Each sample was divided into two\nparts, one for pathologic diagnosis and the other\nfor tissue culture. The exclusion criteria were endometrial\nmalignancies (cancer, hyperplasia, and\npolyps) and patients on hormone therapy or those\nwho used intrauterine devices (IUD) during the\nprevious three months. Only normal endometrium\ndata as reported by a pathologist were chosen for\nfinal analyses.\nEndometrial biopsies were placed in Hank’s balanced\nsalt solution that contained amphotericin\nB (2.5 μg/ml) plus penicillin (50 μg/ml). Biopsies\nwere washed and cut into small fragments\n(approximately 1×1 mm). We used one, 24-well\nculture plate (Orange Scientific) for each biopsy.\nEach row of the culture plate was used for either\nthe control or one of the celecoxib doses ( Fig 1 ).\nSchematic drawing of 24-well culture dish with study\ndesign for one biopsy.\nFibrin gel was formed by the addition of 0.5 ml/\nwell of fibrinogen solution (3 mg/ml in M199) to\neach well and mixed with 10 μL of thrombin (50\nNIH U/ml in 0.15 M NaCl). Endometrial fragments\nwere placed in the center of the wells and covered\nby an additional fibrinogen/thrombin solution that\nformed a second gel layer to hold the endometrial\nexplants between the two clots. M199 supplemented\nwith L-glutamine (2 mM), 5% heat-inactivated FBS, 0.1% ε-amino caproic acid, streptomycin (50\nμg/ml), penicillin (50 IU/ml) and amphotericin B\n(2.5 μg/ml) was added to all wells. Experimental\nwells contained one dose of either 1 μM, 10 μM or\n50 μM of celecoxib ( 10 ,  18 ,  22 ).\nTissues were cultured for 21 days and the culture\nmedia were changed every three days. The explants\nwere cultured at 37˚C in 95% air and 5% CO 2  in a\nhumidified environment. On the first and the last day\nof the culture, explants were photographed for final\ncomparison. At the end of the study, scoring methods\nwere used to determine any tissue growth and morphological\nchanges as observed by two different individuals\nblinded to the analyses ( 21 ) and taking into\nconsideration cellular invasion into the fibrin matrix\nand capillary-like sprouting of endothelial cells.\nThe scoring method was: 0. no growth and tissue\nchanges; 1. growth in less than 25% of the explant;\n2. growth in 26% to 50% of the explant; 3. growth\nin 51% -75% of the explant and 4. growth in more\nthan 75% of each explant. The mean of two scores\nwas used for data analysis. We studied ten biopsies,\neach with a control and experimental groups. We\ncompared the mean score of all control wells from the\nten biopsies with the mean of score of the three different\ncelecoxib doses (1, 10 and 50 μM).\nThe numbers of wells that showed angiogenesis\nwere determined and percent of angiogenesis was\ncalculated. To document endothelial cell sprouting\ninto the fibrin matrix, we fixed the fibrin clots in\n4% paraformaldehyde. Tissue processing was performed\nand microscopic slides immunohistochemically\nstained for endothelial cell marker (CD31)\nwith anti-CD31 antibody ( 21 ).\nThe scores were quantitative and with no normal\ndistribution, thus statistical analyses were\nperformed with the Kruskal-Wallis method using\nSPSS software (version 16). A p value of <0.05\nwas considered significant.\n\nThe mean growth score was 1.37 ± 0.16 for\nthe control. The celecoxib groups had the following\nmean growth scores: 1.97 ± 0.28 (1\nμM), 2.01 ± 0.25 (10 μM ) and 1.17 ± 0.14\n(50 μM). The difference between groups was\nsignificant (p=0.03). The results showed that\n1 and 10 μM concentrations of celecoxib\nstimulated endometrial growth and increased\nthe growth score. The 50 μM concentration\ndid not stimulate endometrial growth and its\ngrowth score approximated that seen with the\ncontrol group ( Fig 2 ).\nThe percent of angiogenesis observed was 25.56\n± 6.72% for the control and 31.89 ± 6.18% (1 μM),\n42.67 ± 7.27% (10 μM) and 23.44 ± 4.03% (50\nμM) for the celecoxib groups, which was not a\nsignificant difference. Angiogenesis was higher at\nthe 1 and 10 μM celecoxib concentrations, which\nrelated to their growth scores ( Fig 3 ).\nComparison of growth score between control and experimental\ngroups. The growth scores were higher at the 1\nand 10 μM celecoxib concentrations; their difference with\ncontrol groups was significant (p<0.05).\nComparison of angiogenesis percent between control\nand experimental groups. The higher angiogenesis percent\nwas seen at the 10 μM celecoxib concentration.\nCellular outgrowths were visualized from\nthe endometrial explant during the culture period.\nThese projections consisted of epithelial,\nendothelial and stromal cell invasions into the\nfibrin matrix that were morphologically distinguishable\nby invert microscope ( Fig 4 ). Endothelial\ncell projections were positive for anti-\nCD31antibody ( Fig 5 ).\nInvert microscopic photographs of endometrial fragment. (A) First day of culture showing no growth and changes in the\nexplant (magnifications ×40). On the 21st day, cellular outgrowths were visualized (B; magnification: ×40) (C) 1 μM (×40 magnification),\n(D) 10 μM (magnification ×40) and (E) 50 μM( magnification ×40). (F) Angiogenesis of explant in 1 μM celecoxib\nat the end of the study (magnifications ×100).\nEndothelial cell projections (brown) that were immunohistochemically\nstained with anti-CD31 antibody.\n\nTo our knowledge, this is the first report of the\neffects of celecoxib on normal human endometrium\nin a 3D culture model. We used three celecoxib\nconcentrations (1, 10 and 50 μM). The 1 and 10\nμM concentrations of celecoxib showed significant\ngrowth and angiogenesis stimulatory effects,\nhowever the 50 μM did not considerably affect\nthe endometrial tissue; its growth score was lower\nthan the control group. We observed a correlation\nbetween growth and angiogenesis in each experimental\ngroup.\nThe 3D culture system is a relatively new, suitable\nmodel for studying endometrial tissue culture.\nThe fibrin matrix provides excellent extracellular\nmatrix for stromal, epithelial and endothelial cell\ninvasion. Gland reconstruction and endothelial\ncell proliferation (angiogenesis) have been visualized\nin this model ( 20 ,  21 ).\nThe human endometrium is a unique, dynamic\nand important tissue that has a central role in uterine\npathophysiology. It has an intensive period of\nproliferation, growth, angiogenesis and remodeling\n( 23 ).\nClinical disorders of the endometrium lead to a\nrange of gynecology problems, particularly infertility\n( 24 ). Endometrial growth and development\nis an important factor in female fertility ( 3 ). Several\nregimens have been introduced to improve a\npoor endometrium, and include estrogen therapy\nand low dose aspirin ( 25 ,  26 ). In recent years, inhibition\nof COX-2 has been researched in numerous\nstudies, particularly cancer ( 19 ). Growing\nevidence, however, suggests that the functional\nsignificance of COX-2 is far beyond what was initially\nrevealed ( 27 ,  28 ).\nCOX-2 is expressed in both eutopic and ectopic\nendometrium, although its mRNA is higher in an\nectopic endometrium. COX-2 expression is directly\nrelated to malignancy (hyperplasia and cancer),\nthus its inhibitor can be used for treatment\nof inflammatory diseases. In previous reports, the\neffects of celecoxib on endometrial stromal cells\n( 10 ), eutopic and ectopic endometrial epithelial\ncells ( 18 ) and endometrial tissue implanted outside\nof the uterus in an animal model ( 17 ) have been\ninvestigated.\nMost studies have used higher concentrations\nof celecoxib (50 μM and more) over a short\nperiod of time (3-5 days). Here, we examined\n50 μM and lower concentrations of celecoxib\n(1 and 10 μM) for longer culture periods (3\nweeks). The growth stimulatory effect of lower\ncelecoxib concentrations in the current study\nhas not been considered in previous research.\nHigher celecoxib concentration induced apoptosis\nin some cell lines ( 29 ,  30 ).\nThe growth stimulatory effect of celecoxib on\nepithelial and endothelial cells (angiogenesis) in\nour study is considerable and it can be used to\nimprove endometrial thickness in an ART cycle.\nThe angiogenic effect of lower celecoxib doses (1\nand 10 μM) in the present study contrasts the antiangiogenic\neffects of a COX-2 inhibitor, as has\nbeen previously reported ( 22 ). An explanation for\nthis contradiction is the higher (50 μM and more)\nconcentration of celecoxib studied in the previous\nresearch. Additional studies are necessary in order\nto define the mechanism of stimulating endometrial\ntissue growth and the angiogenic effect of\ncelecoxib on the human endometrium.\n\nIn this 3D culture model, the lower celecoxib\nconcentrations show stimulatory effects on normal\nhuman endometrium, whereas the higher celecoxib\nconcentration had inhibited endometrial growth.","source_license":"CC-BY-4.0","license_restricted":false}