OoZl-972x/92/7401-0172$03.00/0 Journal of Clinical Endocrinology Copyright 0 1992 by The Endocrine
and
Vol.
Metabolism Society
Printed
Increased Steroid Production by the Ovarian Tissue of Postmenopausal Women with Endometrial Cancer* MANUBAI
NAGAMANI,
CHARLES
A. STUART,
Departments of Obstetrics and Gynecology and Medicine, Galveston. Texas 77550
AND MARK
University
Stromal
G. DOHERTY
Texas Medical
Branch,
stenedione (A), testosterone, and dehydroepiandrosterone than that of women without cancer. Addition of LH resulted in a significant increase in A, testosterone, dehydroepiandrosterone, and progesterone release compared to that with vehicle alone. Addition of insulin stimulated the release of A from the ovarian stroma of women with cancer, but had no effect on the normal postmenopausal ovarian stroma. These results indicate that the ovarian stroma of postmenopausal women with endometrial cancer secrete significantly greater amounts of androgens than those of women without cancer and that both LH and insulin may be important factors contributing to this increase in ovarian steroidogenesis. (J Clin Endocrinol Metab 74: 172%176,1992)
ABSTRACT. An increase in ovarian steroid secretion could play a role in the pathogenesis of endometrial cancer in postmenopausal women. The present study was undertaken to investigate steroid production by isolated ovarian stromal tissues of postmenopausal women with endometrial cancer and to study the effect of LH and insulin on ovarian steroidogenesis in postmenopausal women, Ovarian stromal tissue was obtained from 10 postmenopausal women with endometrial cancer and 8 women without cancer. The stroma was incubated in either the medium alone or the medium to which was added LH (50 ng/ mL) or insulin (500 ng/mL). The ovarian stroma of postmenopausal women with cancer released significantly more andro-
P
of
74, No. 1
in U.S.A.
cancer, 2) to study the responsiveness of the postmenopausal ovarian stroma to LH, and 3) to study the effect of insulin on ovarian steroidogenesis in postmenopausal women.
OSTMENOPAUSAL
ovaries continue to secrete large amounts of testosterone (T) and androstenedione (A) and minimal amounts of estradiol (1, 2). The concentrations of T and A in the ovarian venous serum of postmenopausal women with endometrial cancer have been observed to be significantly higher than those in women without endometrial cancer (3). An increased incidence of stromal hyperplasia and luteinization has been observed in the ovarian stroma of postmenopausal women with endometrial cancer compared to those of women without cancer (4-6). All of these studies suggest that the ovarian stroma of women with endometrial cancer are steroidogenically more active than those of
Materials
and Methods
Patients
Ovarian stromal tissuewasobtained from 10 postmenopausal women with endometrial cancer undergoing total abdominal hysterectomy and oophorectomy. Seven of the cancer patients were obese,weighing more than 20% above their ideal body weights (IBWs). IBWs were obtained from the Metropolitan Life Insurance Co. tables. The mean percent IBW of cancer patients washigher than that of the controls. Their agesranged from 52-67 yr. All patients weremorethan 2 yr postmenopausal
women without cancer. Women with endometrial cancer have insulin resistance and increased immunoreactive insulin levels (7). Previous studies indicate that insulin stimulates ovarian steroidogenesis (8). We undertook the present study 1) to investigate steroid production by isolated ovarian stromal tissues of postmenopausal women with and without endometrial
(range, 2-17 yr). The clinical characteristics of these women are presented in Table 1. Patients with any prior history of estrogen replacement therapy were excluded from the study, since we wanted to study only women who had developed endometrial cancer due to inherent risk factors without any effects of exogenous estrogens. All except one were multiparous and had had regular ovulatory cycles before menopause.One patient was nulliparous and had a history of irregular menstrual cycles before menopause. She had been hirsute since menarche. The peripheral T level in this patient was 2.8 nmol/L, and the dehydroepiandrosterone (DHEA) level was 1.3 nmol/L. In all of the patients, peripheral steroid levels were in the normal
Received February 19, 1991. Address requests for reprints to: Manubai Nagamani, M.D., Department of Obstetrics and Gynecology, University of Texas Medical Branch, Galveston, Texas 77550. *This work was supported by Grants CA-45181 and DK-33749 from the NIH and Grant MOl-RR-00073 from the NIH General Clinical Research Centers. 172
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 03 January 2015. at 02:40 For personal use only. No other uses without permission. . All rights reserved.
OVARIAN TABLE
STEROIDS
IN ENDOMETRIAL
1. Clinical data for women with endometrial cancer and control
women
59 + 4 9+5 98 & 9” 173 + 19” 172 AZ22*
Control women (n = 8) 58 + 4 10 f 6 73 + 6 135 + 8 86 f 9
2762 f 667”
1298 + 208
5.2 -c 0.1
4.6 + 0.1
26.5 + 1.8”
19.9 + 1.3
Endometrial cancer (n = 10)
Age(yr) Yr postmenopausal Wt (kg) %IBW Fasting insulin (pmol/ L) Insulin response (pmol/ LY Fasting glucose (mmol/ L) Glucose response (mmol/L)’
Values are the mean f SE. o P < 0.05. *P < 0.01. c Sum of levels 1, 2, and 3 h after 75 g oral glucose. range and were not significantly different from the levels in the
control women. Mean T and estradiol (E2) levels were slightly higher in women with cancer compared to controls, but the difference was not statistically significant (Table 2). In all of the cancer patients, the diagnosisof endometrial cancer was establishedby fractional curettage of the uterus before the patient was admitted for hysterectomy and oophorectomy.
The
patients
were
staged in accordance
with
the
International Federation of Gynecology and Obstetrics (FIGO) stagingsystem.Eight patients had stageI cancer,one had stage II, and one had stageIII cancer. The histological grade of the tumor wasgrade 2 in six patients, grade 3 in two patients, and grade 1 in two patients. Ovarian stromal tissue was also obtained from eight postmenopausalwomen without endometrial cancer who were undergoing hysterectomy for nonendocrine problems (prolapseduterus, carcinoma in situ of cervix, and stressincontinence) and who matched the cancer patients in age and years sincemenopause.Even though five of the control patients were obese,weighing 20% above their IBWs, their mean percent IBW waslower than that of the womenwith endometrial cancer TABLE
2.
Serum T, A, DHEA, Patient
173
(Table 1). Endometrial histology in these patients varied from proliferative
Clinical findings
CANCER
to inactive
endometrium.
The ovaries of all of the
control and cancer patients wereexamined histologically. Nests of luteinized cells were observedin the ovarian stroma of eight of the womenwith endometrial cancer and two of thosewithout cancer. All women were admitted to the Clinical ResearchCenter the day before surgery, and a standard oral glucosetolerance test was performed. Blood samplesfor glucoseand insulin determinations were obtained before and hourly for 3 h after oral glucose(75 g) administration. Fasting seruminsulin levels and the insulin response(sum of insulin concentrations 1, 2, and 3 h after oral glucoseadministration) were significantly higher in women with endometrial cancer than in control women. None of the patients had overt diabetes.The fasting glucoselevels were normal, and postglucoselevels were below 200 mg/dL. While the fasting glucoselevels were not significanly different between the two groups, the glucoseresponse (sum of glucose levels 1, 2, and 3 h after oral glucose)was higher in women with endometrial cancer (Table 1). The association of increasedinsulin levels and normal glucoselevels indicates that these women with endometrial cancer have insulin resistance. Tissue culture
Immediately after oophorectomy, the ovarian stromal tissue was removed, excluding the outer 2-3 mm of ovarian cortex, and immediately placed in ice-cold medium.Under sterile conditions, the tissue was sectionedto pieceslessthan 1.5 mm in diameter. All dissectionswere performed under a laminar flow hood, and the stromal tissuesobtained from both ovarieswere pooled. A portion of the stromal tissue (-100 mg) was frozen to assessbasal steroid content. Incubation studies were performed as describedpreviously (9). About 100 mg tissuewere placed with 2 mL medium in six-well tissue culture plates (Falcon, Becton-Dickinson, Rutherford, NJ) and incubated at 37 C in 5% carbon dioxide in air. The tissue sections were incubated in either medium alone or medium to which was addedLH (50 ng/mL), insulin (500 ng/mL), or a combination of the two. All incubations were performed in duplicate. These dosesof LH and insulin were selectedon the basisof previous
P, and Es concentrations in endometrial cancer patients T (nmol/L) 2.8 0.9 1.9 0.5 0.8 1.2 0.5 0.4 0.6
A (nmol/L)
DHEA (nmol/L)
A B C D E F G H I J Mean f SE
1.1 + 0.3
4.0 4.1 3.9 4.2 1.4 1.8 4.0 4.0 4.0 2.3 3.4 + 0.3
Control women
0.6 f 0.2
3.1 zk 0.3
P (nmol/L)
1.3 4.0 1.6 2.4 11.6 9.6 2.5 2.5 2.2 2.2 4.0 + 1.1
EO (pmol/L) 172.5 77.1 150.5 102.8 36.7 73.4 73.4 77.0 51.4 161.5 99.1 + 14.7
0.17 0.09 0.09 0.06 0.48 0.08 0.17 0.09 0.06 1.10 0.24 f 0.1
3.9 f 1.2
69.7 + 14.7
0.29 + 0.1
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 03 January 2015. at 02:40 For personal use only. No other uses without permission. . All rights reserved.
174
NAGAMANI,
STUART,
AND
dose-response studies reported in experiments with ovarian stroma (10, 11). After a 24-h incubation, the medium was
60
removedfrom each well and stored at -72 C until analysis. T, A, DHEA, progesterone (P), and E2 were measured in all samples.The tissue sections from each well were weighed at this time, fixed in 10% formalin, and examined histologically with hematoxylin and eosinstain. Tissuefragmentswere found to be viable, with intact cell membranesand nuclei at the end of incubation.
50{
Incubation
JCE & M .1992 Voll4.Nol
DOHERTY ----.
.--0 m
CANCER CONTROL
I
l
5 .-z
40
F
30
x
20
\
medium
10
McCoy’s 5a mediumwith 5% newborn calf serum,penicillin (100U/mL), streptomycin (100pg/mL), and L-glutamine (0.29 mg/mL) was used as medium for all incubations. All media were bubbled in 95% oxygen and sterilized through 0.2- and 0.45-pmfilters before use. Human LH (LER 907) was obtained from the National Hormone and Pituitary Program, and human insulin was obtained from Eli Lilly Co. (Indianapolis, IN). McCoy’s 5a medium and calf serum were supplied by Gibco (Grand Island, NY). The newborn calf serumdid not have any detectable levels of steroids,LH, or insulin. All steroids were measuredby RIA after fractionation by Celite column chromatography, as describedpreviously (12, 13). Serum or incubation medium was extracted three times with 3 vol ether after tritiated tracers were added for recovery calculations.Extracts were dried under nitrogen and subjected to chromatography on micro-Celite columns.Various fractions were collected, dried under nitrogen, and reconstituted with assaybuffer. Different volumes wereusedin the assay,and 0.1 mL wasusedfor recovery calculations. All steroidswere measured by specific RIAs, as previously described(12, 13). Recoveriesrangedfrom 70-90%. All of the steroid assayshad intraassay coefficients of variation between 4.6-a% and interassay coefficients of variation between 6-10%. Serum insulin levels were measuredby a previously describedRIA procedure (14). Plasmaglucosewasmeasuredby the glucoseoxidasemethod. Steroid releaseby the ovarian stroma from the control ovary and that from women with endometrial cancer were compared by Student’s t test. Steroid productions by the tumor tissue under control and different experimental conditions were compared by one-way analysisof variance.
Results The basal steroid content of the tissues at time zero was less than 1 pg/mg tissue for all steroids. All of the stromal tissues studied released measurable amounts of steroids. Figure 1 compares the production of A, T, DHEA, Ez, and P by the isolated ovarian stromal tissue
of postmenopausal women with and without endometrial cancer. The ovarian stroma of postmenopausal women with cancer produced significantly more A [35.0 + 5.6 (k SE) us. 6.5 + 1.6 pg/mg tissue; P C O.OOl), T (23.1 + 5.7 us. 4.8 + 1.1 pg/mg tissue; P < O.Ol), and DHEA (15.1 + 1.5 us. 6.2 + 1.6 pg/mg tissue; P < 0.001) than those from control women. The production of Ez and P was not significantly different in the two groups.
-1
1
0
A
T
E2
P
DHEA
FIG. 1. Steroid production in vitro by ovarian stromal tissue of postmenopausal women with and without endometrial cancer (mean + SE). *, P < 0.001;
‘Or)
**,
r-------
80
P < 0.01.
** T
0
.;’
n
IM
m 6Q
LH INSULIN LH+insulin
60
E” \
40 z 20
0 A
T
E2
P
DHEA
2. Effects of LH and insulin on the release of steroids by the isolated ovarian stromal tissue of postmenopausal women with endometrial cancer (mean zk SE). IM, Incubation medium (control): LH, 50 ng/mL LH; INSULIN, 500 rig/ml insulin; LH + insulin, 50 ng/mL LH and 500 ng/mL insulin. *, P < 0.05; **, P < 0.001 (compared with control). FIG.
Effects of LH and insulin on ovarian stroma from women with endometrial cancer
The addition
of LH to the ovarian stroma of women
with endometrial cancer resulted in a significant increase
in A, T, P, and DHEA release compared to that with the incubation medium alone (Fig. 2). The addition of insulin resulted in a significant increase in A release compared to that in incubation medium alone, but had no effect on the secretion of the other steroids studied. The addition of insulin (500 ng/mL) plus LH (50 ng/mL) resulted in no further significant increase in the release of A compared to that with LH alone (Fig. 2). Effects of LH and insulin on steroid release by ovarian stroma from normal postmenopausal women
In incubations of ovarian stroma from control women, the addition of LH produced a significant increase in A
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 03 January 2015. at 02:40 For personal use only. No other uses without permission. . All rights reserved.
OVARIAN
STEROIDS
IN ENDOMETRIAL
and T release, but had no significant effect on Ez, P, and DHEA production. Insulin, either alone or along with LH, failed to release any of the steroids. The steroid response with LH plus insulin treatment was similar to that observed with LH alone (Fig. 3). Discussion We found a significant increase in the production of the aromatizable androgens A, T, and DHEA by the isolated ovarian stromal tissues of women with endometrial cancer compared to that in women without cancer. These results support previous in uiuo and histochemical studies on ovarian steroid secretion in postmenopausal women with endometrial cancer. We have reported increased ovarian venous A and T levels in women with endometrial cancer (3). Judd et al. (15) observed a significant decrease in A and T levels after oophorectomy in women with endometrial cancer, indicating that the ovaries of postmenopausal women with endometrial cancer secrete significant amounts of A and T. Enzyme histochemistry studies by Novak et al. (16) indicate increased steroidogenic activity in ovaries of postmenopausal women with endometrial hyperplasia and endometrial cancer compared to that in women without any endometrial pathology. Previous in vitro studies on steroidogenesis in postmenopausal ovaries are limited. Incubation studies by Mattingly and Huang (17) with [3H]pregnenolone and [3H(> as labeled precursors revealed that the postmenopausal ovarian stroma produces primarily A, T, and DHEA, and estrogens are synthesized in very limited amounts. Our results are in agreement with this study. Dennefors and his co-workers (18) observed that postmenopausal ovaries with evidence of stromal hyperplasia
653
A
T
E2
P
L!-l+insulin
DHE4
FIG. 3. Effects of LH and insulin on the release of steroids by the isolated ovarian stromal tissue of postmenopausal women without endometrial cancer (mean f SE) IM, Incubation medium (control); LH, 50 ng/mL LH; INSULIN, 500 ng/mL insulin; LH + insulin, 50 ng/mL LH and 500 ng/mL insulin. *, P C 0.01; **, P < 0.03 (compared with control).
CANCER
175
produced more A and Ez than ovarian stroma without stromal hyperplasia. The ovaries of postmenopausal women with endometrial cancer have increased incidence of stromal hyperplasia (4-6). There have not been any previous reports comparing ovarian steroidogenesis in postmenopausal women with and without cancer. The responsiveness of the ovarian stroma to LH that is observed in our present study indicates that ovarian steroidogenesis continues to be gonadotropin dependent during the postmenopausal period. Receptors to LH have been demonstrated in the postmenopausal ovarian stroma (19). Dennefors et al. (18) observed a significant increase in CAMP formation on the addition of hCG to incubations of postmenopausal ovarian stromal tissue. The ovaries of postmenopausal women with endometrial cancer were more responsive to LH than the control ovaries with regard to the amount and number of steroids released. Addition of LH to the ovarian stroma of women with endometrial cancer resulted in a significant increase in the release of A, T, DHEA, and P, while normal ovarian stroma responded to LH with the release of only A and T. Studies by Erickson et al. (20) indicate that the ovarian stromal tissue is heterogeneous and contains primary interstitial cells that have little or no capacity to synthesize steroids and secondary interstitial cells that secrete androgens. There is probably an increase in the number of steroidogenically active secondary interstitial cells in the ovarian stroma of postmenopausal women with endometrial cancer. Postmenopausal women with endometrial cancer are obese and have hyperinsulinemia. We have previously reported that the postmenopausal ovarian stroma have specific binding sites for insulin (7). One of the long term effects of insulin on the ovarian stromal cells might be to induce luteinization and convert them into steroidogenitally active luteinized stromal cells. In vitro studies by Veldhuis et al. (21) indicate that insulin stimulates lipoprotein uptake by granulosa cells. Insulin might have a similar effect on ovarian stromal cells, resulting in luteinization of these cells. We have observed a correlation between peripheral insulin levels and the extent of luteinization of ovarian stroma in premenopausal women with ovarian hyperandrogenism (22). The ovarian stroma of postmenopausal women with endometrial cancer was responsive to insulin with the release of A, while insulin had no effect on the normal ovarian stroma. The acute stimulatory effect of insulin on steroid release is probably effective only on the luteinized stromal cells. The ovarian stroma of women with endometrial cancer responded to the addition of either insulin alone or LH alone with the release of A. However, when LH plus insulin were added to the incubations, insulin had no further additive or synergistic effect on A release compared to that after LH treatment alone. It is possible
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 03 January 2015. at 02:40 For personal use only. No other uses without permission. . All rights reserved.
NAGAMANI,
176
STUART,
that there was already a maximal release of A by the stromal tissue at the dosage of 50 ng/mL LH, and the cell was not able to respond any further to the addition of insulin. A lower dose of LH may have to be used to evaluate the possible interaction between insulin and LH on androgen release. Increased production of aromatizable androgens by the ovarian stroma of women with endometrial cancer would lead to increased prehormone availability for estrogen formation from peripheral conversion. The circulating EP levels in the endometrial cancer patients were slightly higher than those in control women, but the difference was not statistically significant. The endometrial cancer patients were heavier than the control women. Since obesity is associated with lower sex hormone-binding globulin, free EP levels in these women may have been significanly higher than those in control women. Furthermore, conversion of androgens to estrogen could also take place at the endometrial tissue, thereby increasing the estrogen concentrations at the tissue level. Increased estrogen synthesis has been observed in endometrial cancer tissue compared to normal endometrium (23). Insulin has been shown to stimulate aromatase activity in human endometrial glands and stroma (24). Insulin could play a dual role in the development of endometrial cancer: 1) by stimulating the production of the aromatizable androgen A by the ovary, and 2) by increasing the aromatase activity and, therefore, increasing estrogen synthesis at the endometrial tissue level. References 1. Judd HL, Judd GE, Lucas WE, Yen SSC. Endocrine function of the postmenopausal ovary: concentration of androgens and estrogens in ovarian and peripheral vein blood. J Clin Endocrinol Metab. 1974;39:1020-4. 2. Longcope C, Hunter R, Franz C. Steroid secretion by the postmenopausal ovary. Am J Obstet Gynecol. 1980;138:564-8. 3. Nagamani M, Hannigan EV, Dillard Jr EA, Dinh TV. Ovarian steroid secretion in postmenopausal women with and without endometrial cancer. J Clin Endocrinol Metab. 198662508-12. 4. Fienberg R. The stromal theta cell and postmenopausal endometrial carcinoma. Cancer. 1969;24:32-8. 5. Marcus CC. Ovarian cortical stromal hyperplasia and cancer of the endometrium. Obstet Gynecol. 1963;21:175. 6. Sommers SC, Meissner WA. Endocrine abnormalities accompa-
AND DOHERTY
JCE & M. 1992 Voll4*No 1
nying human endometrial cancer. Cancer. 1957;10:516-21. 7. Nagamani M, Hannigan EV, Dinh VT. Stuart CA. Hvnerinsulinemia and stromal luieinization of the ovaries in postmenopausal women with endometrial cancer. J Clin Endocrinol Metab. 1988:67:144-8. 8. Barbieri RL, Makris A, Ryan KJ. Effects of insulin on steroidogenesis in cultured porcine ovarian theta. Fertil Steril. 1983;40:237-41. 9. Nagamani M, Stuart CA, Dinh TV. Steroid biosynthesis in the Sertoli-Leydig cell tumor: effects of insulin and luteinizing hormone. Am J Obstet Gvnecol. 1989:161:1738-43. 10. McNatty KP, Anastasia M, Osathanondh R. Effects of luteinizing hormone on steroidogenesis by thecal tissue from human ovarian follicles in uitro. Steroids. 1980:36:53. 11. Barbieri RL, Makris A, Randall WR, Daniels G, Kistner RW, Ryan KJ. Insulin stimulates androgen accumulation in incubation of ovarian stroma from women with hyperandrogenism. J Clin Endocrinol Metab. 1986:62:904-10. 12. Parker CR, Ellegood JO, Mahesh VB. Methods of multiple steroid radioimmunoassay. J Steroid Biochem. 1975:6:1-8. 13. Nagamani M, McDonough PG, Ellegood JO, Mahesh VB. Maternal and amniotic fluid steroids throughout human -Dregnancy. - Am J Obstet Gynecol. 1979:134:674-80. 14. Soeldner JS. Slone D. Critical variables in the radioimmunoassav of serum insulin using the double antibody technic. Diabetes. 1965:14:771-g. 15. Judd HL, Davidson BJ, Freeman AM, Shamonki MI, Lagasse LD, Ballon SC. Serum androgens and estrogens in postmenopausal women with and without endometrial cancer. Am J Obstet Gynecol. 1974:136:859-71. 16. Novak ER, Goldberg B, Jones SG, O’Toole RV. Enzyme histochemistry of the menopausal ovary associated with normal and abnormal endometrium. Am J Obstet Gynecol. 1965:93:669-82. 17. Mattingly RF, Huang WY. Steroidogenesis of the menopausal and postmenopausal ovarv. Am J Obstet Gvnecol. 1969:103:679-93. 18. Dennefors BL, Jan&n PO, Knutson-F, Hamberger L. Steroid production and responsiveness to gonadotropin in isolated stromal tissue human postmenopausal ovaries. Am J Obstet Gynecol. 1980:136:997-1002. 19. Peluso JJ, Steger RW, Jaszczak S, Hafery ESE. Gonadotropin binding sites in human postmenopausal ovaries. Fertil Steril. 1976:27:789-95. 20. Erickson GF, Magoffin DA, Dyer CA, Hofeditz C. The ovarian androgen producing cells: a review of structure function relationship. Endocr Rev. 1985:6:371-99. 21. Veldhuis JD, Nestler JE, Strauss III JF, Gwynne JJ. Insulin regulates low density lipoprotein metabolism by swine granulosa cells. Endocrinology. 1986:118:2242-53. 22. Nagamani M, Dir& TV, Kelver ME. Hyperinsulinemia in hyperthecosis of the ovaries. Am J Obstet Gvnecol. 1986:154:384-g. 23. Tseng L, Mozella J, Mann WT, Chumas J. Estrogen synthesis in normal and malignant human endometrium. J Clin Endocrinol Metab. 1982:55:1029-31. 24. Randolph JF, Kipersztok S, Ayers JWT, Ansbacker R, Peegal H, Menon KMJ. The effect of insulin on aromatase activity in isolated human endometrial glands and stroma. Am J Obstet Gynecol. 1987:157:1534-g.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 03 January 2015. at 02:40 For personal use only. No other uses without permission. . All rights reserved.
and
Vol.
Metabolism Society
Printed
Increased Steroid Production by the Ovarian Tissue of Postmenopausal Women with Endometrial Cancer* MANUBAI
NAGAMANI,
CHARLES
A. STUART,
Departments of Obstetrics and Gynecology and Medicine, Galveston. Texas 77550
AND MARK
University
Stromal
G. DOHERTY
Texas Medical
Branch,
stenedione (A), testosterone, and dehydroepiandrosterone than that of women without cancer. Addition of LH resulted in a significant increase in A, testosterone, dehydroepiandrosterone, and progesterone release compared to that with vehicle alone. Addition of insulin stimulated the release of A from the ovarian stroma of women with cancer, but had no effect on the normal postmenopausal ovarian stroma. These results indicate that the ovarian stroma of postmenopausal women with endometrial cancer secrete significantly greater amounts of androgens than those of women without cancer and that both LH and insulin may be important factors contributing to this increase in ovarian steroidogenesis. (J Clin Endocrinol Metab 74: 172%176,1992)
ABSTRACT. An increase in ovarian steroid secretion could play a role in the pathogenesis of endometrial cancer in postmenopausal women. The present study was undertaken to investigate steroid production by isolated ovarian stromal tissues of postmenopausal women with endometrial cancer and to study the effect of LH and insulin on ovarian steroidogenesis in postmenopausal women, Ovarian stromal tissue was obtained from 10 postmenopausal women with endometrial cancer and 8 women without cancer. The stroma was incubated in either the medium alone or the medium to which was added LH (50 ng/ mL) or insulin (500 ng/mL). The ovarian stroma of postmenopausal women with cancer released significantly more andro-
P
of
74, No. 1
in U.S.A.
cancer, 2) to study the responsiveness of the postmenopausal ovarian stroma to LH, and 3) to study the effect of insulin on ovarian steroidogenesis in postmenopausal women.
OSTMENOPAUSAL
ovaries continue to secrete large amounts of testosterone (T) and androstenedione (A) and minimal amounts of estradiol (1, 2). The concentrations of T and A in the ovarian venous serum of postmenopausal women with endometrial cancer have been observed to be significantly higher than those in women without endometrial cancer (3). An increased incidence of stromal hyperplasia and luteinization has been observed in the ovarian stroma of postmenopausal women with endometrial cancer compared to those of women without cancer (4-6). All of these studies suggest that the ovarian stroma of women with endometrial cancer are steroidogenically more active than those of
Materials
and Methods
Patients
Ovarian stromal tissuewasobtained from 10 postmenopausal women with endometrial cancer undergoing total abdominal hysterectomy and oophorectomy. Seven of the cancer patients were obese,weighing more than 20% above their ideal body weights (IBWs). IBWs were obtained from the Metropolitan Life Insurance Co. tables. The mean percent IBW of cancer patients washigher than that of the controls. Their agesranged from 52-67 yr. All patients weremorethan 2 yr postmenopausal
women without cancer. Women with endometrial cancer have insulin resistance and increased immunoreactive insulin levels (7). Previous studies indicate that insulin stimulates ovarian steroidogenesis (8). We undertook the present study 1) to investigate steroid production by isolated ovarian stromal tissues of postmenopausal women with and without endometrial
(range, 2-17 yr). The clinical characteristics of these women are presented in Table 1. Patients with any prior history of estrogen replacement therapy were excluded from the study, since we wanted to study only women who had developed endometrial cancer due to inherent risk factors without any effects of exogenous estrogens. All except one were multiparous and had had regular ovulatory cycles before menopause.One patient was nulliparous and had a history of irregular menstrual cycles before menopause. She had been hirsute since menarche. The peripheral T level in this patient was 2.8 nmol/L, and the dehydroepiandrosterone (DHEA) level was 1.3 nmol/L. In all of the patients, peripheral steroid levels were in the normal
Received February 19, 1991. Address requests for reprints to: Manubai Nagamani, M.D., Department of Obstetrics and Gynecology, University of Texas Medical Branch, Galveston, Texas 77550. *This work was supported by Grants CA-45181 and DK-33749 from the NIH and Grant MOl-RR-00073 from the NIH General Clinical Research Centers. 172
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 03 January 2015. at 02:40 For personal use only. No other uses without permission. . All rights reserved.
OVARIAN TABLE
STEROIDS
IN ENDOMETRIAL
1. Clinical data for women with endometrial cancer and control
women
59 + 4 9+5 98 & 9” 173 + 19” 172 AZ22*
Control women (n = 8) 58 + 4 10 f 6 73 + 6 135 + 8 86 f 9
2762 f 667”
1298 + 208
5.2 -c 0.1
4.6 + 0.1
26.5 + 1.8”
19.9 + 1.3
Endometrial cancer (n = 10)
Age(yr) Yr postmenopausal Wt (kg) %IBW Fasting insulin (pmol/ L) Insulin response (pmol/ LY Fasting glucose (mmol/ L) Glucose response (mmol/L)’
Values are the mean f SE. o P < 0.05. *P < 0.01. c Sum of levels 1, 2, and 3 h after 75 g oral glucose. range and were not significantly different from the levels in the
control women. Mean T and estradiol (E2) levels were slightly higher in women with cancer compared to controls, but the difference was not statistically significant (Table 2). In all of the cancer patients, the diagnosisof endometrial cancer was establishedby fractional curettage of the uterus before the patient was admitted for hysterectomy and oophorectomy.
The
patients
were
staged in accordance
with
the
International Federation of Gynecology and Obstetrics (FIGO) stagingsystem.Eight patients had stageI cancer,one had stage II, and one had stageIII cancer. The histological grade of the tumor wasgrade 2 in six patients, grade 3 in two patients, and grade 1 in two patients. Ovarian stromal tissue was also obtained from eight postmenopausalwomen without endometrial cancer who were undergoing hysterectomy for nonendocrine problems (prolapseduterus, carcinoma in situ of cervix, and stressincontinence) and who matched the cancer patients in age and years sincemenopause.Even though five of the control patients were obese,weighing 20% above their IBWs, their mean percent IBW waslower than that of the womenwith endometrial cancer TABLE
2.
Serum T, A, DHEA, Patient
173
(Table 1). Endometrial histology in these patients varied from proliferative
Clinical findings
CANCER
to inactive
endometrium.
The ovaries of all of the
control and cancer patients wereexamined histologically. Nests of luteinized cells were observedin the ovarian stroma of eight of the womenwith endometrial cancer and two of thosewithout cancer. All women were admitted to the Clinical ResearchCenter the day before surgery, and a standard oral glucosetolerance test was performed. Blood samplesfor glucoseand insulin determinations were obtained before and hourly for 3 h after oral glucose(75 g) administration. Fasting seruminsulin levels and the insulin response(sum of insulin concentrations 1, 2, and 3 h after oral glucoseadministration) were significantly higher in women with endometrial cancer than in control women. None of the patients had overt diabetes.The fasting glucoselevels were normal, and postglucoselevels were below 200 mg/dL. While the fasting glucoselevels were not significanly different between the two groups, the glucoseresponse (sum of glucose levels 1, 2, and 3 h after oral glucose)was higher in women with endometrial cancer (Table 1). The association of increasedinsulin levels and normal glucoselevels indicates that these women with endometrial cancer have insulin resistance. Tissue culture
Immediately after oophorectomy, the ovarian stromal tissue was removed, excluding the outer 2-3 mm of ovarian cortex, and immediately placed in ice-cold medium.Under sterile conditions, the tissue was sectionedto pieceslessthan 1.5 mm in diameter. All dissectionswere performed under a laminar flow hood, and the stromal tissuesobtained from both ovarieswere pooled. A portion of the stromal tissue (-100 mg) was frozen to assessbasal steroid content. Incubation studies were performed as describedpreviously (9). About 100 mg tissuewere placed with 2 mL medium in six-well tissue culture plates (Falcon, Becton-Dickinson, Rutherford, NJ) and incubated at 37 C in 5% carbon dioxide in air. The tissue sections were incubated in either medium alone or medium to which was addedLH (50 ng/mL), insulin (500 ng/mL), or a combination of the two. All incubations were performed in duplicate. These dosesof LH and insulin were selectedon the basisof previous
P, and Es concentrations in endometrial cancer patients T (nmol/L) 2.8 0.9 1.9 0.5 0.8 1.2 0.5 0.4 0.6
A (nmol/L)
DHEA (nmol/L)
A B C D E F G H I J Mean f SE
1.1 + 0.3
4.0 4.1 3.9 4.2 1.4 1.8 4.0 4.0 4.0 2.3 3.4 + 0.3
Control women
0.6 f 0.2
3.1 zk 0.3
P (nmol/L)
1.3 4.0 1.6 2.4 11.6 9.6 2.5 2.5 2.2 2.2 4.0 + 1.1
EO (pmol/L) 172.5 77.1 150.5 102.8 36.7 73.4 73.4 77.0 51.4 161.5 99.1 + 14.7
0.17 0.09 0.09 0.06 0.48 0.08 0.17 0.09 0.06 1.10 0.24 f 0.1
3.9 f 1.2
69.7 + 14.7
0.29 + 0.1
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 03 January 2015. at 02:40 For personal use only. No other uses without permission. . All rights reserved.
174
NAGAMANI,
STUART,
AND
dose-response studies reported in experiments with ovarian stroma (10, 11). After a 24-h incubation, the medium was
60
removedfrom each well and stored at -72 C until analysis. T, A, DHEA, progesterone (P), and E2 were measured in all samples.The tissue sections from each well were weighed at this time, fixed in 10% formalin, and examined histologically with hematoxylin and eosinstain. Tissuefragmentswere found to be viable, with intact cell membranesand nuclei at the end of incubation.
50{
Incubation
JCE & M .1992 Voll4.Nol
DOHERTY ----.
.--0 m
CANCER CONTROL
I
l
5 .-z
40
F
30
x
20
\
medium
10
McCoy’s 5a mediumwith 5% newborn calf serum,penicillin (100U/mL), streptomycin (100pg/mL), and L-glutamine (0.29 mg/mL) was used as medium for all incubations. All media were bubbled in 95% oxygen and sterilized through 0.2- and 0.45-pmfilters before use. Human LH (LER 907) was obtained from the National Hormone and Pituitary Program, and human insulin was obtained from Eli Lilly Co. (Indianapolis, IN). McCoy’s 5a medium and calf serum were supplied by Gibco (Grand Island, NY). The newborn calf serumdid not have any detectable levels of steroids,LH, or insulin. All steroids were measuredby RIA after fractionation by Celite column chromatography, as describedpreviously (12, 13). Serum or incubation medium was extracted three times with 3 vol ether after tritiated tracers were added for recovery calculations.Extracts were dried under nitrogen and subjected to chromatography on micro-Celite columns.Various fractions were collected, dried under nitrogen, and reconstituted with assaybuffer. Different volumes wereusedin the assay,and 0.1 mL wasusedfor recovery calculations. All steroidswere measured by specific RIAs, as previously described(12, 13). Recoveriesrangedfrom 70-90%. All of the steroid assayshad intraassay coefficients of variation between 4.6-a% and interassay coefficients of variation between 6-10%. Serum insulin levels were measuredby a previously describedRIA procedure (14). Plasmaglucosewasmeasuredby the glucoseoxidasemethod. Steroid releaseby the ovarian stroma from the control ovary and that from women with endometrial cancer were compared by Student’s t test. Steroid productions by the tumor tissue under control and different experimental conditions were compared by one-way analysisof variance.
Results The basal steroid content of the tissues at time zero was less than 1 pg/mg tissue for all steroids. All of the stromal tissues studied released measurable amounts of steroids. Figure 1 compares the production of A, T, DHEA, Ez, and P by the isolated ovarian stromal tissue
of postmenopausal women with and without endometrial cancer. The ovarian stroma of postmenopausal women with cancer produced significantly more A [35.0 + 5.6 (k SE) us. 6.5 + 1.6 pg/mg tissue; P C O.OOl), T (23.1 + 5.7 us. 4.8 + 1.1 pg/mg tissue; P < O.Ol), and DHEA (15.1 + 1.5 us. 6.2 + 1.6 pg/mg tissue; P < 0.001) than those from control women. The production of Ez and P was not significantly different in the two groups.
-1
1
0
A
T
E2
P
DHEA
FIG. 1. Steroid production in vitro by ovarian stromal tissue of postmenopausal women with and without endometrial cancer (mean + SE). *, P < 0.001;
‘Or)
**,
r-------
80
P < 0.01.
** T
0
.;’
n
IM
m 6Q
LH INSULIN LH+insulin
60
E” \
40 z 20
0 A
T
E2
P
DHEA
2. Effects of LH and insulin on the release of steroids by the isolated ovarian stromal tissue of postmenopausal women with endometrial cancer (mean zk SE). IM, Incubation medium (control): LH, 50 ng/mL LH; INSULIN, 500 rig/ml insulin; LH + insulin, 50 ng/mL LH and 500 ng/mL insulin. *, P < 0.05; **, P < 0.001 (compared with control). FIG.
Effects of LH and insulin on ovarian stroma from women with endometrial cancer
The addition
of LH to the ovarian stroma of women
with endometrial cancer resulted in a significant increase
in A, T, P, and DHEA release compared to that with the incubation medium alone (Fig. 2). The addition of insulin resulted in a significant increase in A release compared to that in incubation medium alone, but had no effect on the secretion of the other steroids studied. The addition of insulin (500 ng/mL) plus LH (50 ng/mL) resulted in no further significant increase in the release of A compared to that with LH alone (Fig. 2). Effects of LH and insulin on steroid release by ovarian stroma from normal postmenopausal women
In incubations of ovarian stroma from control women, the addition of LH produced a significant increase in A
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 03 January 2015. at 02:40 For personal use only. No other uses without permission. . All rights reserved.
OVARIAN
STEROIDS
IN ENDOMETRIAL
and T release, but had no significant effect on Ez, P, and DHEA production. Insulin, either alone or along with LH, failed to release any of the steroids. The steroid response with LH plus insulin treatment was similar to that observed with LH alone (Fig. 3). Discussion We found a significant increase in the production of the aromatizable androgens A, T, and DHEA by the isolated ovarian stromal tissues of women with endometrial cancer compared to that in women without cancer. These results support previous in uiuo and histochemical studies on ovarian steroid secretion in postmenopausal women with endometrial cancer. We have reported increased ovarian venous A and T levels in women with endometrial cancer (3). Judd et al. (15) observed a significant decrease in A and T levels after oophorectomy in women with endometrial cancer, indicating that the ovaries of postmenopausal women with endometrial cancer secrete significant amounts of A and T. Enzyme histochemistry studies by Novak et al. (16) indicate increased steroidogenic activity in ovaries of postmenopausal women with endometrial hyperplasia and endometrial cancer compared to that in women without any endometrial pathology. Previous in vitro studies on steroidogenesis in postmenopausal ovaries are limited. Incubation studies by Mattingly and Huang (17) with [3H]pregnenolone and [3H(> as labeled precursors revealed that the postmenopausal ovarian stroma produces primarily A, T, and DHEA, and estrogens are synthesized in very limited amounts. Our results are in agreement with this study. Dennefors and his co-workers (18) observed that postmenopausal ovaries with evidence of stromal hyperplasia
653
A
T
E2
P
L!-l+insulin
DHE4
FIG. 3. Effects of LH and insulin on the release of steroids by the isolated ovarian stromal tissue of postmenopausal women without endometrial cancer (mean f SE) IM, Incubation medium (control); LH, 50 ng/mL LH; INSULIN, 500 ng/mL insulin; LH + insulin, 50 ng/mL LH and 500 ng/mL insulin. *, P C 0.01; **, P < 0.03 (compared with control).
CANCER
175
produced more A and Ez than ovarian stroma without stromal hyperplasia. The ovaries of postmenopausal women with endometrial cancer have increased incidence of stromal hyperplasia (4-6). There have not been any previous reports comparing ovarian steroidogenesis in postmenopausal women with and without cancer. The responsiveness of the ovarian stroma to LH that is observed in our present study indicates that ovarian steroidogenesis continues to be gonadotropin dependent during the postmenopausal period. Receptors to LH have been demonstrated in the postmenopausal ovarian stroma (19). Dennefors et al. (18) observed a significant increase in CAMP formation on the addition of hCG to incubations of postmenopausal ovarian stromal tissue. The ovaries of postmenopausal women with endometrial cancer were more responsive to LH than the control ovaries with regard to the amount and number of steroids released. Addition of LH to the ovarian stroma of women with endometrial cancer resulted in a significant increase in the release of A, T, DHEA, and P, while normal ovarian stroma responded to LH with the release of only A and T. Studies by Erickson et al. (20) indicate that the ovarian stromal tissue is heterogeneous and contains primary interstitial cells that have little or no capacity to synthesize steroids and secondary interstitial cells that secrete androgens. There is probably an increase in the number of steroidogenically active secondary interstitial cells in the ovarian stroma of postmenopausal women with endometrial cancer. Postmenopausal women with endometrial cancer are obese and have hyperinsulinemia. We have previously reported that the postmenopausal ovarian stroma have specific binding sites for insulin (7). One of the long term effects of insulin on the ovarian stromal cells might be to induce luteinization and convert them into steroidogenitally active luteinized stromal cells. In vitro studies by Veldhuis et al. (21) indicate that insulin stimulates lipoprotein uptake by granulosa cells. Insulin might have a similar effect on ovarian stromal cells, resulting in luteinization of these cells. We have observed a correlation between peripheral insulin levels and the extent of luteinization of ovarian stroma in premenopausal women with ovarian hyperandrogenism (22). The ovarian stroma of postmenopausal women with endometrial cancer was responsive to insulin with the release of A, while insulin had no effect on the normal ovarian stroma. The acute stimulatory effect of insulin on steroid release is probably effective only on the luteinized stromal cells. The ovarian stroma of women with endometrial cancer responded to the addition of either insulin alone or LH alone with the release of A. However, when LH plus insulin were added to the incubations, insulin had no further additive or synergistic effect on A release compared to that after LH treatment alone. It is possible
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 03 January 2015. at 02:40 For personal use only. No other uses without permission. . All rights reserved.
NAGAMANI,
176
STUART,
that there was already a maximal release of A by the stromal tissue at the dosage of 50 ng/mL LH, and the cell was not able to respond any further to the addition of insulin. A lower dose of LH may have to be used to evaluate the possible interaction between insulin and LH on androgen release. Increased production of aromatizable androgens by the ovarian stroma of women with endometrial cancer would lead to increased prehormone availability for estrogen formation from peripheral conversion. The circulating EP levels in the endometrial cancer patients were slightly higher than those in control women, but the difference was not statistically significant. The endometrial cancer patients were heavier than the control women. Since obesity is associated with lower sex hormone-binding globulin, free EP levels in these women may have been significanly higher than those in control women. Furthermore, conversion of androgens to estrogen could also take place at the endometrial tissue, thereby increasing the estrogen concentrations at the tissue level. Increased estrogen synthesis has been observed in endometrial cancer tissue compared to normal endometrium (23). Insulin has been shown to stimulate aromatase activity in human endometrial glands and stroma (24). Insulin could play a dual role in the development of endometrial cancer: 1) by stimulating the production of the aromatizable androgen A by the ovary, and 2) by increasing the aromatase activity and, therefore, increasing estrogen synthesis at the endometrial tissue level. References 1. Judd HL, Judd GE, Lucas WE, Yen SSC. Endocrine function of the postmenopausal ovary: concentration of androgens and estrogens in ovarian and peripheral vein blood. J Clin Endocrinol Metab. 1974;39:1020-4. 2. Longcope C, Hunter R, Franz C. Steroid secretion by the postmenopausal ovary. Am J Obstet Gynecol. 1980;138:564-8. 3. Nagamani M, Hannigan EV, Dillard Jr EA, Dinh TV. Ovarian steroid secretion in postmenopausal women with and without endometrial cancer. J Clin Endocrinol Metab. 198662508-12. 4. Fienberg R. The stromal theta cell and postmenopausal endometrial carcinoma. Cancer. 1969;24:32-8. 5. Marcus CC. Ovarian cortical stromal hyperplasia and cancer of the endometrium. Obstet Gynecol. 1963;21:175. 6. Sommers SC, Meissner WA. Endocrine abnormalities accompa-
AND DOHERTY
JCE & M. 1992 Voll4*No 1
nying human endometrial cancer. Cancer. 1957;10:516-21. 7. Nagamani M, Hannigan EV, Dinh VT. Stuart CA. Hvnerinsulinemia and stromal luieinization of the ovaries in postmenopausal women with endometrial cancer. J Clin Endocrinol Metab. 1988:67:144-8. 8. Barbieri RL, Makris A, Ryan KJ. Effects of insulin on steroidogenesis in cultured porcine ovarian theta. Fertil Steril. 1983;40:237-41. 9. Nagamani M, Stuart CA, Dinh TV. Steroid biosynthesis in the Sertoli-Leydig cell tumor: effects of insulin and luteinizing hormone. Am J Obstet Gvnecol. 1989:161:1738-43. 10. McNatty KP, Anastasia M, Osathanondh R. Effects of luteinizing hormone on steroidogenesis by thecal tissue from human ovarian follicles in uitro. Steroids. 1980:36:53. 11. Barbieri RL, Makris A, Randall WR, Daniels G, Kistner RW, Ryan KJ. Insulin stimulates androgen accumulation in incubation of ovarian stroma from women with hyperandrogenism. J Clin Endocrinol Metab. 1986:62:904-10. 12. Parker CR, Ellegood JO, Mahesh VB. Methods of multiple steroid radioimmunoassay. J Steroid Biochem. 1975:6:1-8. 13. Nagamani M, McDonough PG, Ellegood JO, Mahesh VB. Maternal and amniotic fluid steroids throughout human -Dregnancy. - Am J Obstet Gynecol. 1979:134:674-80. 14. Soeldner JS. Slone D. Critical variables in the radioimmunoassav of serum insulin using the double antibody technic. Diabetes. 1965:14:771-g. 15. Judd HL, Davidson BJ, Freeman AM, Shamonki MI, Lagasse LD, Ballon SC. Serum androgens and estrogens in postmenopausal women with and without endometrial cancer. Am J Obstet Gynecol. 1974:136:859-71. 16. Novak ER, Goldberg B, Jones SG, O’Toole RV. Enzyme histochemistry of the menopausal ovary associated with normal and abnormal endometrium. Am J Obstet Gynecol. 1965:93:669-82. 17. Mattingly RF, Huang WY. Steroidogenesis of the menopausal and postmenopausal ovarv. Am J Obstet Gvnecol. 1969:103:679-93. 18. Dennefors BL, Jan&n PO, Knutson-F, Hamberger L. Steroid production and responsiveness to gonadotropin in isolated stromal tissue human postmenopausal ovaries. Am J Obstet Gynecol. 1980:136:997-1002. 19. Peluso JJ, Steger RW, Jaszczak S, Hafery ESE. Gonadotropin binding sites in human postmenopausal ovaries. Fertil Steril. 1976:27:789-95. 20. Erickson GF, Magoffin DA, Dyer CA, Hofeditz C. The ovarian androgen producing cells: a review of structure function relationship. Endocr Rev. 1985:6:371-99. 21. Veldhuis JD, Nestler JE, Strauss III JF, Gwynne JJ. Insulin regulates low density lipoprotein metabolism by swine granulosa cells. Endocrinology. 1986:118:2242-53. 22. Nagamani M, Dir& TV, Kelver ME. Hyperinsulinemia in hyperthecosis of the ovaries. Am J Obstet Gvnecol. 1986:154:384-g. 23. Tseng L, Mozella J, Mann WT, Chumas J. Estrogen synthesis in normal and malignant human endometrium. J Clin Endocrinol Metab. 1982:55:1029-31. 24. Randolph JF, Kipersztok S, Ayers JWT, Ansbacker R, Peegal H, Menon KMJ. The effect of insulin on aromatase activity in isolated human endometrial glands and stroma. Am J Obstet Gynecol. 1987:157:1534-g.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 03 January 2015. at 02:40 For personal use only. No other uses without permission. . All rights reserved.
