Measurement of human epidermal growth factor receptor in the endometrium during the menstrual cycle Vladimir Troche, MD,' Dennis M. O'Connor, MD,' and R. Paul Schaudies, PhDb Washington, D.C., and Bethesda, Maryland To evaluate a potential physiologic role of the epidermal growth factor receptor in the endometrium, we measured the receptor content at different times in the menstrual cycle. Endometrial biopsy specimens were obtained from 28 normal women during the proliferative or secretory phase of the menstrual cycle, and the epidermal growth factor receptor content was determined. When the number of epidermal growth factor binding sites were evaluated as a function of time within each phase, a difference between phases became evident. The level of receptor increased during the proliferative phase with a maximum just before ovulation (p = 0.0128, r = 0.748). The epidermal growth factor receptor level decreased during the secretory phase, reaching a minimum before menses (p = 0.0001. r = 0.843). We conclude that the endometrial epidermal growth factor receptor content is cycle dependent. being maximal during the periovulatory period and minimal just before or during menses. These findings further suggest a physiologic role for epidermal growth factor in the proliferation and differentiation of the endometrium. (AM J OBSTET GYNECOL 1991;165:1499-503.)
Key words: Epidermal growth factor receptor, endometrium, menstrual cycle Epidermal growth factor (EGF), a 53-amino-acid single-chain polypeptide, is a potent cellular mitogen for a variety of cell types.' Studies of EGF administration in vivo or in vitro have been shown to result in accelerated proliferation and differentiation of a multitude of tissues of ectodermal origin. 2 In vitro EGF administration results in increased cellular protein, ribonucleic acid and deoxyribonucleic acid synthesis, and enhanced cell multiplication." EGF exerts its biologic activity through interaction with specific transmembrane receptors, which have been found to be overexpressed in various human malignant tumors, such as squamous cell carcinomas of the head, neck, lung, and skin,' and tumors of the thyroid: brain,' breast,6 bladder," and colon. s The presence of these abnormally elevated numbers of EGF receptors have been associated with more aggressive tumors and a poor prognosis. 6 . 7 This overexpl'ession of the EGF receptor sugests a disrup-
From the Department of Obstetrics and Gynecology" and the Department of Clinical Investigation,' Walter Reed Army Medical Center, and the Uniformed Services University of the H eallh Sciences.' Supported iJy the Department of Clinical Investigation. Waller Reed Army Medical Center, Washington. D. C. Presented in part at the Thirty-eighth Annual !vI eeting of the Society for Gynecologic Investigation, San Antonio, Texas. MaTch 20-23,
1991.
The opinions and assn"tions expressed herein are those of the authot"S and are not to be construed as official or as representing the views of the Department of the Army 01' the Department of Defense. Received for publication January 9, 1991," revised April 10, 1991; accepted April 24, 1991. Reprint requests: Vladimir Troche, MD, Department of Obstetrics and Gynecology. Walter Reed Army Medical Center, Washington, DC 20307-5001. 611 130683
tion of the normal regulatory mechanisms of cellular growth in which cell proliferation is controlled by regulation of the EGF receptor in the tissue. The EGF receptor was identified in the human myometrium, endometrium, and leiomyomas." These findings support a role for EGF and its receptor in the proliferation and differentiation of the endometrium during the normal menstrual cycle. Research in this area has produced conflicting results, with no agreement as to the expression of the EGF receptor in the endometrium as a function of the menstrual cycle. EGF receptor content has been reported as follows: (I) being not different between the proliferative and secretory phases of the menstrual cycle,lo. II (2) being higher in the proliferative than the secretory phase,I2 and (3) exhibiting cyclic variation during the menstrual cycle. J3 We performed endometrial biopsies in 28 women of reproductive age at documented time points in the proliferative or secretory phase of the menstrual cycle. The endometrial tissues were assayed for the presence of the EGF receptor to investigate its expression in endometrium and establish a correlation for EGF receptor expression during the proliferative and secretory phases of the menstrual cycle.
Material and methods Materials. Chloramine-T and Coomassie Brilliant Blue G were purchased from Sigma Chemical Company, St. Louis; minimum essential medium was obtained from Hazelton Biologics, Inc., Lenexa, Kan.; bovine serum albumin (fraction 5) was purchased from Gibco Laboratories, Grand Island, N.Y.; recombinant human EGF was supplied by Upstate Biotechnology,
1499
November 1991 Am J Obstet Gynecol
1500 Troche, O'Connor, and Schaudies
Inc., Lake Placid, N.Y., and carrier-free sodium-iodine 125 was obtained from E.I. Dupont de Nemours & Co., Wilmington, Del. Patients. The study population consisted of women of reproductive age undergoing diagnostic laparoscopy or laparoscopic tubal sterilization at our ambulatory surgery center. The human use committee of our institution approved the research protocol before patient recruitment started. The participants were recruited from the infertility and family planning services at Walter Reed Army Medical Center, Washington, D.C. They were taking no medications, and their menstrual histories were normal. Patients with amenorrhea, abnormal menstrual cycles, abnormal uterine bleeding, recent use of oral contraceptives (within the last 4 months), pelvic inflammatory disease, or pregnancy were excluded from this study. Before the scheduled surgery the research protocol was explained to all candidates, and informed consent was obtained from participants. On the day ofthe scheduled laparoscopic procedures, endometrial biopsies were performed with a Novak curette while the patients were under anesthesia. The biopsy specimens were taken in an uniform manner from the anterior, posterior, and lateral walls of the uterine fundus. A representative sample of each endometrial biopsy specimen was forwarded to our gynecologic pathologist for histologic dating, and the reminder of the biopsy specimens were immediately frozen in liquid nitrogen. The endometrial tissue was stored at - 70° C until assayed. All participants were contacted within 2 to 4 weeks of surgery to document the date of onset of the next menstrual cycle. This information, together with the histologic results, was used to date the endometrial tissue. Preparation of murine EGF and iodination of human EGF. Pure murine EGF was isolated by the method of Savage and Cohen. '4 Recombinant human EGF was radioiodinated by the chloramine-T technique with carrier-free sodium- '25 !.'5 The specific activity of the radiolabeled human EGF was adjusted to 1.5 x 10 6 counts/mining by the addition of nonlabeled human EGF. Preparation of endometrial membranes. Endometrial tissues were homogenized in five volumes of icecold 10 mmol/L Tris hydrochloride, ph 7.5, containing 25 mmol/L sucrose and 1 mmol/L ethylenediaminetetraacetic acid with 15 slow strokes in a 7 ml Dounce homogenizer. The homogenates were centrifuged in a Beckman JA-20 rotor at 800 g at 4° C for 10 minutes. Nuclear pellets were discarded, and the resultant clear supernatants were centrifuged in a Beckman 60Ti rotor at 120,000 g at 4° C for 60 minutes. Supernatants were discarded and the resultant endometrial membrane pellets were resuspended in minimal essential medium containing 0.2% bovine serum albumin. The
protein content of the resuspended membrane pellets was determined by the Bradford method,16 and the final endometrial membrane protein concentration was adjusted to 6 mg/ml. Radioreceptor assay. This assay is based on the observation that nonlabeled EGF competes in an equimolar basis with radiolabeled EGF for the plasma membrane receptor and the fact that both murine and human EGF elicit nearly identical biologic responses. J. 2 Binding assays were conducted in a total volume of 200 f.L1 in 1.5 ml polyallomer microcentrifuge tubes. In all assays, 25 f.L1 (150 f.Lg) of endometrial membrane suspension was added to each tube. The binding reaction was initiated by the addition of 125 1 human EGF at nine separate concentrations ranging from 0.125 to 35 ng/ml to the microcentrifuge tubes. The incubations were performed in duplicate at room temperature for 90 minutes with gentle shaking. After incubation, 1 ml of ice-cold minimal essential medium containing 0.2% bovine serum albumin was added to each tube, followed by vortex blending. The tubes were immediately centrifuged in a Surespin centrifuge at 9500 g for 5 minutes. The supernatants were aspirated and the bound radioactivity in the pellets was counted in an LKB 1274 automatic "V-counter with a counting efficiency of 73%. Specific binding was obtained from the differences between total and nonspecific binding that was determined in parallel in the presence of a 500-fold excess of nonlabeled murine EGF. Nonspecific binding accounted for 20% of the total binding. Scatchard 17 plot analysis of the specific binding was performed for each endometrial membrane preparation, the number of binding sites, and the dissociation constant were determined. Statistical analysis of the data was performed by (I) the two-sample t test for independent samples, when the binding data of the proliferative and secretory phases were compared with each other and (2) simple linear regression, for evaluation of the binding data within each phase of the menstrual cycle.
Results The participants in the study had a mean age of 30.2 years, with a range of 25 to 38 years. A total of 28 endometrial samples were obtained from these patients, 10 of which were in the proliferative phase and 18 in the secretory phase of the menstrual cycle. In Table I the participants in the study are divided according to the phase of menstrual cycle, and their characteristics are presented. The EGF radioreceptor assay was performed in all of these samples, and the EGF receptor content and dissociation constants were calculated for each endometrial sample and for each phase of the menstrual cycle. The binding of 125 1 human EGF to the endometrial membrane preparations was specific, because it was largely abolished by excess nonla-
Epidermal growth factor receptor in endometrium
Volume 165 umber 5, Part I
1501
Table I. Characteristics of the participants in the endometrial EGF binding study Characteristic
Age (yr) Gravidity Parity Abortion Elective termination of pregnancy Cycle length (days)
Follicular phase*
Secretory phase*
(n = 10)
(n = 18)
30.3 2.0
(N = 28)
30.11 1.5 0.68 0.6 0.78 27.39
l.l 0.5 0.4 30.0
Combined phased 30.19 1.68 0.82 0.2 0.64 28.32
± ± ± ± ± ±
4.13 1.96 1.28 0.63 1.22 4.23
*Mean values. tMean values ± SEM.
Table II. Binding characteristics in menstrual cycle CIUlracteristic
K.. (mol/L)
Mean SEM EGF binding sites (fmol/mg of protein) Mean SEM
Proliferative phase (11 = 10) 8.80 3.10
X
]0"10
X
10- 10
24.56 8.73
Secretory phase (n = 18)
p Value*
8.34 X 10- 111 3.80 X 10- 10
0.74
24.31 15.31
0.96
*Results from comparison of proliferative and secretory phases with two-sample t test for independent samples.
beled murine EGF. The specific binding of 1251 human EGF also was saturable, with similar concentrations needed to attain saturation within each phase of the menstrual cycle. The calculated within-assay and between-assay variabilities for the EGF binding assay were 9.17% and 10.47%, respectively. In the proliferative phase endometrium, the mean (± SEM) dissociation constant calculated from the individual Kd values obtained from the Scatchard plots was 8.80 ± 3.10 X 10- 10 moUL. The mean (±SEM) total number of EGF binding sites present in the endometrial membranes, as calculated from the binding data from each endometrial sample, was 24.56 ± 8.73 fmol/mg of protein, with a minimum and maximum number of 15.84 and 39.52 fmol/mg of protein, respectively. In the secretory phase endometrium, the mean (± SEM) dissociation constant calculated was 8.34 ± 3.80 x 10- 10 mol/L, and the mean total number of available EGF binding sites was 24.31 ± 15.31 fmol/ mg of protein, with a minimum and maximum number of 3.43 and 58.69 fmol/mg of protein, respectively. The mean Kd and mean number of EGF binding sites available in the membrane preparations from each phase of the menstrual cycle were compared with each other by means of the two-sample t test for independent samples. No significant differences were found between the proliferative and secretory phases in mean Kd values and mean number of EGF binding sites, with p = 0.74 and 0.96, respectively. These data are presented in Table II. However, when the number of EGF binding sites was
evaluated as a function of time within each phase of the menstrual cycle, a difference between proliferative and secretory phases was evident. With advancing proliferative phase an increase in number of EGF binding sites was observed, which was maximal before the expected time of ovulation. Simple linear regression, performed with the EGF receptor content (number of EGF binding sites) of the endometrial membranes as the dependent variable and the proliferative phase day in which the sample was obtained as the independent variable, revealed a significant positive linear correlation with l' = 0.748andp = 0.0128 (Fig. 1). In the secretory phase the number of EGF binding sites was maximal at the beginning, decreasing with advancing secretory phase, being minimal before the time of expected menses. Simple linear regression performed with the EGF receptor content as the dependent variable and the secretory phase day as the independent variable revealed a very highly significant negative linear correlation with r = 0.843 and p = 0.0001 (Fig. 2). No significant linear correlation was seen between the dissociation constant and phase day in the proliferative and secretory phases of the menstrual cycle, with p = 0.99 and 0.8, respectively.
Comment Our study demonstrates the cyclic variation in endometrial EGF receptor content during the menstrual cycle as a function of time, confirming the previous report by Taketani and Mizuno. 13 In addition, our data are in agreement with reports by Sheets et al. 10 and
502
Troche, O'Connor, and Schaudies
1'
Key words: Epidermal growth factor receptor, endometrium, menstrual cycle Epidermal growth factor (EGF), a 53-amino-acid single-chain polypeptide, is a potent cellular mitogen for a variety of cell types.' Studies of EGF administration in vivo or in vitro have been shown to result in accelerated proliferation and differentiation of a multitude of tissues of ectodermal origin. 2 In vitro EGF administration results in increased cellular protein, ribonucleic acid and deoxyribonucleic acid synthesis, and enhanced cell multiplication." EGF exerts its biologic activity through interaction with specific transmembrane receptors, which have been found to be overexpressed in various human malignant tumors, such as squamous cell carcinomas of the head, neck, lung, and skin,' and tumors of the thyroid: brain,' breast,6 bladder," and colon. s The presence of these abnormally elevated numbers of EGF receptors have been associated with more aggressive tumors and a poor prognosis. 6 . 7 This overexpl'ession of the EGF receptor sugests a disrup-
From the Department of Obstetrics and Gynecology" and the Department of Clinical Investigation,' Walter Reed Army Medical Center, and the Uniformed Services University of the H eallh Sciences.' Supported iJy the Department of Clinical Investigation. Waller Reed Army Medical Center, Washington. D. C. Presented in part at the Thirty-eighth Annual !vI eeting of the Society for Gynecologic Investigation, San Antonio, Texas. MaTch 20-23,
1991.
The opinions and assn"tions expressed herein are those of the authot"S and are not to be construed as official or as representing the views of the Department of the Army 01' the Department of Defense. Received for publication January 9, 1991," revised April 10, 1991; accepted April 24, 1991. Reprint requests: Vladimir Troche, MD, Department of Obstetrics and Gynecology. Walter Reed Army Medical Center, Washington, DC 20307-5001. 611 130683
tion of the normal regulatory mechanisms of cellular growth in which cell proliferation is controlled by regulation of the EGF receptor in the tissue. The EGF receptor was identified in the human myometrium, endometrium, and leiomyomas." These findings support a role for EGF and its receptor in the proliferation and differentiation of the endometrium during the normal menstrual cycle. Research in this area has produced conflicting results, with no agreement as to the expression of the EGF receptor in the endometrium as a function of the menstrual cycle. EGF receptor content has been reported as follows: (I) being not different between the proliferative and secretory phases of the menstrual cycle,lo. II (2) being higher in the proliferative than the secretory phase,I2 and (3) exhibiting cyclic variation during the menstrual cycle. J3 We performed endometrial biopsies in 28 women of reproductive age at documented time points in the proliferative or secretory phase of the menstrual cycle. The endometrial tissues were assayed for the presence of the EGF receptor to investigate its expression in endometrium and establish a correlation for EGF receptor expression during the proliferative and secretory phases of the menstrual cycle.
Material and methods Materials. Chloramine-T and Coomassie Brilliant Blue G were purchased from Sigma Chemical Company, St. Louis; minimum essential medium was obtained from Hazelton Biologics, Inc., Lenexa, Kan.; bovine serum albumin (fraction 5) was purchased from Gibco Laboratories, Grand Island, N.Y.; recombinant human EGF was supplied by Upstate Biotechnology,
1499
November 1991 Am J Obstet Gynecol
1500 Troche, O'Connor, and Schaudies
Inc., Lake Placid, N.Y., and carrier-free sodium-iodine 125 was obtained from E.I. Dupont de Nemours & Co., Wilmington, Del. Patients. The study population consisted of women of reproductive age undergoing diagnostic laparoscopy or laparoscopic tubal sterilization at our ambulatory surgery center. The human use committee of our institution approved the research protocol before patient recruitment started. The participants were recruited from the infertility and family planning services at Walter Reed Army Medical Center, Washington, D.C. They were taking no medications, and their menstrual histories were normal. Patients with amenorrhea, abnormal menstrual cycles, abnormal uterine bleeding, recent use of oral contraceptives (within the last 4 months), pelvic inflammatory disease, or pregnancy were excluded from this study. Before the scheduled surgery the research protocol was explained to all candidates, and informed consent was obtained from participants. On the day ofthe scheduled laparoscopic procedures, endometrial biopsies were performed with a Novak curette while the patients were under anesthesia. The biopsy specimens were taken in an uniform manner from the anterior, posterior, and lateral walls of the uterine fundus. A representative sample of each endometrial biopsy specimen was forwarded to our gynecologic pathologist for histologic dating, and the reminder of the biopsy specimens were immediately frozen in liquid nitrogen. The endometrial tissue was stored at - 70° C until assayed. All participants were contacted within 2 to 4 weeks of surgery to document the date of onset of the next menstrual cycle. This information, together with the histologic results, was used to date the endometrial tissue. Preparation of murine EGF and iodination of human EGF. Pure murine EGF was isolated by the method of Savage and Cohen. '4 Recombinant human EGF was radioiodinated by the chloramine-T technique with carrier-free sodium- '25 !.'5 The specific activity of the radiolabeled human EGF was adjusted to 1.5 x 10 6 counts/mining by the addition of nonlabeled human EGF. Preparation of endometrial membranes. Endometrial tissues were homogenized in five volumes of icecold 10 mmol/L Tris hydrochloride, ph 7.5, containing 25 mmol/L sucrose and 1 mmol/L ethylenediaminetetraacetic acid with 15 slow strokes in a 7 ml Dounce homogenizer. The homogenates were centrifuged in a Beckman JA-20 rotor at 800 g at 4° C for 10 minutes. Nuclear pellets were discarded, and the resultant clear supernatants were centrifuged in a Beckman 60Ti rotor at 120,000 g at 4° C for 60 minutes. Supernatants were discarded and the resultant endometrial membrane pellets were resuspended in minimal essential medium containing 0.2% bovine serum albumin. The
protein content of the resuspended membrane pellets was determined by the Bradford method,16 and the final endometrial membrane protein concentration was adjusted to 6 mg/ml. Radioreceptor assay. This assay is based on the observation that nonlabeled EGF competes in an equimolar basis with radiolabeled EGF for the plasma membrane receptor and the fact that both murine and human EGF elicit nearly identical biologic responses. J. 2 Binding assays were conducted in a total volume of 200 f.L1 in 1.5 ml polyallomer microcentrifuge tubes. In all assays, 25 f.L1 (150 f.Lg) of endometrial membrane suspension was added to each tube. The binding reaction was initiated by the addition of 125 1 human EGF at nine separate concentrations ranging from 0.125 to 35 ng/ml to the microcentrifuge tubes. The incubations were performed in duplicate at room temperature for 90 minutes with gentle shaking. After incubation, 1 ml of ice-cold minimal essential medium containing 0.2% bovine serum albumin was added to each tube, followed by vortex blending. The tubes were immediately centrifuged in a Surespin centrifuge at 9500 g for 5 minutes. The supernatants were aspirated and the bound radioactivity in the pellets was counted in an LKB 1274 automatic "V-counter with a counting efficiency of 73%. Specific binding was obtained from the differences between total and nonspecific binding that was determined in parallel in the presence of a 500-fold excess of nonlabeled murine EGF. Nonspecific binding accounted for 20% of the total binding. Scatchard 17 plot analysis of the specific binding was performed for each endometrial membrane preparation, the number of binding sites, and the dissociation constant were determined. Statistical analysis of the data was performed by (I) the two-sample t test for independent samples, when the binding data of the proliferative and secretory phases were compared with each other and (2) simple linear regression, for evaluation of the binding data within each phase of the menstrual cycle.
Results The participants in the study had a mean age of 30.2 years, with a range of 25 to 38 years. A total of 28 endometrial samples were obtained from these patients, 10 of which were in the proliferative phase and 18 in the secretory phase of the menstrual cycle. In Table I the participants in the study are divided according to the phase of menstrual cycle, and their characteristics are presented. The EGF radioreceptor assay was performed in all of these samples, and the EGF receptor content and dissociation constants were calculated for each endometrial sample and for each phase of the menstrual cycle. The binding of 125 1 human EGF to the endometrial membrane preparations was specific, because it was largely abolished by excess nonla-
Epidermal growth factor receptor in endometrium
Volume 165 umber 5, Part I
1501
Table I. Characteristics of the participants in the endometrial EGF binding study Characteristic
Age (yr) Gravidity Parity Abortion Elective termination of pregnancy Cycle length (days)
Follicular phase*
Secretory phase*
(n = 10)
(n = 18)
30.3 2.0
(N = 28)
30.11 1.5 0.68 0.6 0.78 27.39
l.l 0.5 0.4 30.0
Combined phased 30.19 1.68 0.82 0.2 0.64 28.32
± ± ± ± ± ±
4.13 1.96 1.28 0.63 1.22 4.23
*Mean values. tMean values ± SEM.
Table II. Binding characteristics in menstrual cycle CIUlracteristic
K.. (mol/L)
Mean SEM EGF binding sites (fmol/mg of protein) Mean SEM
Proliferative phase (11 = 10) 8.80 3.10
X
]0"10
X
10- 10
24.56 8.73
Secretory phase (n = 18)
p Value*
8.34 X 10- 111 3.80 X 10- 10
0.74
24.31 15.31
0.96
*Results from comparison of proliferative and secretory phases with two-sample t test for independent samples.
beled murine EGF. The specific binding of 1251 human EGF also was saturable, with similar concentrations needed to attain saturation within each phase of the menstrual cycle. The calculated within-assay and between-assay variabilities for the EGF binding assay were 9.17% and 10.47%, respectively. In the proliferative phase endometrium, the mean (± SEM) dissociation constant calculated from the individual Kd values obtained from the Scatchard plots was 8.80 ± 3.10 X 10- 10 moUL. The mean (±SEM) total number of EGF binding sites present in the endometrial membranes, as calculated from the binding data from each endometrial sample, was 24.56 ± 8.73 fmol/mg of protein, with a minimum and maximum number of 15.84 and 39.52 fmol/mg of protein, respectively. In the secretory phase endometrium, the mean (± SEM) dissociation constant calculated was 8.34 ± 3.80 x 10- 10 mol/L, and the mean total number of available EGF binding sites was 24.31 ± 15.31 fmol/ mg of protein, with a minimum and maximum number of 3.43 and 58.69 fmol/mg of protein, respectively. The mean Kd and mean number of EGF binding sites available in the membrane preparations from each phase of the menstrual cycle were compared with each other by means of the two-sample t test for independent samples. No significant differences were found between the proliferative and secretory phases in mean Kd values and mean number of EGF binding sites, with p = 0.74 and 0.96, respectively. These data are presented in Table II. However, when the number of EGF binding sites was
evaluated as a function of time within each phase of the menstrual cycle, a difference between proliferative and secretory phases was evident. With advancing proliferative phase an increase in number of EGF binding sites was observed, which was maximal before the expected time of ovulation. Simple linear regression, performed with the EGF receptor content (number of EGF binding sites) of the endometrial membranes as the dependent variable and the proliferative phase day in which the sample was obtained as the independent variable, revealed a significant positive linear correlation with l' = 0.748andp = 0.0128 (Fig. 1). In the secretory phase the number of EGF binding sites was maximal at the beginning, decreasing with advancing secretory phase, being minimal before the time of expected menses. Simple linear regression performed with the EGF receptor content as the dependent variable and the secretory phase day as the independent variable revealed a very highly significant negative linear correlation with r = 0.843 and p = 0.0001 (Fig. 2). No significant linear correlation was seen between the dissociation constant and phase day in the proliferative and secretory phases of the menstrual cycle, with p = 0.99 and 0.8, respectively.
Comment Our study demonstrates the cyclic variation in endometrial EGF receptor content during the menstrual cycle as a function of time, confirming the previous report by Taketani and Mizuno. 13 In addition, our data are in agreement with reports by Sheets et al. 10 and
502
Troche, O'Connor, and Schaudies
1'
