| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Departments of Obstetrics/Gynecology (R.T.C., A.S.G., E.T.M.), Physiology (R.T.C.), Preventive Medicine (I.B.H., P.H.G.), and the Robert H. Lurie Comprehensive Cancer Center (R.T.C., A.S.G., E.T.M., I.B.H., P.H.G.), Feinberg School of Medicine, Northwestern University, Chicago, Illinois 60611
Address all correspondence and requests for reprints to: Robert T. Chatterton, Ph.D., Department of Ob/Gyn, 710 North Fairbanks Court, Room 8408, Chicago, Illinois 60611. E-mail: chat{at}northwestern.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Premenopausal women were those described in previous reports (1, 7). This group of 47 women formed the basis for comparison of the other groups. The age of the women ranged from 20–40 yr. They had menstrual cycles of 25–35 d, and they were not lactating and had not taken OCs or other drugs that may affect hormone metabolism within the last 6 months. The other groups were 13 women taking OCs, 18 postmenopausal women (including three who had an oophorohysterectomy) who were not taking HRT, and seven women who were taking HRT.
The mean age of the group taking OCs was 30.4 ± 5.1 yr (range, 22–36 yr) with a mean body mass index (BMI) of 25.2 ± 8.0 kg/m2. They had been taking OCs for at least 4 months before entry into the study. They came to the General Clinical Research Center at Northwestern Memorial Hospital on three occasions within a month: the first visit was 6 d after completing a cycle of active drug; the next visit was 14 d later; and the last visit was 21 d after the first visit. All women were taking tricyclic OCs, Tri-Norinyl (Searle), Ortho-Novum (Ortho-McNeil), Tri-Levlen (Berlex), Triphasil (Wyeth-Ayerst), or Otho Tri-Cyclen (Ortho-McNeil). None were taking any medications, other than the OCs, that might interfere with the hormone measurements to be made, nor did any of these subjects have any chronic diseases that might compromise metabolism. Of 50 subjects on whom NAF collection was attempted, only 13 (26%) produced more than 50 µl adequate for analyses.
The mean age of postmenopausal women not taking HRT was 50.9 ± 4.8 yr (range, 43–66 yr) with a mean BMI of 27.0 ± 5.5 kg/m2. They had not had a menstrual period for more than 1 yr before sampling or had had an oophorohysterectomy more than 1 yr before sampling. Serum levels of estradiol and progesterone (see Table 2
) were in the range expected for postmenopausal women. The mean 5-yr Gail score for breast cancer risk was 0.87 ± 0.37. None had had a menstrual period within the previous 6 months. None were taking any medications that might interfere with the hormone measurements to be made, nor did any of these subjects have any chronic diseases that might compromise metabolism. Three samples were collected from each volunteer not closer than 2 wk apart. Of 60 subjects on whom NAF collection was attempted, 18 (30%) produced sufficient volume for analyses.
|
All procedures involving human subjects were approved by the Institutional Review Board of Northwestern University.
Sample collection
Blood and NAF were obtained at each visit for hormone measurements. NAF was collected by warming and massaging the breast and aspirating a small volume of liquid from the nipple using a specially designed vacuum device (Cytyc Corp., Boxborough, MA). The NAF was collected in a calibrated capillary tube, the ends were sealed, and the sample was stored at –20 C.
Analytical methods
The procedure for analysis of the NAF has been described in detail recently (1). Briefly, NAF was diluted with PBS (pH 7.4) and extracted with ethyl acetate-hexane (3:2). The organic extract was evaporated and partitioned between isooctane and 0.4 N aqueous NaOH to separate the respective phenolic from nonphenolic steroids. The proteins and steroid sulfates [EGF, cathepsin D, estrone sulfate, and dehydroepiandrosterone (DHEA) sulfate] were measured in the remaining aqueous phase after ethyl acetate-hexane extraction; DHEA, androstenedione, and progesterone were measured in the nonphenolic fraction, and estradiol was measured in the phenolic fraction. All measurements were performed by standard immunoassays in the reconstituted fractions (1). The cross-reaction of ethynyl estradiol in the estradiol assay was less than 0.1%, and that of medroxyprogesterone acetate in the progesterone assay was less than 0.01%. The intraassay coefficients of variation (CVs) for the assays were as follows: EGF, 7.5%; cathepsin D, 12.6%; estrone sulfate, 6.9%; DHEA sulfate, 8.0%; DHEA, 4.6%; androstenedione, 15.2%; and progesterone, 16.0%. Insufficient batches of each analyte were assayed for estimation of interassay CV. Analytes with concentrations close to the limit of sensitivity of the method had higher CVs. Serum hormones and other products were measured by standard immunoassays as described in detail previously (1, 8).
Statistical methods
All data were transformed to their natural logs for comparisons. This transformation led to adequate normalization of the variables. Mean skewness values, calculated by the SAS, version 8 (SAS Institute Inc., Cary, NC), univariate procedure on logs of the variables were all less than ± 1.0 except for EGF, which had a value of ±1.6. Significant skewness with an N of 35 per comparison by this procedure is ±1.07. This brings some additional uncertainty in the estimates of the significance of comparisons with EGF. The significance of the difference between values in each group from those of the premenopausal subjects was determined by group t tests adjusted by the Bonferroni procedure for multiple comparisons. The means and SD values presented are the antilogs of the calculated values. Rank order correlations were calculated by the Spearman method. Intraclass correlation coefficients (ICCs) were calculated as described by Pinheiro and Bates (9), with confidence limits determined by the procedure of Jovanovic and Viana (10).
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
No significant difference in levels of progesterone occurred between the first, second, or third samples collected from each subject, although the first sample was collected 6 d after the last active medication was taken; the mean concentrations in the first and third samples were 0.53 ± 1.44 (SE) ng/ml (1.7 ± 4.6 nmol/liter) and 0.41 ± 1.16 (SE) ng/ml (1.3 ± 3.7 nmol/liter), respectively. Estradiol in the first and third samples was, however, different (P < 0.01) with a doubling from the 21-d sample 15.2 ± 1.20 (SE) pg/ml (55.8 ± 4.4 pmol/liter) to the sixth day after cessation of OC administration [33.2 ± 1.2 (SE) pg/ml, i.e. 121.9 ± 4.4 pmol/liter].
Progesterone was suppressed by OCs to insignificant levels in both serum and NAF. The correlation of serum and NAF progesterone was 0.18 in controls and 0.00 with OCs. The concentrations of estradiol in serum and NAF of premenopausal women not taking OCs (control group) were almost identical (Tables 1 and 2
). Compared with controls, women taking OCs had almost identical, significantly lower levels of estradiol in serum and NAF (71 and 68% lower, averaged across the three time periods). Although hormone concentrations varied across visits in OC users, the average reasonably reflects integrated levels during the course of 1 month of OC use. Nevertheless, the correlations between estradiol concentrations in serum and NAF, measured in samples collected simultaneously, were only 0.25 and 0.22 in controls and women with OCs, respectively. Whereas the concentration of estrone sulfate was 150-fold greater in NAF than serum, the suppression by OCs was proportionally similar, 79% in serum and 89% in NAF. None of the other steroids in NAF were affected by the treatment. Cathepsin D, EGF, and IL-6 in NAF were also not significantly altered by the combination of endogenous and exogenous hormones.
|
The three samples from each postmenopausal woman were taken as a single estimate to represent that subject. The mean progesterone concentration was only 6.7 nmol/liter in the postmenopausal women (Table 1). However, individuals varied greatly. Eight of 18 women had at least one breast with a value of 37.7 ng/ml (120 nmol/liter) or more. The mean of the 13 breasts with the highest values was 145 ± 2.5 (SE) ng/ml (462 ± 7.9 nmol/liter). The increase was specific to a breast. For example, the right breast of one woman had 454 and 350 ng/ml (1444 and 1113 nmol/liter) on two occasions, whereas her left breast had less than 0.9 ng/ml (<3.0 nmol/liter). Serum progesterone levels in these women averaged 0.38 ± 1.19 (SE) ng/ml (1.2 ± 3.8 nmol/liter). The correlation between serum and NAF progesterone was only 0.13.
The mean progesterone level of the women with high BMI (those with >37.7 ng/ml, i.e. >120 nmol/liter) was significantly greater than that of the other postmenopausal women, 10.0 ± 1.8 (SE) and 7.9 ± 1.2 (SE) ng/ml (31.7 ± 5.8 and 25.3 ± 3.8 nmol/liter), respectively (P = 0.01). The subjects with high progesterone did not have higher estradiol levels than the group as a whole.
Compared with the premenopausal control group, mean NAF estradiol concentrations were 62% lower in the postmenopausal women (Table 1). However, this was not significant despite the highly significantly lower (71% lower) serum levels in the postmenopausal women (Table 2). The low levels of estradiol were associated with a negligible correlation between serum and NAF in this group of –0.02. The concentration of estrone sulfate was significantly lower in serum of postmenopausal women but not in NAF, and the correlation between serum and NAF in the postmenopausal women was relatively high at 0.30. In confirmation of the continued presence of estrogen in NAF, no changes in cathepsin D or EGF were observed from values in the premenopausal women. Samples collected but not analyzed in an earlier study of premenopausal women (11) were assayed for cathepsin D and EGF to compare with the postmenopausal women; the concentrations were comparable to those of the premenopausal women in the present study: 1628 ng/ml (32 nmol/liter) and 234 ng/ml (36 nmol/liter), respectively. IL-6 was significantly higher in the postmenopausal women, but no other analytes were different from those found in premenopausal women.
The ICCs with their 95% confidence limits are shown in Table 3. Values for premenopausal women are shown for comparison. The estimates in the postmenopausal women were based on three samples over a period of 2–3 months, whereas those for premenopausal women were estimated from 47 women over a period of 15 months. For analytes in which both NAF and serum ICCs could be calculated, the NAF ICCs were in the same range or slightly less. Progesterone was too low for a meaningful estimate in the postmenopausal group. Most ICCs were similar to those of the premenopausal group with values for estrone sulfate, EGF, and DHEA sulfate being highest.
|
The three samples from each postmenopausal woman on HRT were taken as a single estimate to represent that subject. Estradiol in NAF was seven times greater in the women on HRT than in the premenopausal women and 18 times greater than in postmenopausal control women (Table 1). By comparison, serum estradiol was only 2.5-fold higher in HRT users, and serum estrone was 3.7-fold higher (Table 2). With the exception of the control postmenopausal women, the correlation between serum and NAF, 0.20, was similar to the other groups. The mean concentration of estrone sulfate, a component of the medication, was significantly elevated in serum but was nonsignificantly lower in NAF of subjects receiving HRT (P = 0.069 with adjustment for multiple comparisons). The correlation of serum and NAF estrone sulfate was 0.02. Progesterone remained at essentially zero levels in NAF. The elevated levels of estradiol in NAF were actually associated with a highly significant decrease in the concentration of cathepsin D in NAF and no change in EGF from control levels.
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Progesterone was suppressed to insignificant concentrations in both serum and NAF by OCs: the levels were undetectable after 7 d of OC use and remained undetectable 6 d after the last pill was taken. The combined OCs inhibit both LH and FSH secretion from the pituitary, suppressing follicular development and ovulation (12). The increase in estradiol in NAF observed in the present study after 6 d off OCs indicates that suppression of gonadotropins is lessening by 6 d, but maturation of the follicle and ovulation would have to occur before progesterone levels could increase. In this respect, the concentrations in NAF reflect those in serum. By comparison with the premenopausal control group, OCs completely suppressed progesterone in both serum and NAF and, therefore, no correlation could be calculated for the OC group.
Estradiol concentrations in serum and NAF were lower by similar proportions in OC compared to nonusers but remained detectable in both fluids. However, the correlations between serum and NAF values in the OC and control groups were only 0.22 and 0.25, respectively. The reason for the apparent discrepancy between the proportional decline and the poor correlation in the levels of estradiol in the two fluids is that serum estradiol concentrations fluctuate greatly day by day and hour by hour (13), whereas NAF estradiol concentrations are much more stable (11). In addition, the efficiency of uptake of estrogens may differ with the serum level of the hormone (14). Although it has not yet been proven, the integrated level of serum estradiol over 1 d or more should correlate closely with the levels of estradiol in NAF although the levels within a single blood sample correlate weakly.
Previous studies suggest that a significant proportion of estradiol in NAF is produced locally from androgen and estrogen sulfate precursors (1, 11, 15, 16). The fact that the suppression of estradiol in NAF was almost exactly proportional to that in serum suggests that serum estradiol relates closely to estradiol in NAF. Nevertheless, we have observed that the concentration of estradiol in NAF is highly significantly associated with local concentrations of both androgen and estrone sulfate in NAF (6) of premenopausal women. Obviously, we cannot determine from these experiments the exact proportion of NAF estradiol that is derived from circulating estradiol and how much is derived from conversion of precursors in the breast. Nevertheless, it is apparent that both contribute to the estradiol in the breast.
The proportional decrease in estrone sulfate in NAF attributable to OCs was similar to or greater than that in serum, yet the concentration in NAF was 150-fold greater than in serum. This is evidence for facilitated uptake from serum against a concentration gradient as described by Pizzagalli et al. (17) and for an equilibrium condition between serum and NAF. The same process applies to the uptake of DHEA sulfate, which was more than 500-fold greater in NAF than in serum. Similarly, high concentrations of the sulfonated steroids have been found in breast cysts (18, 19). The pool of estrone sulfate may be converted to estrone by the action of aryl sulfatase and then to estradiol by 17ß-hydroxysteroid dehydrogenase type 1 in the breast epithelium (11, 20). Similarly, DHEA sulfate may serve as an estrogen precursor (21).
It is important to note that the concentrations of EGF and cathepsin D were not less in the NAF of the women on OCs. EGF has been shown to be related to serum levels of estradiol within individuals (22), and cathepsin D has an estrogen response element in the promoter region of the gene, and its synthesis is promoted by estradiol (23, 24). Thus, the levels of these estrogen-responsive proteins in NAF indicate that the ethynyl estradiol in the OCs maintained but did not elevate the levels of these products in the breast. Interestingly, the lack of increases in either cathepsin D or EGF in OC users is consistent with results of numerous epidemiological studies that have failed to detect a consistent elevation of breast cancer risk associated with OC use (25, 26, 27).
Postmenopausal women
Participants in the postmenopausal group were relatively young but had not had a menstrual period for at least 6 months. The mean of the natural log values of NAF progesterone was only 5% of the mean values obtained in the midluteal phase of the menstrual cycles of the premenopausal women. Nevertheless, the number of women with substantially higher values without corresponding increases in serum progesterone suggests that progesterone may be synthesized in the breast under some conditions, particularly in obese women. The frequency of persistently high levels in only one breast also suggests that the synthesis is local. Pregnenolone sulfate is a potential precursor that circulates in the blood of postmenopausal women, and it has been identified in high concentration in breast cyst fluid (28, 29). Determining the factors promoting progesterone biosynthesis in the breast may be important from the point of view of prevention of breast cancer. Sulfatase has been demonstrated in the human breast (11, 16), as has 3ß-hydroxysteroid dehydrogenase (30, 31).
The estradiol levels were less than in the premenopausal group but not significantly so. However, the difference between pre- and postmenopausal women would have been greater by comparison with our earlier study of premenopausal women (11). Using the same methodology, the earlier group of premenopausal women had a mean NAF estradiol that was 3.4-fold greater than that of the premenopausal group of this study. The mean serum estradiol concentration was, however, not higher. The reason for the lower NAF estradiol in this study is not known, but we speculated, as discussed earlier (1), that it may be because the subjects in the present study had a high level of physical activity. It is also possible that although the methodology was the same, there was a difference in the processing of the samples that may explain the lower levels of estradiol in the present study. The apparent decline in NAF estradiol from pre- to postmenopause in either study was only half as great as the decline in serum estradiol, supporting the contention that NAF estrogen is synthesized in the breast of postmenopausal women and not strictly dependent on estradiol in the circulation.
The products of estrogen action, EGF and cathepsin D, in the postmenopausal women were similar to those of the premenopausal women of the present study as well as those of the previous study. Thus, the levels of estradiol in the two studies were not closely associated with levels of EGF and cathepsin D. The concentrations of estrone sulfate were much more closely related to these products, suggesting that it may play an important role in determining the overall estrogenic activity present in NAF. The increase in IL-6 in the postmenopausal women may be a result of the known effect of aging on plasma levels of this cytokine (32). The elevated IL-6 levels may promote the synthesis of estradiol in the breast by activating 3ß-hydroxysteroid dehydrogenase 5,4-isomerase (33) and aromatase (34, 35).
The ICCs of the postmenopausal group are comparable to those of the premenopausal women. The calculation of ICCs in this article takes into account the several covariables of the study and is more accurate than the procedure used previously, and, therefore, the values reported here and in the earlier article differ slightly. In particular, the ICC for estradiol is significantly higher. Overall, the ICCs of both pre- and postmenopausal women are high and indicate that the analytes in NAF represent stable variables that have predictive value.
HRT group
Hormone therapy with a combination of equine-conjugated estrogens and medroxyprogesterone acetate resulted in 2.5 times higher levels of serum estradiol and 18 times higher levels of estradiol in NAF on average than were found in postmenopausal women. Thus, increases in circulating estradiol clearly have an effect on the levels of estradiol within the breast. The relation to serum estrone sulfate is more complex. Serum levels were significantly higher by 3.7-fold, but levels in NAF were nonsignificantly lower by 51%. No reason for this is evident. Salivary estrone sulfate levels in women on HRT were also lower by 26% (Chatterton, R., I. Helenowski, and P. Gann, unpublished results). To establish the significance of the apparent decrease in tissue fluids, a larger number of subjects will be required. Chetrite et al. (36) have shown that progestins inhibit sulfatase activity in the breast cancer cells. If this applies to the normal breast, the presence of medroxyprogesterone acetate in the HRT medication should increase the level of estrone sulfate in NAF by inhibiting its breakdown. This cannot explain the results obtained. We speculate that more of the estrone sulfate is being converted to estradiol in the breast of these women than in the postmenopausal women not taking HRT, thereby lowering the level of estrone sulfate in NAF. Another indication of the increased sulfatase activity is the apparently decreased level of DHEA sulfate and the increase in unconjugated DHEA in NAF of women on HRT.
The levels of EGF were not increased by the high levels of estradiol, and cathepsin D levels were unexplainably low. Although estrogen stimulates the transcription of cathepsin D mRNA, an effect of progesterone has not been demonstrated (37). Nevertheless, it is possible that medroxyprogesterone acetate in the HRT medication acts indirectly to suppress the effect of the elevated estradiol on cathepsin D by inhibiting the expression of the estrogen receptor. This would suggest that EGF is less dependent on estrogen for its expression in the breast than is cathepsin D. However, the number of observations in this group are small, and verification of this finding in a larger study is needed before any conclusions can be drawn.
It is concluded that: 1) potential precursors of estradiol, both androgens and estrone sulfate, remain at comparable levels in the breast after menopause; 2) local synthesis is important for maintenance of estradiol levels in NAF of postmenopausal women but less important for progesterone; 3) local synthesis of progesterone may be related to adiposity; and 4) correlations between serum and NAF concentrations of hormones measured simultaneously are weak but, with sufficient time changes in serum hormone levels, lead to similar changes in NAF. Considered as a whole, demonstration of similarities in response of hormone levels to menopause, OCs, and HRT in serum and NAF validate the physiological relevance of these NAF compounds as biomarkers. At the same time, differences in response such as the continued elevation of estrogens in NAF after menopause point to opportunities to use NAF biomarkers to improve understanding of the hormonal etiology of breast cancer. The relatively low success rate in obtaining NAF, particularly from postmenopausal women, is a limiting factor if analyses of NAF were to be used as a screening procedure. However, Sauter et al. (38) have developed a technique by which they have successfully obtained NAF in essentially all adult women.
|
Footnotes |
|---|
First Published Online November 30, 2004
Abbreviations: BMI, Body mass index; CV, coefficient of variation; DHEA, dehydroepiandrosterone; EGF, epidermal growth factor; HRT, hormone replacement therapy; ICC, intraclass correlation coefficient; NAF, nipple aspirate fluid; OC, oral contraceptive.
Received September 20, 2004.
Accepted November 22, 2004.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
5-4 isomerase type 1 gene transcription in human breast cancer cell lines and in normal mammary epithelial cells by interleukin-4 and interleukin-13. Mol Endocrinol 13:66–81This article has been cited by other articles:
 |
|
 |
|
  Y. Huang, K. E. Anderson, M. Nagamani, J. J. Grady, and L.-J. W. Lu Dietary Intake of Lactose as a Strong Predictor for Secretor Status of Nipple Aspirate Fluid in Healthy Premenopausal Nonlactating Women Clin. Cancer Res., March 1, 2008; 14(5): 1386 - 1392. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. E. Cribb, M. J. Knight, D. Dryer, J. Guernsey, K. Hender, M. Tesch, and T. M. Saleh Role of polymorphic human cytochrome p450 enzymes in estrone oxidation. Cancer Epidemiol. Biomarkers Prev., March 1, 2006; 15(3): 551 - 558. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. S. Dua, C. M. Isacke, and G. P.H. Gui The Intraductal Approach to Breast Cancer Biomarker Discovery J. Clin. Oncol., March 1, 2006; 24(7): 1209 - 1216. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. H. Gann, A. S. Geiger, I. B. Helenowski, E. F. Vonesh, and R. T. Chatterton Estrogen and Progesterone Levels in Nipple Aspirate Fluid of Healthy Premenopausal Women: Relationship to Steroid Precursors and Response Proteins Cancer Epidemiol. Biomarkers Prev., January 1, 2006; 15(1): 39 - 44. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S A Khan, D Bhandare, and R T Chatterton Jr The local hormonal environment and related biomarkers in the normal breast Endocr. Relat. Cancer, September 1, 2005; 12(3): 497 - 510. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
