Immunohistochemical detection of type I, III, and IV collagen in endometriosis implants*t

Transcription

1 FERTILITY AND STERILITY Vol. 57, No.5, May 1992 Copyright 1992 The American Fertility Society Printed on acid-free paper in U.S.A. Immunohistochemical detection of type I, III, and IV collagen in endometriosis implants*t Dale W. Stovall, M.D.* Joyce A. Anners, M.A. Jouko Halme, M.D., Ph.D. Department of Obstetrics and Gynecology, Division of Reproductive Endocrinology and Fertility, University of North Carolina School of Medicine, Chapel Hill, North Carolina Objective: To determine if types I, III, or IV collagen are present in ectopic endometrium and to determine which type(s) of collagen are present in the connective tissue surrounding ectopic endometrial implants. Design: Paired intrauterine and ectopic endometrial samples were obtained for study at the time of laparoscopy from women in the proliferative and secretory phase of the menstrual cycle. Connective tissue surrounding each ectopic implant was also obtained for study. Setting: Academic research environment with institutional review board approval. Patients: Six patients without endometriosis were used as controls. Ten additional patients with stage II or III endometriosis were studied. Only women on no medications participated in the study. Main Outcome Measures: Immunohistochemical techniques were used to identify the presence of collagen in biopsied specimens. Results: Each collagen type studied was identified in the intrauterine endometrium of patients with and without endometriosis. All collagen types were also identified in each of the ectopic endometrial implants studied. The distribution of collagen in ectopic endometrial implants was similar to the distribution of collagen seen in intrauterine endometrium obtained from patients with or without endometriosis. Collagenous tissue that contained type I collagen was identified at the periphery of deep ectopic implants. Conclusions: This study demonstrates the presence of type I, III, and IV collagen in the intrauterine and ectopic endometrium of patients with endometriosis. Type I collagen was the predominant collagen present in the surrounding collagenous tissue associated with deep, ectopic endometrial implants. Fertil Steril 1992;57:984-9 Key Words: Endometriosis, endometrium, collagen, immunohistochemistry, connective tissue Collagen is an important extracellular matrix protein. At least 13 different types of collagen exist. Received November 19, 1991; revised and accepted January 30,1992. * Supported by grant R01 HD21546 from the National Institutes of Health, Bethesda, Maryland. t Presented at the 38th Annual Meeting of the Society for Gynecologic Investigation, San Antonio, Texas, March 20 to 23, :t: Reprint requests and present address: Dale W. Stovall, M.D., Department of Obstetrics and Gynecology, Division of Reproductive Endocrinology and Infertility, University of Pittsburgh Magee Women's Hospital, Forbes Avenue and Halket Street, Pittsburgh, Pennsylvania Stovall et al. Collagen in endometriosis Each type's structure is tailored to the function it performs. Type I and III collagen are often found in combination. They commonly function as structural proteins in tissue stroma. Type IV collagen is found almost exclusively in basement membranes. It has been identified in the basement membranes of blood vessels and glands. The extracellular matrix components of the intrauterine endometrium have been studied (1). Several types of collagen have been identified in both the proliferative and secretory endometrium including types I, III, and IV. However, only the presence of a basement membrane, which contains type IV Fertility and Sterility

2 collagen, has been documented in ectopic endometrium (2). In addition to endometrial glands and stroma, endometriosis implants contain other structural and functional components that distinguish them from intrauterine endometrium. Ectopic endometrial implants are often encompassed by a chronic inflammatory process unlike the inflammation that occurs in intrauterine endometrium only at the time of menses. Ectopic endometrium is further characterized by an excessive production of connective tissue. It has been suggested that this abundant connective tissue may contribute to the impaired fertility associated with this disease. One possible mechanism is the interference between the interaction of the ovum and fallopian tube, including the impedance of ovum pick-up or release. Also, the overabundant connective tissue matrix could influence the production of cytokines and/or vice versa. Identification of the specific collagens contained within endometriotic implants is fundamental to the elucidation of these interactions. Using immunohistochemical techniques, we sought to answer two general questions. First, are the major collagenous components of ectopic endometrium similar to that of eutopic endometrium? Second, of the collagen subtypes present in ectopic endometrium, which are most abundant in the surrounding connective tissue? With this information, more appropriate in vitro conditions can be developed to study the interactions between specific collagen subtypes and the cellular components of ectopic endometrium. MATERIALS AND METHODS Patients and Tissue Sampling Consents for study participation were obtained from subjects undergoing laparoscopy for known or suspected endometriosis. A group of subjects undergoing laparoscopic tubal ligation were recruited to serve as controls. Only women of reproductive age without surgical or hormonal therapy for the preceding 6 months were eligible for entry into the study. Patients with previous major surgical procedures for treatment of endometriosis or removal of pelvic organs were ineligible. Women in either the proliferative or secretory phase of the menstrual cycle were studied. The timing of each biopsy was based on day of last menses and confirmed with histologic dating. Paired tissue samples were acquired from each patient. An aspiration biopsy cannula (Pipelle or Karman cannula) was used to obtain intrauterine endometrial tissue. Biopsies of ectopic endometrium were obtained with laparoscopic instruments using a punch-type biopsy. Only endometriosis specimens containing glands, stroma, and surrounding connective tissue were used. All biopsies were taken from peritoneal surfaces, usually from the posterior cul-de-sac or from the uterosacral ligaments. The endometriosis lesions were all of the pigmented, puckered type and were either deep or superficial lesions. Primary Antibodies Antibodies were obtained for types I, III, and IV collagen. Type I and IV antibodies were the immunoglobulin (Ig)G sub fractions from polyclonal, rabbit antihuman antibodies (Biodesign International, Kennebunkport, ME). Type III antibody was an IgG monoclonal, mouse antihuman antibody (Heyltex, Houston, TX). Dot blotting was performed to determine appropriate antibody dilutions and to assess antibody specificity. In brief, 2.5 JlL of purified types I, III, or IV antigen (Sera-Lab LTD., London, United Kingdom) from a O.l-Jlg/mL stock solution were allowed to dry on nitrocellulose paper. Nonfat dry milk was used to block nonspecific receptors. Serial dilutions of primary antibody were then placed not only on the corresponding antigen but on both of the nonspecific antigens. Preabsorption of Primary Antibody Initial dot blots and tissue staining revealed crossreactivity between types I and III antibody and their respective antigens. Therefore, a preabsorption procedure was performed to purify these two antibodies before further tissue staining. In brief, antibody was prepared at two times the concentration initially used for staining. After addition of the opposing antigen the solution was incubated at 37 C for 1 hour and then allowed to precipitate overnight at 4 C. The solution was centrifuged at 3,000 X g for 5 minutes. Dot blots performed with the supernatant proved that virtually all cross-reactivity had been removed. All of the tissue staining was performed with preabsorbed antibody except for type IV antibody, which was without cross-reactivity. Immunohistochemistry Biopsy specimens were snap frozen in liquid nitrogen. Microscope slides were pretreated by dipping into a subbing solution. The subbing solution con- Vol. 57, No.5, May 1992 Stovall et al. Collagen in endometriosis 985

3 sisted of 1 g of 275 bloom granular gelatin (Fisher Scientific, Fair Lawn, NJ) dissolved in 1 liter deionized H 20 supplemented with 0.1 g chromium KS0 4 Cryostat tissue sections were cut at 8 Ilm and placed on the pretreated slides. The tissue was fixed with acetone, and endogenous peroxidase was blocked with 3% hydrogen peroxide in methanol. The slides were washed with TRIS-buffered solution between each step. Receptor blockage was accomplished by incubation with 200 ill of a 1:10 dilution of serum from the species of the secondary antibody (either horse or goat) for 30 minutes in a humidity chamber at 37 C. The primary antibodies were diluted in TRIS-buffered solution and supplemented with 2% serum from the species of the appropriate secondary antibody. The tissue sections were incubated with primary antibody overnight at 4 C in a humidity chamber. The Vectastain ABC kit (Vector Laboratories, Inc., Burlingame, CA) was used to detect the presence of primary antibody with diaminobenzidine as the enzyme substrate. The tissue was counterstained with Gills' hematoxylin no. 2 (Sigma, St. Louis, MO), dehydrated, and mounted with Permount (Fisher Scientific Co.). An Olympus BH-2 microscope (Olympus Optical Co., Lake Success, NY) was used to evaluate tissue sections. The Vectastain kit used to detect the primary antibodies stains brown in color and the counterstain, Gills' hematoxylin, stains blue. RESULTS Initial staining procedures were performed on intrauterine tissue specimens obtained after consent from six normally cycling women undergoing laparoscopic tubal ligation. As confirmed by endometrial dating, three of these women were in the secretory phase of the menstrual cycle and three were in the proliferative phase. All six women had normal pelvic anatomy as confirmed by laparoscopy and served as a control group for the intrauterine tissue findings obtained in the patients with endometriosis. Paired tissue samples were obtained from 10 reproductive age women. Of these 10 women, 5 were in the proliferative phase of their cycle and 5 were in the secretory phase. Six patients had a history of infertility and pelvic pain. Four patients underwent laparoscopy for chronic pelvic pain alone and were either fertile or had not had their fertility tested. Both endometrial glands and stroma were identified in all 10 eutopic and ectopic tissue samples. Of the 10 ectopic samples, abundant, surrounding collagen was found in only eight. The two biopsy specimens without excess collagenous tissue were obtained from superficial lesions in infertile women with moderate dysmenorrhea and stage II disease (3). The 8 ectopic samples with abundant, surrounding collagenous tissue were all from deeper lesions in patients with stage II or III disease. Staining of types I, III, and IV collagen in proliferative, intrauterine endometrium is shown in Figure la, B, and C, respectively. These sections are representative of the staining patterns seen in all patients. Samples from secretory endometrium demonstrated similar staining patterns. Tissue fragmentation, as seen in these sections, was common. All intrauterine specimens, including those from the control group and from the paired samples, contained all three types of collagen with similar distributions. Type I collagen was found both in close proximity to glandular structures and within the stroma. The sporadic staining pattern seen in Figure 1A occurred more frequently in fragmented sections and may be an artifact. Type III collagen stained in a characteristic fibrillar pattern and was found only within the stroma. Type IV collagen was present only within the basement membrane of glands and arterioles in both the proliferative and secretory tissue samples. Figure 2A, B, and C illustrate representative staining of types I, III, and IV collagen from ectopic implants. Both glandular structures and endometrial stroma can be identified in each section. Similar staining patterns and distribution of collagen, as established in the intrauterine tissue, are evident. As seen in Figure 2A, type I collagen was present in a more diffuse pattern in ectopic endometrial tissue as compared with eutopic tissue. Because ofthe vascularity of most implants, type IV collagen was identified throughout most specimens. Each of the three collagen types were demonstrated in all 10 paired samples. Staining for types I, III, and IV collagen was performed on the surrounding collagenous tissue identified in eight of the ectopic implants. Only type I collagen was identified in each of these implants. If the surrounding tissue was vascularized, type IV collagen was present (within the basement membrane of arterioles). Type III collagen was not identified in the surrounding area of any of these tissue samples. DISCUSSION We were able to confirm the presence of each collagen type in the ectopic endometrium of patients 986 Stovall et al. Collagen in endometriosis Fertility and Sterility

4 Figure 1 Proliferative intrauterine endometrium stained with anticollagen antibody. The brown stain represents the presence of anticollagen antibody. Cell nuclei are stained with hematoxylin. (A), Antitype I collagen antibody; (B), antitype III collagen antibody; (C), anti-type IV collagen antibody. Each tissue section is magnified at X400. Figure 2 Ectopic endometrial tissue obtained during the proliferative phase and stained with anticollagen antibody. The brown stain represents the presence of anticollagen antibody. Cell nuclei are stained with hematoxylin. (A), Anti-type I collagen antibody; (B), antitype III collagen antibody; (C), antitype IV collagen antibody. The magnification is X400 for A and C and X600 for B. Vol. 57, No.5, May 1992 Stovall et al. Collagen in endometriosis 987

5 in both the proliferative and secretory phases of the menstrual cycle. Because collagen is a stable molecule, one would not expect to see a major change in collagen composition throughout the menstrual cycle. Aplin et al. (1) found that in intrauterine tissue, type III collagen was present throughout the menstrual cycle and in early pregnancy. These findings are consistent with our data. Aplin et al. (1) also noted a change in the consistency of type III collagen from tightly packed fibrils in the proliferative phase to a system of matrix channels within decidual cells. In this study, there was a noticeable change in the pattern of type I collagen from intrauterine to ectopic tissue. Type I collagen was present in a diffuse, fibrillar pattern in ectopic endometrium and was more tightly packed in eutopic endometrial samples. As previously discussed, the differences in the distribution of staining seen in these tissue sections may be secondary to tissue fragmentation. When connective tissue immediately surrounding an endometriosis implant was examined, only type I collagen was identified (unless the tissue included arterioles). The presence of surrounding connective tissue was more common in deeply invading lesions as compared with superficial lesions. Although our sample size is too small to determine the prevalence of this occurrence, it suggests that deeply invasive lesions may be more active in their production of collagen. It seems likely that the excess type I collagen at the periphery of the implant represents a reaction of the surrounding tissue to limit the growth or influence of the implant. Types I and III collagen are commonly found in combination. Tissue from the skin, gingiva, and, as demonstrated in this study, endometrium contain both type I and III collagen. Healing skin wounds also contain both type I and III collagen (4). One might expect to find both of these collagen types in the collagenous tissue surrounding ectopic endometrium. Surprisingly, this tissue was composed of type I collagen alone. This may be an effect of cytokines, which are associated with ectopic endometrium. Moriyama (5) has demonstrated the stimulatory effects of growth factors on scar fibroblasts in vitro. When transforming growth factor beta was added to fibroblasts in culture, an increase in both collagen and fibronectin resulted (6). Similar paracrine effects might modulate the production of collagen in endometriosis implants. The evaluation of adhesions secondary to disease processes other than endometriosis would answer this question. If both type I and III collagen were present in these adhesions, this would suggest that cytokines do play a role in the collagen production of endometriosis implants. Normal intrauterine endometrium is not characterized by an abundant production of collagenous tissue, yet ectopic endometrium can be. Implantation of endometrial tissue into peritoneal tissue may in itself stimulate this transformation. Recombination experiments by Cunha et al. (7) have shown that the coculturing of vaginal stroma and endometrial epithelium caused the endometrial epithelium to keratinize and secrete mucin not unlike vaginal epithelium. Because these are both hormonally responsive tissues, one might suggest that this analogy is inappropriate. Yet, there is some evidence to suggest that the peritoneum and/or the subperitoneal mesenchymal cells are also responsive to hormonal stimulation. Coelomic transformation has long been heralded as a potential cause of endometriosis. The presence of endometriosis in the bladder of men receiving estrogen (E) supports this hypothesis (8). A rare condition, known as leiomyomatosis peritonealis disseminata in which subperitoneal smooth muscle nodules form usually in response to E, supports the hypothesis that either the peritoneum or subperitoneal mesenchymal cells are responsive to hormonal stimulus (9). Pelvic pain is more common in patients with deeply invasive endometriosis lesions (10). The excessive production of collagen by these lesions could in theory contribute to their related symptomatology. However, because medical treatment of endometriosis is known to decrease implant size and pelvic pain but to have little effect on adhesions or fibrosis (11), this relationship may have little clinical significance with regard to pain. Previous authors have suggested that some endometriosis implants may not contain a basement membrane (12). This information might be useful in determining which lesions are more aggressive and should be treated accordingly. Evers and Willebrand (2) previously have documented the presence of basement membrane in ectopic endometrial implants. He was unable to demonstrate any lesions, invasive or otherwise, that did not contain basement membrane. Our findings are consistent with that of Evers. We found type IV collagen (present in basement membrane) in both deep and superficial endometriotic lesions. In any immunohistochemical investigation, it is important to document the sensitivity and specificity of the antibodies used in antigen detection. Staining patterns in our initial immunohistochemical studies (before using antibody preabsorption) suggested that 988 Stovall et al. Collagen in endometriosis Fertility and Sterility

6 significant cross-reactivity existed between type I and type III collagen and their respective antibodies. A dot-blot procedure was used to confirm crossreactivity between these two collagen antibodies and the need for preabsorption. Dot-blot analysis confirmed that preabsorption of the nonspecific antibody in conjunction with appropriate antibody dilution had resolved this problem. The specificity of the antibodies was not tested against all other known collagen types. Not only would this be an cost-limiting proposition, but because of the specialized nature of most collagen types, their presence would be very unlikely. Staining patterns were used to confirm the positive identification of each collagen type in extrauterine tissue. The verification of each collagen was first established in intrauterine tissue. To be certain that patients with endometriosis had similar collagenous tissue to patients without the disease, a control group was studied. The actual pattern of staining, for example, a fibrillar pattern as opposed to a solid pattern, may vary from one antibody to another especially when polyclonal antibodies are used. However, the distribution of collagen, such as that found in the stroma or in the basement membrane, should remain relatively consistent. This consistency in staining distribution was used as an indication of staining specificity. When types I, III, and IV collagen were studied in human gingival tissue, type I collagen was principally organized in large bundles, type III collagen was found in a fibrillar pattern, and type IV was found only in basement membranes (13). These findings are consistent with our data. The staining distributions seen in the extrauterine tissue were virtually duplicated when we studied intrauterine tissue. Numerous mediators of cell growth and differentiation may be responsible for the increased production of collagenous proteins in ectopic endometrial implants. Nevertheless, we have demonstrated that the collagenous extracellular matrix proteins contained within intrauterine endometrial tissue of normal controls and patients with endometriosis are qualitatively similar to those of ectopic endometrial tissue and that type I collagen is the predominant collagen type contained within the abundant surrounding connective tissue associated with deeper implants. Further study is necessary to determine the relationship between paracrine factors and extracellular matrix production in ectopic endometrium and to determine the effects of collagen production on infertility and pelvic pain. REFERENCES 1. Aplin JD, Charlton AK, Ayad S. An immunohistochemical study of human endometrial extracellular matrix during the menstrual cycle and the first trimester of pregnancy. Cell Tissue Res 1988;253: Evers JLH, Willebrand D. The basement membrane in endometriosis. Fertil Steril1987;47: The American Fertility Society. Revised American Fertility Society classification of endometriosis: Fertil Steril 1985;43: Haukipuro K. Synthesis of collagen types I and III in reincised wounds in humans. Br J Surg 1991;78: Moriyama K, Shimokawa H, Sucumi T, Sucumi S, Kuruda T. Effects of growth factors on mucosal scar fibroblasts in culture-a possible role of growth factors in scar formation. Matrix 1991;11: Ignotz RA, Massague J. Transforming growth factor-{3 stimulates the expression of fibronectin and collagen and their incorporation into the extracellular matrix. J Bioi Chern 1986;261: Cunha GR, Bigsby RM, Cooke PS, Sugimura Y. Stromalepithelial interactions in adult organs. Cell Differ 1985;17: Oliker AJ, Harris AE. Endometriosis of the bladder in a male patient. J Urol 1971;106: Fujii S, Okamura H, Nakashima N, Bann C, Aso T, Nishimura T. Leiomyomatosis peritonealis disseminata. Obstet Gynecol 1980;55:79S Koninckx PR, Meuleman C, Demeyere S, Lesaffre E, Comillie FJ. Suggestive evidence that pelvic endometriosis is a progressive disease, whereas deeply infiltrating endometriosis is associated with pelvic pain. Fertil Steril 1991;55: Kennedy SH, Williams la, Brodribb J, Barlow DH, Shaw RH. A comparison of nafarelin acetate and danazol in the treatment of endometriosis. Fertil Steril1990;53: Mettler L, Semm K. Three-step therapy of genital endometriosis in cases of human infertility with Iynestrenol, danazol or gestrinone administration in the second step. In: Raynaud JP, Ojasoo T, Martini L, editors. Medical management of endometriosis. New York: Raven Press, 1984: Chavrier C, Couble ML, Magloire H, Grimaud JA. Connective tissue organization of healthy human gingiva. J Periodont Res 1984;19: Vol. 57, No.5, May 1992 Stovall et al. Collagen in endometriosis 989