Detection of a unique 32-kd protein in the peritoneal fluid of women with endometriosis*t

Size: px

Start display at page:

Download "Detection of a unique 32-kd protein in the peritoneal fluid of women with endometriosis*t"

Noel Curtis
5 years ago
Views:

1 FERTILITY AND STERILITY Vol. 61, No, 2, February The American Fertility Society Printed on acid-free paper in U. S. A. Detection of a unique 32-kd protein in the peritoneal fluid of women with endometriosis*t Warren B. Nothnick, Ph,D.:j: Thomas E. Curry, Jr., Ph.D.11 Ken N. Muse, M.D. II Steve N. London, M.D. II Michael W. Vernon, Ph.D.:j:~ University of Kentucky Medical Center, and Women's Speciality Center, Central Baptist Hospital, Lexington, Kentucky Objective: To test the hypothesis that endometriotic tissue secretes endometriotic-specific proteins into the peritoneal fluid (PF) of women with endometriosis. Design: A prospective design was utilized in this study. Setting: Tertiary care, university-based center and reproductive endocrinology laboratory. Participants: Women of reproductive age who were infertile with endometriosis (n = 19), as well as without endometriosis (n = 7), and fertile women undergoing tubal ligation (n = 6). Interventions: Collection of PF fluid via laparoscopy. Main Outcome Measures: Peritoneal fluid proteins were isolated and assessed by two-dimensional polyacrylamide gel electrophoresis. Results: Two-dimensional electrophoresis of PF proteins isolated a group of proteins (M r = 32 to 40 kd, pi = 4.5 to 5.2) in all PF samples that was similar to the rat endometriotic implant-specific protein, Endo-l. This group of proteins consisted of 5 to 12 individual proteins with endometriosis PF containing a significantly higher number of proteins (median = 11) compared with either PF from infertile women without endometriosis (median = 8) or from women undergoing tubal ligation (median = 7). In addition, one protein (M r = 32 kd,pi = 5.8), termed EPF -32, was detected predominantly (18 of 19 samples analyzed) in PF from women with endometriosis. This protein was also detected in PF from infertile women without endometriosis (2 of 7 samples) but not in the PF of fertile women undergoing tubal ligation (0 of 6 samples). The appearance of this protein was not associated with the severity of endometriosis. Conclusion: It is concluded from this study that PF from women with endometriosis predominantly contains a 32-kdprotein (EPF-32) compared with the PF of women without the disease. The role of EPF -32 in the pathophysiology of endometriosis is not established but this protein may function as a diagnostic marker for endometriosis. Fertil Steril 1994;61: Key Words: Endometriosis, endometriotic implant, Endo-l, peritoneal fluid, protein analysis Received July 9, 1993; revised and accepted October 21, * Supported in part by grant HD21962 from the National Institutes of Health, Bethesda, Maryland. t Presented in part at the 48th Annual Meeting of The American Fertility Society, New Orleans, Louisiana, October 30 to November 5, :j: Department of Physiology, University of Kentucky Medical Center. Reprint requests: Department of Obstetrics and Gynecology, Center for Women's Health, Room C366, University of Kentucky, Lexington, Kentucky Endometriosis is a gynecologic disease characterized by pelvic pain, dysmenorrhea, and infertility (1, 2). Currently, endometriosis can only be detected by invasive procedures such as laparoscopy, with definitive diagnosis requiring visual or biopsy confirmation. Even then, microscopic or occult endometriosis may be misdiagnosed due to the inability to visualize the disease (3). In addition, invasive procedures have been associated with the accidental development of endometriosis (4, 5). Therefore, II Department of Obstetrics and Gynecology, University of Kentucky Medical Center. 11 Women's Speciality Center. 288 Nothnick et al. Endometriosis-associated protein Fertility and Sterility

the benefits of a noninvasive, biochemical diagnostic marker for the detection of endometriosis are evident. However, to date there is no such biochemical diagnostic marker.

2 the benefits of a noninvasive, biochemical diagnostic marker for the detection of endometriosis are evident. However, to date there is no such biochemical diagnostic marker. Studies attempting to detect a peritoneal fluid or plasma constituent unique to endometriosis patients have been inconclusive. The tumor-associated antigen, CA-125, has been shown to be elevated in the plasma and peritoneal fluid (PF) of women with endometriosis (6, 7), but this antigen is also used as a marker for ovarian cancer (8). Like CA-125, placental protein 14 (PP-14) has also been shown to be elevated in endometriosis patients (9). However, both the levels ofca-125 (10), and PP-14 fluctuate with the menstrual cycle (11). Taken together, it is clear that neither CA-125, nor PP-14 can be associated solely with the presence of endometriosis. Recently, two unique groups of proteins (known as Endo-l and Endo-2) have been shown to be synthesized and released in vitro by rat endometriotic implants (12). Specifically, Endo-l is a group of 8 to 12 glycosylated proteins that span the molecular weight range of 40 to 50 kd with an isoelectric point of 4.0 to 5.2. These proteins differ in molecular weight and net charge because of their extent of glycosylation and are thus referred to as isovariants. Endo-2 is a group of 9 to 12 nonglycosylated proteins that have an apparent isoelectric point of 7.5 to 9.0 and span the molecular weight range of 28 to 32 kd. Both groups of proteins appear to be produced only by the rat implant and not the uterus, and their production does not seem to be modulated by the reproductive cycle (12). In addition, a group of proteins displaying a similar molecular weight and isoelectric point to that of rat Endo-2 has been detected in culture media of human endometriotic tissue (Sharpe KL, Zimmer RL, Penney LL, Griffin WT, abstract), suggesting that human implants may produce similar proteins to those produced by rat implant tissue. If so, proteins such as Endo-l and -2 may serve as potential diagnostic markers for endometriosis because these proteins theoretically should be present only in women with the disease. Detection of such "disease-specific" proteins in vivo would be the first step toward the development of these proteins as diagnostic markers for endometriosis. Therefore, we wished to test the hypothesis that endometriotic tissue secretes endometriotic-specific proteins into the PF of women with endometriosis to verify the in vivo presence of implant-specific proteins in humans. To do so, we analyzed the protein composition of PF from both fertile and infertile women without endometriosis, as well as infertile women with endometriosis by two-dimensional gel electrophoresis coupled with ultrasensitive silver staining of the proteins. Patients MATERIALS AND METHODS Peritoneal fluid was obtained from 32 women of reproductive age seeking tubal ligation (n = 6; all secretory phase of the menstrual cycle) or undergoing a diagnostic laparoscopy for infertility (n = 26). This latter group of infertile women was further divided into two groups on the basis of the presence of endometriosis (n = 19, of which 15 were in their secretory phase and 4 were in their proliferative phase) or absence of the disease (n = 7, of which 6 were in their secretory phase and 1 was in her proliferative phase). None of the women included in this study had received any hormonal treatment within 3 months before this study. The day of the menstrual cycle was determined by the first day of their last menstrual period. Endometriosis was confirmed by pathological examination of biopsied tissue and the extent of the disease was staged according to the revised American Fertility Society (AFS) classification (13). Of the 19 women found to have endometriosis, 7 were classified as stage I, 4 as stage II, 4 as stage III, and 4 as stage IV. Processing of Peritoneal Fluid Peritoneal fluid was collected via laparoscopy. To minimize blood contamination, PF samples were collected immediately after insertion of the laparoscope. Peritoneal fluid was then placed on ice and transported to the laboratory. After measuring PF volume, the hematocrit was determined by loading an aliquot (approximately 50 ill) of PF into a hematocrit tube and determining the percent of red blood cells present. The remaining PF was then centrifuged at 1875 X g for 15 minutes at 4 C to remove cellular debris. After centrifugation, the PF supernatant was dialyzed (molecular weight cut-off = 6,000 to 8,000 MW) against 1.0 mm tris(hydroxymethyl)aminomethane HCI, ph 8.2, for 48 hours with three changes of 2 L each to remove electrolytes and other compounds that interfere with isoelectric focusing. Protein concentration was then determined for each sample by the protein -dye binding procedure as described by Bradford (14). Vol. 61, No.2, February 1994 Nothnick et al. Endometriosis-associated protein 289

Two-Dimensional Gel Electrophoresis Two-dimensional polyacrylamide gel electrophoresis (PAGE) was performed as described by Roberts and colleagues (15). Briefly, 500 J.

3 Two-Dimensional Gel Electrophoresis Two-dimensional polyacrylamide gel electrophoresis (PAGE) was performed as described by Roberts and colleagues (15). Briefly, 500 J.lg of nondialyzable PF protein were resuspended in 25 J.lL of 5 mm K 2 C0 3 containing 9.4 M urea, 2% (v/v) Nonidet P-40 and 0.5% (w/v) dithiothreitol (Sigma Chemical Co., St. Louis, MO) and applied to each first dimension isoelectric focusing (IEF) gel. A mixture of ampholines (Pharmacia LKB Biotechnology, Inc., Piscataway, NJ) was utilized to create a ph gradient of 3.0 to 9.0 for the IEF. Molecular weight markers (14.4 to 94.0 kd; Pharamacia LKB) were added to the second dimension slab gels as molecular weight references. Silver Staining of Gels The polyacrylamide (12 %) second dimension gels were fixed overnight in a fixative solution consisting of 5% (vol/vol) glacial acetic acid, 50% (vol/vol) absolute ethanol, and 45% water. The next morning gels were silver stained using the ultrasensitive method as described by Sammons and coworkers (16). After staining, gels were photographed and protein patterns were visually evaluated. Statistical Analysis Protein concentration data were first evaluated for homogeneity of variance using Bartlett's test. No significant difference was detected, which allowed data to then be analyzed by one-way analysis of variance (17). The number of proteins in the Endo-1-like group for each of the test groups were analyzed by the Kruskal-Wallis test for nonparametric data. Significant differences among the three test groups were then determined using the Dunn procedure (17). Specificity and sensitivity of EPF -32 for the detection of endometriosis was calculated using standard formulations (17). Briefly, specificity was calculated using the following formula: specificity = 1 - (the number of samples containing EPF-32 from nonendometriosis patients/the total number of samples from nonendometriosis patients) X 100%. Sensitivity was calculated using the following formula: sensitivity = (the number of samples containing EPF -32 from women with endometriosis/the total number of samples from women with endometriosis) X 100%. RESULTS No differences in PF protein concentration or protein separation patterns existed because of the Protein Concentration (mg/ml) 120r-~~~~~~~~~~~~~~~~~ o~~~~~~~--~~~~~~~~ Tubal ligations Without endometriosis With endometriosis Figure 1 Total protein concentration (mg/ml) in PF from tubal ligation and infertile women with and without endometriosis. The median values (- ) of total PF protein concentration for the tubal ligation and infertile women with and without endometriosis (28.5, 57.0, and 48.0 mg/ml, respectively) groups were not statistically different. stage of the menstrual cycle, therefore all data from patients in their secretory and proliferative stages were pooled in each study group. Likewise, no differences were detected because of the severity of endometriosis, so all data grouped according to stage of endometriosis were pooled within this study group. In addition, hematocrits for all PF samples were <1.0%, indicating minimal blood contamination of the PF samples at the time oflaparoscopy. Analysis of total protein concentration revealed no significant difference (P > 0.10) among the three test groups (Fig. 1). Two-dimensional analysis of PF proteins revealed that several hundred similar proteins were present in the PF of all three study groups. However, qualitative analysis revealed that all PF contained a group of proteins with a similar pi (4.5 to 5.2) and two-dimensional PAGE separation pattern to Endo-1 produced by rat ectopic endometrium (Fig. 2A and B). The molecular weight for this group of human proteins was 32 to 40 kd compared with 40 to 50 kd for rat Endo-1. The overall number of proteins contained in the Endo-1-like group of proteins ranged from 5 to 12 (see Fig. 2C for schematic compilation of the two-dimensional PAGE separation pattern of all 12 proteins) with endometriosis PF containing a significantly higher number of proteins (median = 11; range = 9 to 12) compared with PF from either infertile women 290 Nothnick et al. Endometriosis-associated protein Fertility and Sterility

4 Number of Proteins 14r---~~----~----~ r OL L--~------~ ~ Tubal ligation Without endometriosis With endometriosis Figure 3 Number of proteins in the Endo-1-like group detected in the PF of each of the three study groups. Groups are defined in Figure l. The median number of proteins for tubal ligation and infertile women with and without endometriosis were 7, 11, and 8, respectively. c Mr -3 (X 10 ) pi 4.0 > 5.8 Figure 2 Two-dimensional gel containing 500 Ilg of PF protein from an infertile patient (A) found to have endometriosis (stage II, revised AFS classification) and an infertile patient found to be disease-free (B)_ (A), The arrow indicates the presence of EPF -32. (B), The angled bracket indicates the core proteins common to all 32 PF samples, and the arrow indicates the absence of EPF-32 in this sample. (e), A schematic composite (from all 32 samples) of all proteins in the Endo-1-like group (numbered 1 to 12 with protein number 13 being EPF-32). without endometriosis (median = 8; range = 5 to 11, P = ) or from fertile women undergoing tubal ligation (median = 7; range = 5 to 8, P = 0.001) PF (Fig. 3). No significant difference (P = 0.197) was detected in the number of proteins present in this Endo-l-like of proteins between PF from either in- 12 fertile women without endometriosis (median = 8) or from fertile women undergoing tubal ligation (median = 7; Fig. 3). Of the 12 proteins in the Endo I-like group, 4 "core" proteins (the group indicated by the bracket in Figure 2B and numbered 1 and 3 to 5 in Fig. 2C) were detected in all PF samples regardless of study group_ The remaining proteins were detected in no specific fashion. However, one protein was detected predominantly in PF from women with endometriosis (18 of the 19 endometriosis PF samples analyzed) (Table 1, Fig. 2A and C)_ This protein had an apparent pi of 5_8 and Mr of 32 kd (now termed EPF -32) and was detected in only 2 of the 7 PF samples analyzed from infertile, nonendometriosis patients, but was absent in PF samples from fertile women undergoing sterilization (Table 1, Fig_ 2B and C). The specificity and sensitivity of EPF -32 for endometriosis was 84.6% and 94.7%, respectively. We were unable to detect any proteins showing Table 1 Occurrence of EPF -32 in the Peritoneal Fluid of the Study Groups Study group Tubal ligation Infertile, nonendometriosis Endometriosis Number of positive samples per group 0/6 2/7 18/19 Percentage positive Vol. 61, No.2, February 1994 Nothnick et al Endometriosis-associated protein 291

5 characteristics (pi = 7.5 to 9.0; Mr = 28 to 32 kd) similar to that of the previously reported rat (12) and human (Sharpe KL, Zimmer RL, Penney LL, Griffin WT, abstract) forms of the endometriotic implant-specific group of proteins Endo-2. DISCUSSION The present study demonstrates that a 32-kd protein, EPF -32, (pi = 5.8) was found predominantly in the PF of women with endometriosis. The presence of this protein did not appear to be correlated with the severity of the disease, suggesting that some factor or factors common to all stages of the disease may influence the production or be the source of EPF -32. The fact that EPF -32 was detected in a portion (2/7) of PF from infertile women without the disease may suggest that these infertile women possessed a common factor with those women having endometriosis that influenced the production of EPF -32. Alternatively, these two infertile women may have actually had endometriosis. It is quite plausible that microscopic or occult endometriosis was present in these women but was not detected at the time of laparoscopy (3). To reiterate, EPF -32 was found in all but one PF sample from women with endometriosis. This finding may be explained by the absence of certain factor(s) that were present in the PF of the other 18 patients with endometriosis, as well as in the PF of the 2 infertile patients containing EPF -32, that influenced the presence of this protein. These factors that may influence the production of EPF -32 remain to be determined. The finding of a group of proteins in all study groups that had a similar isoelectric point and twodimensional PAGE separation pattern to the rat endometriotic implant-specific group of proteins Endo-l was unexpected, but is of considerable interest. Sharpe and Vernon (12) have reported that Endo-l is a group of glycosylated proteins (M r = 40 to 50 kd, pi = 4.0 to 5.2) consisting of 8 to 12 isovariants synthesized and released exclusively by rat endometriotic implants (12). In the present study we found a group of proteins consisting of 5 to 12 isovariants spanning the molecular weight range of 32 to 40 kd with a pi = 4.0 to 5.2, which we postulate may be the human homologue of rat Endo-l. The lower molecular weight of this group of human proteins may be explained by species difference between the human and rat. Also, it is plausible that the in vivo form of Endo-l differs in molecular weight due to processing of the proteins, which does not occur in vitro and thus allows the higher molecular weight. If this group of proteins in our study is in fact Endo-l, it would be of concern as to why these proteins appear in the PF of women without endometriosis. One possibility may be that a core group of proteins is synthesized by the endometriotic implant as well as other tissue sources, but only in endometriosis do we see the other isovariant forms of these core proteins expressed. In fact, Sammons and colleagues (16) have detected a group of approximately eight proteins in plasma that has a similar Mr and pi to that of the Endo-l-like proteins detected in PF from fertile women in the present study. More importantly, the additional 1 to 4 proteins (as well as EPF-32) detected in PF from women with endometriosis in the present study were not observed in the plasma samples in the study of Sammons and colleagues (16). These findings may suggest that these additional Endo-l-like proteins and EPF -32 are associated with endometriosis, but the core group of Endo-l-like proteins are common plasma proteins. Thus, the finding of Sammons and colleagues (16) may explain why in the present study we see some of the Endo-l-like proteins in PF from fertile women but see significantly more Endo-l-like proteins in PF of women with endometriosis. Although we were able to detect an Endo-l-like group of proteins. in PF in this study, we were unable to detect an Endo-2-like group of proteins in the same PF samples. Endo-2 is a nonglycosylated group of 9 to 12 proteins with an apparent Mr of 28 to 32 kd and a pi of 7.5 to 9.0 produced by both rat (12) and human endometriotic tissue (Sharpe KL, Zimmer RL, Penney LL, Griffin WT, abstract). Both Endo-l and -2 are secreted in similar quantities in vitro as compared with total implant protein secretion in the rat (12). Thus, the detection of one of these proteins and not the other was surprising. However, this may indicate that there are additional sources of Endo-l in vivo besides the endometriotic implant that lead to a greater quantity of these proteins in PF. This suggestion is again strengthened by the finding of Endo-l-like proteins in PF of nonendometriosis patients. The apparent absence ofendo-2 in human PF may also be related to the in vivo dynamics of this group of proteins. It is plausible that Endo-2 may be secreted and rapidly taken up by the lymphatics or peritoneum, thereby leaving minimal levels in the PF. Also, the production of Endo-2 may be suppressed in vivo 292 Nothnick et al. Endometriosis-associated protein Fertility and Sterility

6 and therefore only detected in an in vitro culture system. In summary, we report for the first time, the detection of a group of proteins similar to the rat endometriotic implant-specific protein, Endo-l in PF from women with and without endometriosis. Individual constituents that comprise this group of proteins are more abundant in PF from women with endometriosis. Specifically, a 32-kd protein, EPF- 32, was detected predominantly in patients with endometriosis, and the detection of this protein did not appear to be associated with the severity of the disease. Future identification of this protein may provide further insight into the understanding of the pathophysiology of endometriosis, as well as lead to the development of a noninvasive diagnostic marker for the disease. REFERENCES 1. Muse KN. Endometriosis and infertility. In: Wilson E, editor. Endometriosis. New York: Liss, 1987:91~1l0. 2. Burns WN, Schenken RS. Pathophysiology. In: Schenken RS, editor. Endometriosis: contemporary concepts in clinical management. Philadelphia: Lippincott, 1989:83~ Murphy AA, Green WR, Bobbie D, dela Cruz ZC, Rock JA. Unsuspected endometriosis documented by scanning electron microscopy in visually normal peritoneum. Fertil Steril 1986;46:522~4. 4. Stroup PE. Endometriosis in laparotomy scars. J Reprod Med 1976;16:85~6. 5. Rovito P, Gittleman M. Two cases of endometrioma in cesarean scars. Surgery 1986;100:1l8~ Barbieri RL, Niloff JM, Bast RC Jr, Schaetzl E, Kistner RW, Knapp RC. Elevated serum concentrations of CA-125 in patients with advanced endometriosis. Fertil Steril 1986;45:630~4. 7. Koninckx PR, Riittinen L, Seppala M, Co millie FJ. CA-125 and placental protein 14 concentrations in plasma and peritoneal fluid of women with deeply infiltrating pelvic endometriosis. Fertil Steril 1992;57:523~ Jacobs I, Bast RC Jr. The CA-125 tumour-associated antigen: a review of the literature. Hum Reprod 1989;4:1 ~ Telimaa S, Kauppila A, Ronnberg L, Suikkasi AM, Seppala M. Elevated serum levels of endometrial secretory protein PP14 in patients with advanced endometriosis. Suppression by treatment with danazol and high-dose medroxyprogesterone acetate. Am J Obstet GynecoI1989;161:866~ Jager W, Meier C, Wildt L, Sauerbrei W, Lang N. CA-125 serum concentrations during the menstrual cycle. Fertil Steril 1988;50:223~ Julkunen M, Apter D, Seppala M, Stenman D-H, Bohn H. Serum levels of placental protein 14 reflect ovulation in non-conceptional menstrual cycles. Fertil Steril 1986;45: 47~ Sharpe KL, Vernon MW. Polypeptides synthesized and released by rat endometriotic tissue differ from those of uterine endometrium in culture. Bioi Reprod 1993;48:1334~ The American Fertility Society. Revised American Fertility Society classification of endometriosis. Fertil Steril 1985; 43:351~ Bradford MM. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein -dye binding. Anal Biochem 1976;72:248~ Roberts RM, Baumbach GA, Buhi WC, Denny FB, Fitzgerald LA, Babelyn SF, Morst MN. Analysis of membrane polypeptides by two-dimensional polyacrylamide gel electrophoresis. In: Ventor JC, Harrison LC, editors. Molecular and chemical characterization of membrane receptors. New York: Liss, 1984:61~1l Sammons DW, Adams LD, Nishizawa EE. Ultra-sensitive silver-based color staining of polypeptides in polyacrylamide gels. Electrophoresis 1981;2:135~ Rosner B. Fundamentals of biostatistics. 3rd. ed. Boston: PWS-Dent,1990. Vol. 61, No.2, February 1994 Nothnick et al. Endometriosis-associated protein 293

Menstrual cyclicity of CA-125 in patients with endometriosist

FERTILITY AND STERILITY Copyright 1992 The American Fertility Society Printed on acid-free paper in U.S.A. Menstrual cyclicity of CA-125 in patients with endometriosis*t* Mark D. Hornstein, M.D. II Phaedra