ORIGINAL ARTICLE: ENDOMETRIOSIS

Possible involvement of signaltransducer and activator oftranscription-3 in cell–cellinteractions of peritonealmacrophages and endometrialstromal cells in human endometriosis

Fumiko Itoh, M.D.,a,b Yoshihiro Komohara, M.D., Ph.D.,b Kiyomi Takaishi, M.D., Ph.D.,a

Rituo Honda, M.D., Ph.D.,a Hironori Tashiro, M.D., Ph.D.,a Satoru Kyo, M.D., Ph.D.,c

Hidetaka Katabuchi, M.D., Ph.D.,a and Motohiro Takeya, M.D., Ph.D.b

a Department of Obstetrics and Gynecology and b Department of Cell Pathology, Graduate School of Medical Sciences,Kumamoto University, Kumamoto, Japan; and c Department of Obstetrics and Gynecology, Graduate School of MedicalScience, Kanazawa University, Kanazawa, Japan

Objective: To investigate interactions between peritoneal macrophages and endometrial stromal cells (ESCs) involved in the develop-ment of endometriosis.Design: Clinicopathologic and in vitro studies.Setting: Department of Obstetrics and Gynecology and Department of Pathology, Kumamoto University Hospital.Patient(s): Women undergoing laparoscopy or laparotomy to treat endometriosis or other benign gynecologic conditions.Intervention(s): We collected samples of peritoneal fluid (ascites), endometrium, and endometriotic tissues. We cocultured ESCsin vitro with or without human macrophages.Main Outcome Measure(s): Macrophage phenotypes in peritoneal fluid were determined via immunostaining. Proliferation of ESCsand activation of signal transducer and activator of transcription-3 (Stat3) in cocultures were evaluated.Result(s): The endometriosis group had a significantly higher total number of macrophages in ascites compared with the control group,but the ratios of CD163þ alternatively activated macrophages (M2) in the two groups did not differ significantly. Coculture with M2macrophages significantly up-regulated ESC proliferation and Stat3 activation in ESCs in vitro. Proliferation of ESCs was suppressedafter Stat3 was down-regulated by small interfering RNA. Stat3 was activated in epithelial cells and ESCs in human endometriotic

Use your smartphone

lesions.Conclusion(s): Interactions between M2 macrophages and ESCs via Stat3 activation may playan important role in the development of endometriosis. (Fertil Steril� 2013;-:-–-. �2013by American Society for Reproductive Medicine.)Key Words: Endometriosis, macrophages, endometrial stromal cells, Stat3

Discuss: You can discuss this article with its authors and with other ASRM members at http://fertstertforum.com/itohf-stat3-macrophages-endometriosis/

to scan this QR codeand connect to thediscussion forum forthis article now.*

* Download a free QR code scanner by searching for “QRscanner” in your smartphone’s app store or app marketplace.

Received September 3, 2012; revised December 8, 2012; accepted January 21, 2013.F.I. has nothing to disclose. Y.K. has nothing to disclose. K.T. has nothing to disclose. R.H. has nothing

to disclose. H.T. has nothing to disclose. S.K. has nothing to disclose. H.K. has nothing to disclose.M.T. has nothing to disclose.

This studywas supported in part byGrants-in-Aid for Scientific Research (C21592137 to H.K., 21790388to M.T.) from the Ministry of Education, Culture, Sports, Science, and Technology of Japan.

F.I. and Y.K. contributed equally to this work.Reprint requests: Yoshihiro Komohara, M.D., Ph.D., Kumamoto University, Graduate School of

Medical Sciences, Department of Cell Pathology, Honjo 1-1-1, Chuou-ku, Kumamoto 860-8556,Japan (E-mail: ycomo@kumamoto-u.ac.jp).

Fertility and Sterility® Vol. -, No. -, - 2013 0015-0282/$36.00Copyright ©2013 American Society for Reproductive Medicine, Published by Elsevier Inc.http://dx.doi.org/10.1016/j.fertnstert.2013.01.133

VOL. - NO. - / - 2013

E ndometriosis is a common gyne-cologic disorder in which growthof endometrial-like tissue occurs

outside the uterine cavity (1). Thisdisorder affects at least 5%–10% of allwomen of reproductive age and up to30% of infertile women (1–3). Surgicalremoval of lesions and hormonalsuppression are the current goldstandards of therapy, but both

1

ORIGINAL ARTICLE: ENDOMETRIOSIS

approaches are associated with various side effects and a highincidence of relapse (4, 5).

The importance of peritoneal macrophages in thedevelopment of endometriosis has been known for the past3 decades (6–12). Peritoneal fluid volume is increased andthe total number of macrophages is elevated in ascites frompatients with peritoneal endometriosis (6–8). Severalmacrophage-derived cytokines, such as interleukin (IL)-1b,IL-6, IL-8, monocyte chemoattractant protein-1 (MCP-1),tumor necrosis factor-a, and transforming growth factor-b,also have elevated levels in ascites from patients with pelvicendometriosis (13–17). These cytokines stimulateendometrial stromal cells (ESCs) and are believed to beassociated with endometriosis development (18–20). Inaddition, after macrophages are stimulated with solublefactors derived from ESCs, they secrete many factors,including IL-8, that activate ESCs (21–23). These findingsindicate that interactions between macrophages and ESCsmay be associated with the development of endometriosis.

A novel concept of macrophage phenotypes was recentlyproposed. Interferon-g induces the classic activation ofmacrophages, whereas anti-inflammatory cytokines, suchas IL-10, IL-4, and IL-13, stimulate alternative activation(24–26). These macrophage phenotypes are called M1 andM2, respectively. CD163 is a marker of M2 macrophagesthat is believed to be polarized to the alternatively activatedor anti-inflammatory phenotype (24–27). M2 peritonealmacrophages were closely related to the development ofendometriosis in a murine endometriosis model in that theyinduced angiogenesis in lesions (28).

We previously demonstrated that activation of signaltransducer and activator of transcription-3 (Stat3) is involvedin interactions betweenmacrophages and ovarian cancer cells(29). Stat3 is a key signal transducer and regulator of macro-phage activation associated with several oncogenic signalingpathways, such as proliferation, survival, angiogenesis, andimmunosuppression (30). On the basis of all these data, we hy-pothesized that interactions between macrophages and ESCsmay be important in the development of endometriosis andthat Stat3 activation may be associated with interactions be-tween macrophages and ESCs.

MATERIALS AND METHODSHuman Samples

Samples of ascites were collected from 25 female patientswith endometriosis and a control group of 15 female patientswithout endometriosis (but with uterine fibroma or ovariandermoid cyst) who were scheduled to undergo laparoscopyor laparotomy at Kumamoto University Hospital. Womenwho had received steroid hormone therapy within 3 monthsof surgery were excluded from the study. Peritoneal fluidsamples were transported to the laboratory under sterileconditions and were used immediately. Endometrial sampleswere cut from uteri collected by laparotomy from womenwithout endometriosis. Endometriotic lesions were obtainedafter surgical removal by laparotomy. All samples manifestedsuperficial peritoneal endometriosis (red and black lesions).The women with endometriosis all had peritoneal lesions

2

with endometriotic tissue. The presence or absence ofendometriosis was confirmed visually at surgery and byhistologic analysis. The extent of endometriosis was scoredaccording to the American Society for Reproductive Medicineclassification (31). The phase of the menstrual cycle wasdetermined according to the most recent menstrual periodand was confirmed by visualization of a corpus luteum ordominant follicle, or by histologic dating of the endometrium.All patients provided written informed consent for participa-tion in the study. The Kumamoto University Hospital ReviewBoard approved the study protocol (Project 1370).

Immunostaining of Macrophages in PeritonealFluid

Cytospin (Shandon) specimens of ascites on glass slides werefixed in 4% paraformaldehyde for 10 minutes and immedi-ately kept at �80�C. Immunostaining was started at thesame time. Slides were incubated with mouse anti-CD68(PG-M1; Dako) or anti-CD163 (AM-3K; TransGenic)antibodies. After use of the secondary antibody horseradishperoxidase–labeled goat anti-mouse immunoglobulin(Nichirei), macrophages were visualized with diaminobenzi-dine. CD163 expression was scored semiquantitativelyaccording to staining intensity (1, weak; 2, intermediate; 3,strong). Normal mouse IgG and rabbit IgG (Dako) were usedas negative controls, and these control antibodies detectedno positive signals. Although M1 macrophages also expressCD163, CD163 protein expression by M1 macrophages is tooweak to be detected by routine immunohistochemicalmethods(32). We therefore used CD163 as the M2 marker in this study.

ESCs and Cell Lines

PrimaryESCs (pESCs)were isolatedbydigestionof endometrialtissue fragmentswith 0.5%collagenase as previously described(33). Samples of endometrium at the proliferative phase wereobtained from premenopausal patients who had undergonehysterectomy for uterine fibromas and had no evidence ofendometriosis. To establish immortalized ESCs (iESCs), stromalcells isolated from noncancerous endometrium were purifiedand transduced with the human telomerase reverse transcrip-tase gene by using retroviral infections as described previously(34). We confirmed that these cells expressed humantelomerase reverse transcriptase, maintained the structuralcharacteristics of ESCs, and continued to divide at least overPD50. Primary ESCs and iESCs were cultured in Dulbecco’smodified Eagle’s medium (DMEM) supplemented with 10%fetal bovine serum (FBS). Primary ESCs in monolayer cultureafter the third passage were>99% pure according to immuno-cytochemical staining with antibodies (all from Dako) tovimentin (V9), CD10 (SS2/36), cytokeratin factor, andleukocyte common antigen (2B11 þ PD7/26).

Coculture Assays

Peripheral blood mononuclear cells were obtained from threehealthy adult female volunteer donors without endometriosis.CD14þ monocytes were purified from peripheral bloodmononuclear cells by positive selection, via magnetic

VOL. - NO. - / - 2013

Fertility and Sterility®

activated cell sorting technology (Miltenyi Biotec).Monocytes (5 � 106 cells per 10-cm dish, 5 � 105 cells perwell in 12-well culture plates, and 5 � 104 cells per well in96-well culture plates) were cultured with granulocytemacrophage colony-stimulating factor (GM-CSF) (5 ng/mL;Wako) and macrophage colony-stimulating factor (M-CSF)(50 ng/mL;Wako) for up to 7 days to induceM2macrophages.After monocyte-derived macrophages were washed inphosphate-buffered saline, they were cocultured with pESCsor iESCs (1:1 ratio) in DMEM/F12 medium without FBS.

Preparation of Conditioned Medium

Macrophages and iESCswere cocultured inDMEM/F12withoutFBS for 3 days. Coculture conditioned medium (cCM) was col-lected and stored at�80�C. Macrophages and iESCs were alsocultured separately in DMEM/F12 without FBS for 3 days, andmacrophage CM (mCM) and iESC CM (eCM) were collected.

Counting Cocultured Cells

Primary ESCs or iESCs were cocultured for 3 days with or with-outhumanmacrophages in12-well plates. Cocultured cellswereimmunostained for CD204, which is expressed by macrophagesbut not ESCs; we then counted CD204� ESCs. Anti-CD204antibody (clone SRA-E5) was purchased from TransGenic.

5-Bromo-2-deoxyuridine Labeling Assay

Immortalized ESCs were cocultured for 1 day with or withouthuman macrophages in 12-well culture plates and were thenincubated with 5-bromo-2-deoxyuridine (BrdU; Roche) for 2hours. The cells were fixed in cold acetone, stained withanti-CD204 antibody, and visualized by using diaminobenzi-dine. After anti-CD204 antibody was detached in glycinebuffer (pH 2.2), cells were immersed in FixDenat solution(Roche), incubated with anti-BrdU antibody (clone IIB5;Santa Cruz Biotechnology) and anti-horseradish peroxi-dase–labeled mouse IgG (Fab fragment; GE Healthcare), andvisualized using HistoGreen (Linaris).

Labeling with BrdU was also investigated using ELISA.Cells were seeded in 96-well plates and incubated for 24 hoursbefore culture with BrdU for 2 hours. Fixed cells wereevaluated with a BrdU ELISA kit (Roche).

Cell Proliferation Assay

Immortalized ESCs were seeded in 96-well plates and thenstimulated by CMs with or without the Stat3 inhibitorcorosolic acid (35) for 2 days. Cells were evaluated by usingthe WST assay (Dojindo) according to the manufacturer’sinstructions. Neutralizing anti–GM-CSF polyclonal antibodywas purchased from GeneTex.

Evaluation of Stat3 Phosphorylation

Stat3 activation in iESCs was evaluated by means of Westernblotting, and Stat3 activation in cocultured cells was investi-gated by double immunostaining as described earlier (27). Inbrief, paraffin-embedded blocks of cells fixed in 4% parafor-maldehyde were cut and stained with anti-pStat3 (Tyr 705)

VOL. - NO. - / - 2013

(rabbit monoclonal, clone D3A7; Cell Signaling) and anti-CD10 (mouse monoclonal, clone 56C6; Nichirei) antibodies,after which Stat3 phosphorylation was visualized via diami-nobenzidine (Vector Laboratories) and Fast Red solution.

Small Interfering RNA in iESCs

Immortalized ESCs were transfected with small interferingRNA (siRNA) against human Stat3 (Santa Cruz Biotechnol-ogy) by using Lipofectamine RNAiMAX (Invitrogen). ControlsiRNA (Santa Cruz Biotechnology) was the negative control.

Immunohistochemistry

Primary Stat3 expression in human samples was investigatedby immunostaining as follows: 3-mm-thick paraffin-embedded tissues of eutopic and ectopic endometrium weredeparaffinized and processed for antigen retrieval. Sectionswere stained with mouse anti-CD163 (10D6; Novocastra)and rabbit anti-pStat3 (Tyr 705) (rabbit monoclonal, cloneD3A7; Cell Signaling) antibodies. The reaction was visualizedvia diaminobenzidine (Vector Laboratories). Two observers(Y.K. and F.I.) evaluated the findings in blind fashion.

Cytokine Array

The soluble factors in CM were quantified by means of theCytokine Array Kit (R&D Systems) according to the manufac-turer’s instructions.

Polymerase Chain Reaction

Total RNA from Gr1þ cells extracted with RNAiso Plus(TaKaRa) was reverse-transcribed using the PrimeScript RTreagent kit (TaKaRa) and was amplified by polymerase chainreaction (PCR) using DNA polymerase (TaKaRa) and thefollowing primers: GM-CSF: forward 50-CACTGCTGCTGAGATGAATG-30 and reverse 50-AATCTGGGTTGCACAGGAAG-30; GM-CSFRa: 50-GGAGATCCGGTGTCCTTATTAC-30

and 50-AACGTACGGTGACATTGCTG-30; GM-CSFRb: 50-AGACCGAGACCCTCCAGAAC-30 and 50-TGAGGAAGATCACGATGAGG-30; C-C chemokine receptor type 2 (CCR2): 50- TACCAACGAGAGCGGTGAAG-30 and 50-AAAGATGAACACCAGCGAGTAGAG-30; and b-actin: 50-CTGTGGCATCCACGAAACTAC-30 and 50-CTGATCCACATCTGCTGGAAG-30.

Statistics

All in vitro data represent two or three independent experi-ments (n ¼ 3 or 4 samples each) and are expressed as means� SD. Pairwise comparisons were analyzed using theMann-Whitney U test study and Student’s t test. A P valueof < .05 was considered statistically significant.

RESULTSAscites Macrophages Polarized to the M2Phenotype

We first evaluated macrophage phenotypes in ascites fromage-matched patients with endometriosis (endometriosisgroup) and those without endometriosis and cancer (control

3

ORIGINAL ARTICLE: ENDOMETRIOSIS

group) at the follicular phase (n ¼ 12 and n ¼ 7, respectively)and the luteal phase (n ¼ 13 and n ¼ 8, respectively). Theendometriosis group had a significantly elevated ascitesvolume compared with the controls (control group 5.5 � 3.1mL, endometriosis group 15.3 � 12.1 mL; P< .01). CD68and CD163 were used as pan-macrophage and M2 macro-phage markers, respectively. Immunostaining identifiedmany CD68þ and CD163þ macrophages in ascites fromboth groups (Fig. 1A). The total number of CD68þ andCD163þ macrophages in ascites from the endometriosisgroup was significantly higher than that in ascites from thecontrol group (Fig. 1A). Although the concentration of bothmacrophage types and the ratio of CD163þ to CD68þmacro-phages did not differ significantly, the endometriosis grouptended to have a higher ratio (Fig. 1B). The volume of ascites

FIGURE 1

Phenotypes of macrophages in ascites from patients with and withoutendometriosis contained many CD68þ and CD163þ macrophages. Theliving cells in the ascites and the percentage of CD68þ or CD163þ cells i(C) CD163 expression was scored, and the difference between the endothe Mann-Whitney U test.Itoh. Stat3 and macrophages in endometriosis. Fertil Steril 2013.

4

and the number of macrophages were not associated with thestage of endometriosis or menstrual cycle (data not shown),and CD163 staining intensity did not differ significantly be-tween the two groups (Fig. 1C).

ESC Proliferation in Coculture with Macrophages

We investigated the significance of macrophages for prolifer-ation of ESCs by using cocultured cells. M2macrophages weredifferentiated from monocytes by stimulation with GM-CSFand M-CSF and were then cultured with iESCs or pESCs for3 days. For evaluation of the ESC number in coculture, cellswere immunostained with CD204 (a macrophage marker).The numbers of iESCs and pESCs increased significantly aftercoculture with macrophages (Fig. 2A). The percentage of

endometriosis. (A) Cytospin specimens of ascites from patients withtotal number of macrophages was calculated from the number of

n the living cells. (B) CT ¼ control group; EM ¼ endometriosis group.metriosis and control groups was evaluated. Data were analyzed via

VOL. - NO. - / - 2013

FIGURE 2

Proliferation of ESCs cocultured with M2 macrophages (Mac). (A) Immortalized ESCs or pESCs were cocultured with and without M2 macrophages.Specific CD204 immunostaining distinguished macrophages from ESCs in cocultures. (B, C) Immortalized ESCs were cocultured with and withoutM2 macrophages. Endometrial stromal cell proliferation was evaluated via BrdU labeling. Ratios (%) of BrdUþ ESCs were counted after doubleimmunostaining for BrdU and CD204. In addition, BrdU incorporation into ESCs was evaluated using ELISA. (D) After culture with CM from culturedmacrophages (mCM) or cocultured cells (cCM), iESC proliferation was evaluated via WST assays. Data were analyzed using Student’s t test. *P< .05.Itoh. Stat3 and macrophages in endometriosis. Fertil Steril 2013.

Fertility and Sterility®

BrdU-labeled iESCs also significantly increased after cocul-ture with macrophages (Fig. 2B and C).

To determine the effects of macrophage-derived solublefactors, iESCs were incubated with mCM or cCM. Both mediasignificantly enhanced iESC proliferation, but cCM did so toa greater extent (Fig. 2D).

Stat3 Activation of ESCs in Coculture withMacrophages

Stat3 is an important signal molecule that is related to cellproliferation, so we evaluated Stat3 activation in coculturedcells. Immunostaining revealed activated Stat3 in ESCs and

VOL. - NO. - / - 2013

macrophages (Fig. 3A). In addition, cCM induced Stat3activation in cocultured ESCs (Fig. 3B). To determine thesignificance of Stat3 for ESC proliferation, siRNA againstStat3 was added to iESC cultures. Proliferation of ESCscultured with cCM was significantly suppressed after Stat3was down-regulated with siRNA (Fig. 3C). The Stat3 inhibitorcorosolic acid inhibited ESC proliferation induced by cCM butnot in the absence of cCM (Fig. 3D).

Stat3 Activation in Human Endometriotic Lesions

Stat3 activation in human endometriosis was evaluated viaimmunostaining. M2 macrophages infiltrated around stromal

5

FIGURE 3

Stat3 activation in ESCs coculturedwithM2macrophages (Mac). (A) Cocultured cells and control iESCswere double-immunostained for pStat3 andCD10 (an ESC marker). (B) Immortalized ESCs were stimulated with CM from cultured macrophages (mCM) or cocultured cells (cCM), and Stat3phosphorylationwas analyzed usingWestern blotting. (C) Following Stat3, proteinwas silenced in iESC, coculturewas started, and total numbers ofiESC were evaluated. (D) The inhibitory effect of a Stat3 inhibitor, corosolic acid (CA; final concentration, 10 mM), on ESC proliferation wasinvestigated using WST assays. Data were analyzed using Student’s t test. *P< .05.Itoh. Stat3 and macrophages in endometriosis. Fertil Steril 2013.

ORIGINAL ARTICLE: ENDOMETRIOSIS

cells, and Stat3 was activated in both epithelial and stromalcells in endometriotic lesions (Fig. 4A). Both epithelial cellsand ESCs evidenced Stat3 activation during all phases ofthe menstrual cycle in normal endometrium, but activationwas notably higher in both cell types during the proliferativephase (Fig. 4B).

GM-CSF Associated with ESC Proliferation

The results of the cytokine array showed significantlyup-regulated GM-CSF, RANTES (regulated upon activation,normal T-cell expressed, and secreted), IL-1 receptor antago-nist (IL-1RA), and MCP-1 production in cCM (SupplementalFig. 1A, available online). Monocyte chemoattractantprotein-1 activates Stat3 signals via CCR2 (36), but in this

6

study CCR2 expression was undetectable in ESCs(Supplemental Fig. 1B). We then investigated GM-CSF,because it is thought to act as a growth factor in some typesof cancers, and it activates Stat5 and Stat3 in neutrophils(37, 38). Macrophages and ESCs expressed GM-CSF and GM-CSF receptor (GM-CSFR) mRNAs (Supplemental Fig. 1B), andGM-CSF significantly and dose-dependently induced ESCproliferation (Supplemental Fig. 1C). However, GM-CSF didnot induce Stat3 activation in ESCs (data not shown).

DISCUSSIONImplanted endometrial tissue, remnants of the m€ulleriansystem, and metaplasia of the coelomic epithelium arebelieved to cause endometriosis (39). However, the etiology

VOL. - NO. - / - 2013

FIGURE 4

Immunostaining of pStat3 and CD163 in human endometriotic lesions and normal endometrium. (A) Tissue sections of pelvic endometriotic lesionswere stained with anti-pStat3 and anti-CD163 antibodies. (B) Sections of normal endometrium from patients with uterine fibromas were stainedwith anti-pStat3 antibody. Data were analyzed using Student’s t test.Itoh. Stat3 and macrophages in endometriosis. Fertil Steril 2013.

Fertility and Sterility®

and pathophysiology of the disorder remain poorly under-stood. Endometriosis frequently develops in pelvic organs(e.g., the ovaries, Douglas’ pouch, and peritoneum), whichsuggest that the peritoneal environment that contains variousfree-floating cells such as macrophages and endometrial cellsderived from refluxed menstrual blood may contribute to theestablishment and progression of endometriosis.

We previously reported that the proportion of activemacrophages was elevated in ascites from patients withendometriosis (8, 10), and in another study, activatedmacrophages reportedly released several cytokines into theperitoneal cavity (12). In the present study we first analyzedmacrophage phenotypes in the peritoneal cavity of womenwith and without endometriosis. Flow cytometry(fluorescence-activated cell sorting) has identifiedsignificantly increased CD163 expression in peritonealmacrophages from patients with endometriosis (28). Ourpresent immunostaining techniques indicated that the totalnumber of CD163þ macrophages in ascites from theendometriosis group was significantly higher than thatof the control group. However, the ratio of CD163þ toCD68þ macrophages in the two groups did not differsignificantly. Our observations thus did not agree with thefindings of the previous study. The methodology used mayexplain this discrepancy: fluorescence-activated cell sortingdetects molecules expressed on the cell surface, whereasimmunostaining detects intracytoplasmic proteins in additionto surface molecules.

Endometrial stromal cells as well as macrophages canproduce and secrete growth factors and inflammatorymediators into the peritoneal fluid. Changes in cellularfunctions such as proliferation in ESCsmay have considerableeffects on the peritoneal environment.We previously reportedthe presence of glandular cell metaplasia of the mesothelium

VOL. - NO. - / - 2013

when human ovarian surface epithelial cells and ESCs werecocultured with 17b-estradiol (40), which suggests thatESCs are a major factor contributing to the development ofendometriosis. On the basis of these findings, we investigatedwhether peritoneal macrophages may play a principal role inESC proliferation and lead to development of pelvicendometriosis.

Mitogen-activated protein kinase and nuclear factor-kBtranscriptional activities are known to regulate cell prolifer-ation in endometrial cells (41–44). The transcription factorStat3, which is crucial for survival, proliferation, anddifferentiation of cells, is activated by many cytokines andgrowth factors. Stat3 is also activated in endometrialglandular epithelium and stroma during the secretoryphase and is associated with cell adhesion (45). However,the role of the Stat3 signaling pathway in regulation ofESC proliferation and in development of endometriosisremains unknown. Takaishi et al. (29) found that manyM2 macrophages occurred in ascites in advanced epithelialovarian cancer and that M2 macrophages activated Stat3in ovarian cancer cells, which resulted in theirproliferation. In this study we observed similar interactionsbetween ESCs and M2 macrophages. In fact, theproliferation and Stat3 activation of ESCs weresignificantly up-regulated by coculture with M2 macro-phages; cCM also induced ESC proliferation and Stat3 acti-vation. These findings suggest that soluble factors derivedfrom macrophages and activated ESCs are involved in inter-actions between macrophages and ESCs. We therefore inves-tigated which cytokines may cause Stat3 activation in ESCs.In addition, because these findings were obtained mainlyfrom an in vitro study, additional investigations are neces-sary to demonstrate the significance of macrophages andStat3 in endometriosis development.

7

ORIGINAL ARTICLE: ENDOMETRIOSIS

The results of our cytokine array showed up-regulation ofGM-CSF production in CM from ESCs cocultured with macro-phages. The activation and differentiation of hematopoieticmyeloid cells are closely associated with GM-CSF (46), whichbinds to a dimeric receptor (GM-CSFR) consisting of a- and b-chain subunits. GM-CSFR is expressed in myeloid cells andother tumor cell types (47, 48). Many kinds of tumor cellsexpressing GM-CSFR secrete GM-CSF, which stimulates tu-mor cell proliferation in an autocrine manner (48, 49). Herewe demonstrated that GM-CSF was associated with interac-tions between ESCs and macrophages. Our current PCR anal-ysis showed that macrophages and ESCs expressed GM-CSFand GM-CSFR mRNAs, which suggests that GM-CSF that issecreted by both cell types induces ESC proliferation via anautocrine or a paracrine loop. However, a neutralizing anti-body against GM-CSF did not suppress the proliferation ofcCM-stimulated ESCs (data not shown). Unknown soluble fac-tors other than GM-CSF may also be associated with interac-tions between ESCs and macrophages.

We also found that RANTES and IL-1RA were up-regulated in cocultured cells. The effect of IL-1RA on endome-triosis was not reported until rather recently. Macrophagesstimulate RANTES production from ESCs (50), and secretedRANTES induces progression of endometriosis by means ofmacrophage recruitment (51). In this study, we did not iden-tify which cytokines could induce Stat3 activation. Addi-tional studies are needed to clarify details of themechanisms of interactions between ESCs and macrophages.

In the present study, we focused on interactions betweenESCs and macrophages; however, we also detected Stat3 acti-vation in epithelial cells in endometriotic lesions. Jasper et al.(52) and Nakamura et al. (53) recently investigated interactionsbetween endometrial epithelial cells and macrophages andfound that macrophage-derived IL-6 and leukemia inhibitoryfactor stimulated epithelial cells and promoted endometrial im-plantation in early pregnancy. Because these cytokines arewell-known molecules involved in Stat3 activation, similarmechanismsmay pertain in the development of endometriosis.

In conclusion, our findings from in vitro and in vivo stud-ies indicate that the interactions between M2 macrophagesand ESCs via Stat3 activation may play indispensable rolesin the development of endometriosis. Targeting Stat3 signalsor the regulation of macrophage function may aid the treat-ment of patients with endometriosis.

Acknowledgments: The authors thank Prof. Kaei Nasu andDr. Yukiko Kawano (Department of Obstetrics and Gynecol-ogy, Faculty of Medicine, Oita University, Oita, Japan) fortechnical support with the primary ESC culture; and Ms.Emi Kiyota, Mr. Osamu Nakamura, and Ms. Yui Hayashida(Department of Cell Pathology, Graduate School of MedicalSciences, Kumamoto University, Kumamoto, Japan) for theirtechnical assistance.

REFERENCES1. Galle PC. Clinical presentation and diagnosis of endometriosis. Obstet Gyne-

col Clin North Am 1989;16:29–42.2. Bergqvist A. Different types of extragenital endometriosis: a review. Gynecol

Endocrinol 1993;7:207–21.

8

3. Simoens S, Hummelshoj L, D’Hooghe T. Endometriosis: cost estimates andmethodological perspective. Hum Reprod Update 2007;13:395–404.

4. Vercellini P, Fedele L, Aimi G, De Giorgi O, Consonni D, Crosignani PG. Re-productive performance, pain recurrence and disease relapse after conser-vative surgical treatment for endometriosis: the predictive value of thecurrent classification system. Hum Reprod 2006;21:2679–85.

5. Bergqvist A. Current drug therapy recommendations for the treatment ofendometriosis. Drugs 1999;58:39–50.

6. Bulun SE. Endometriosis. New Engl J Med 2009;360:268–79.7. Halme J, Becker S, HammondMG, Raj MH, Raj S. Increased activation of pel-

vic macrophages in infertile women with mild endometriosis. Am J ObstetGynecol 1983;145:333–7.

8. Miyamura S, Okamura H, Naito M, Takahashi K. Intracytochemical localiza-tion of peroxidase activity in human peritoneal macrophages. J Clin ElectronMicroscop 1987;20:449–50.

9. Fukumatsu Y, Katabuchi H, Miyamura S, Matsuura K, Okamura H. Perito-neal macrophages in endometriosis. In: Hafez ESE, editor. Assisted humanreproductive technology. New York: Hemisphere Publishing; 1991:112–9.

10. Fukumatsu Y, Katabuchi H, Miyamura S, Matsuura K, Okamura H, Naito M,et al. Activatedmacrophages in the peritoneal fluid of womenwith endome-triosis: examination of the intracytoplasmic localization of endogenous per-oxidase and interleukin-1. Acta Obstet Gynaecol Jap 1992;44:529–36.

11. Gazvani R, Templeton A. Peritoneal environment, cytokines and angio-genesis in the pathophysiology of endometriosis. Reproduction 2002;123:217–26.

12. Okamura H, Katabuchi H, Kanzaki H. Macrophages in reproductive biology.In: Burke B, Lewis CE, editors. The macrophage. 2nd ed. New York: OxfordUniversity Press; 2002:549–76.

13. Khan KN, Kitajima M, Hiraki K, Fujishita A, Sekine I, Ishimaru T, et al.Immunopathogenesis of pelvic endometriosis: role of hepatocyte growthfactor, macrophages and ovarian steroids. Am J Reprod Immunol 2008;60:383–404.

14. Defrere S, Gonzalez-Ramos R, Lousse JC, Colette S, Donnez O, Donnez J,et al. Insights into iron and nuclear factor-kappa B (NF-kappaB) involvementin chronic inflammatory processes in peritoneal endometriosis. Histol Histo-pathol 2011;26:1083–92.

15. Mori H, Sawairi M, NakagawaM, Itoh N, Wada K, Tamaya T. Peritoneal fluidinterleukin-1 beta and tumor necrosis factor in patients with benign gyneco-logic disease. Am J Reprod Immunol 1991;26:62–7.

16. Koyama N, Matsuura K, Okamura H. Cytokines in the peritoneal fluid of pa-tients with endometriosis. Int J Gynaecol Obstet 1993;43:45–50.

17. Arici A, Tazuke SI, Attar E, Kliman HJ, Olive DL. Interleukin-8 concentration inperitoneal fluid of patients with endometriosis and modulation ofinterleukin-8 expression in human mesothelial cells. Mol Hum Reprod1996;2:40–5.

18. Yih S, Katabuchi H, Araki M, Matsuura K, Takeya M, Takahashi K, et al. Ex-pression of monocyte chemoattractant protein-1 in peritoneal endometri-otic cells. Virchows Arch 2001;438:70–7.

19. Nishida M, Nasu K, Narahara H. Role of chemokines in the pathogenesis ofendometriosis. Front Biosci (Elite Ed) 2011;3:1196–204.

20. Gazvani R, Smith L, Fowler PA. Effect of interleukin-8 (IL-8), anti-IL-8, and IL-12 on endometrial cell survival in combined endometrial gland and stromalcell cultures derived from women with and without endometriosis. FertilSteril 2002;77:62–7.

21. Bullimore DW. Endometriosis is sustained by tumour necrosis factor-alpha.Med Hypotheses 2003;60:84–8.

22. Khan KN, Masuzaki H, Fujishita A, Kitajima M, Hiraki K, Sekine I, et al. Inter-leukin-6- and tumour necrosis factor alpha-mediated expression of hepato-cyte growth factor by stromal cells and its involvement in the growth ofendometriosis. Hum Reprod 2005;20:2715–23.

23. Eyster KM, Hansen KA, Winterton E, Klinkova O, Drappeau D, Mark-Kappeler CJ. Reciprocal communication between endometrial stromal cellsand macrophages. Reprod Sci 2010;17:809–22.

24. Gordon S. Alternative activation of macrophages. Nat Rev Immunol 2003;3:23–35.

25. Pollard JW. Tumour-educated macrophages promote tumour progressionand metastasis. Nat Rev Cancer 2004;4:71–8.

VOL. - NO. - / - 2013

Fertility and Sterility®

26. Sica A, Schioppa T, Mantovani A, Allavena P. Tumour-associated macro-phages are a distinct M2 polarised population promoting tumour progres-sion: potential targets of anti-cancer therapy. Eur J Cancer 2006;42:717–27.

27. XuW, Schlagwein N, Roos A, van den Berg TK, Daha MR, van Kooten C. Hu-man peritoneal macrophages show functional characteristics ofM-CSF-drivenanti-inflammatory type 2 macrophages. Eur J Immunol 2007;37:1594–9.

28. Bacci M, Capobianco A, Monno A, Cottone L, Di Puppo F, Camisa B, et al.Macrophages are alternatively activated in patients with endometriosis andrequired for growth and vascularization of lesions in a mouse model of dis-ease. Am J Pathol 2009;175:547–56.

29. Takaishi K, Komohara Y, Tashiro H, Ohtake H, Nakagawa T, Katabuchi H,et al. Involvement of M2-polarized macrophages in the ascites from ad-vanced epithelial ovarian carcinoma in tumor progression via Stat3 activa-tion. Cancer Sci 2010;101:2128–36.

30. Yu H, Kortylewski M, Pardoll D. Crosstalk between cancer and immune cells:role of STAT3 in the tumour microenvironment. Nat Rev Immunol 2007;7:41–51.

31. Revised American Society for Reproductive Medicine classification of endo-metriosis: 1996. Fertil Steril 1997;67:817–21.

32. Komohara Y, Hirahara J, Horikawa T, Kawamura K, Kiyota E, Sakashita N,et al. AM-3K, an anti-macrophage antibody, recognizes CD163, a moleculeassociated with an anti-inflammatory macrophage phenotype. J HistochemCytochem 2006;54:763–71.

33. Kawano Y, Nasu K, Li H, Tsuno A, Abe W, Takai N, et al. Application of thehistone deacetylase inhibitors for the treatment of endometriosis: histonemodifications as pathogenesis and novel therapeutic target. Hum Reprod2011;26:2486–98.

34. Kyo S, Nakamura M, Kiyono T, Maida Y, Kanaya T, Tanaka M, et al. Success-ful immortalization of endometrial glandular cells with normal structural andfunctional characteristics. Am J Pathol 2003;163:2259–69.

35. Fujiwara Y, Komohara Y, Ikeda T, Takeya M. Corosolic acid inhibits glioblas-toma cell proliferation by suppressing the activation of signal transducer andactivator of transcription-3 and nuclear factor-kappa B in tumor cells andtumor-associated macrophages. Cancer Sci 2011;102:206–11.

36. Mellado M, Rodríguez-Frade JM, Aragay A, del Real G, Martín AM, Vila-Coro AJ, et al. The chemokine monocyte chemotactic protein 1 triggers Ja-nus kinase 2 activation and tyrosine phosphorylation of the CCR2B receptor.J Immunol 1998;161:805–13.

37. FortinCF, Larbi A, DupuisG, LesurO, Fulop T Jr.GM-CSFactivates the Jak/STATpathway to rescuepolymorphonuclear neutrophils from spontaneous apopto-sis in young but not elderly individuals. Biogerontology 2007;8:173–87.

38. Revoltella RP, Menicagli M, Campani D. Granulocyte-macrophage colony-stimulating factor as an autocrine survival-growth factor in human gliomas.Cytokine 2012;57:347–59.

39. Nisolle M, Donnez J. Peritoneal endometriosis, ovarian endometriosis, andadenomyotic nodules of the rectovaginal septum are three different entities.Fertil Steril 1997;68:585–96.

40. Ohtake H, Katabuchi H,Matsuura K, Okamura H. A novel in vitro experimen-tal model for ovarian endometriosis: the three-dimensional culture of

VOL. - NO. - / - 2013

human ovarian surface epithelial cells in collagen gels. Fertil Steril 1999;71:50–5.

41. Yoshino O, Osuga Y, Hirota Y, Koga K, Hirata T, Harada M, et al. Possiblepathophysiological roles of mitogen-activated protein kinases (MAPKs) inendometriosis. Am J Reprod Immunol 2004;52:306–11.

42. Gonzalez-Ramos R, Van Langendonckt A, Defrere S, Lousse JC, Colette S,Devoto L, et al. Involvement of the nuclear factor-kappaB pathway in thepathogenesis of endometriosis. Fertil Steril 2010;94:1985–94.

43. Yamanaka K, Xu B, Suganuma I, Kusuki I, Mita S, Shimizu Y, et al. Dienogestinhibits aromatase and cyclooxygenase-2 expression and prostaglandin E(2)production in human endometriotic stromal cells in spheroid culture. FertilSteril 2012;97:477–82.

44. Li MQ, Li HP, Meng YH, Wang XQ, Zhu XY, Mei J, et al. Chemokine CCL2enhances survival and invasiveness of endometrial stromal cells in an auto-crine manner by activating Akt andMAPK/Erk1/2 signal pathway. Fertil Steril2012;97:919–29.

45. Marwood M, Visser K, Salamonsen LA, Dimitriadis E. Interleukin-11and leukemia inhibitory factor regulate the adhesion of endometrial ep-ithelial cells: implications in fertility regulation. Endocrinology 2009;150:2915–23.

46. Metcalf D. The granulocyte-macrophage colony-stimulating factors. Science1985;229:16–22.

47. Miyagawa K, Chiba S, Shibuya K, Piao YF, Matsuki S, Yokota J, et al. Fre-quent expression of receptors for granulocyte-macrophage colony-stimulat-ing factor on human nonhematopoietic tumor cell lines. J Cell Physiol 1990;143:483–7.

48. Harris RJ, Pettitt AR, Schmutz C, Sherrington PD, Zuzel M, Cawley JC, et al.Granulocyte-macrophage colony-stimulating factor as an autocrine survivalfactor for mature normal and malignant B lymphocytes. J Immunol 2000;164:3887–93.

49. Tachibana M, Miyakawa A, Tazaki H, Nakamura K, Kubo A, Hata J,et al. Autocrine growth of transitional cell carcinoma of the bladder in-duced by granulocyte-colony stimulating factor. Cancer Res 1995;55:3438–43.

50. Lebovic DI, Chao VA, Taylor RN. Peritoneal macrophages induce RANTES(regulated on activation, normal T cell expressed and secreted) chemokinegene transcription in endometrial stromal cells. J Clin Endocrinol Metab2004;89:1397–401.

51. Wang XQ, Yu J, Luo XZ, Shi YL, Wang Y, Wang L, et al. The high levelof RANTES in the ectopic milieu recruits macrophages and induces theirtolerance in progression of endometriosis. J Mol Endocrinol 2010;45:291–9.

52. Jasper MJ, Care AS, Sullivan B, Ingman WV, Aplin JD, Robertson SA. Macro-phage-derived LIF and IL1B regulate alpha(1,2)fucosyltransferase 2 (Fut2) ex-pression inmouse uterine epithelial cells during early pregnancy. Biol Reprod2011;84:179–88.

53. Nakamura H, Jasper MJ, Hull ML, Aplin JD, Robertson SA. Macrophages reg-ulate expression of alpha1,2-fucosyltransferase genes in human endome-trial epithelial cells. Mol Hum Reprod 2012;18:204–15.

9

SUPPLEMENTAL FIGURE 1

Cytokine secretion from cocultured cells and control ESCs. (A) Cytokine secretion was evaluated with the cytokine array kit. Granulocytemacrophage–CSF, RANTES, IL-1RA, and MCP-1 production showed greater up-regulation in CM from cocultured cells (cCM) compared withcultured iESC medium (eCM). (B) Expressions of CCR2, GM-CSF, and GM-CSFR mRNAs were analyzed by PCR. Macrophages and iESCsevidenced GM-CSF and GM-CSFR mRNAs. (C) Immortalized ESCs were cultured with or without GM-CSF, and the percentage of total cells wasevaluated with WST assays. Mac ¼ macrophage. Data were analyzed using Student’s t test. *P< .05.Itoh. Stat3 and macrophages in endometriosis. Fertil Steril 2013.

ORIGINAL ARTICLE: ENDOMETRIOSIS

9.e1 VOL. - NO. - / - 2013