Estrogen and Progesterone Receptors in Colon Tumors

Anatomic Pathology / ESTROGEN AND PROGESTERONE RECEPTORS IN COLON TUMORS

Estrogen and Progesterone Receptors in Colon Tumors Martha L. Slattery, PhD, MPH,1 Wade S. Samowitz, MD,2 and Joseph A. Holden, MD2 Key Words: Adenocarcinoma; Colon cancer; Estrogen receptors; Hormone replacement therapy; Progesterone receptors

Abstract Existing data suggest that there is a hormonal basis to the cause of colon cancer. In the present study, we evaluated the presence of estrogen receptors (ERs) and progesterone receptors (PRs) in colonic tumors from 156 women diagnosed with colon cancer in Utah from September 1991 through September 1994. Immunohistochemical staining with antibodies to ERs and PRs was performed on histologic sections prepared from paraffin blocks. None of the tumors were considered ER-positive; 1 tumor was PR-positive. Use of hormone replacement therapy was not associated with PR-positive tumors. These data do not support previous reports that suggest that colon tumors frequently have receptors for estrogen, progesterone, or both.

A hormonal basis to the cause of colon cancer is supported by many factors: age-adjusted incidence rates for colon cancer are lower in women than in men1; subgroups of the population with low parity, such as nuns, have higher rates of colon cancer than populations in which parity is high2; and epidemiologic studies suggest that using hormone replacement therapy (HRT) decreases the risk of developing colon cancer.3 Furthermore, estrogen receptors (ERs) and progesterone receptors (PRs) have been found in normal and malignant colon tissues.4 Most studies evaluating ERs and PRs in colonic tissue have included few samples, and results are inconclusive for the prevalence of ER- and PR-positive colon tumors. Issa and colleagues5 have shown that the promoter region of the ER gene is silenced in colon tumors. However, whether HRT alters tumor ER or PR status is unknown. In the present study, we evaluated ER and PR status in tumors and normal tissue to determine the prevalence of ER- and PR-positive tumors and their association with HRT, reproductive, and demographic factors.

Methods Data for this study come from a population-based casecontrol study of colon cancer originally designed to assess associations between diet, activity, and other lifestyle factors as they relate to the cause of colon cancer. Utah women diagnosed with a first primary adenocarcinoma of the colon from September 1991 through September 1994 were included in the present study. Study eligibility included age between 30 and 79 years at the time of diagnosis; white, black or Hispanic; mentally competent to complete the interview; and living in 1 of 8 Utah counties (Salt Lake, Davis, Wasatch, 364

Am J Clin Pathol 2000;113:364-368

© American Society of Clinical Pathologists

Anatomic Pathology / ORIGINAL ARTICLE

Utah, Weber, Tooele, Morgan, and Summit). Cases were identified by using a rapid-reporting system in which study abstractors visited hospital pathology laboratories on a monthly basis; complete case ascertainment was verified by using data from the Utah Cancer Registry. The average time between diagnosis and identification was 27 days. Of all the women identified with colon cancer, 61.9% participated. Primary reasons for nonparticipation were death between identification and interview (8.5%); physician refusal (4.4%); too ill to participate (2.6%); unable to locate or moved (1.5%); and refusal (20.7%). A total of 159 women with valid interview data were targeted for the present study. Data Ascertainment Data were collected by trained and certified interviewers using laptop computers. The referent year for the study was the year 2 years before diagnosis. Data collected included detailed diet, activity, and reproductive history along with information on body size, family history of cancer, demographic information, and other general lifestyle questions. Of interest to the present study were data collected as part of the reproductive history, including number of pregnancies (including live births, miscarriages, and stillbirths); use of HRT (if HRT had been used, year stopped, age started); and usual adult height and weight during the referent year.6 Data from the Utah Cancer Registry were used to classify tumor site and stage at time of diagnosis. Estrogen Receptors and Progesterone Receptors Formalin-fixed paraffin-embedded colonic adenocarcinoma tissue was available for 156 of the 159 eligible women. Immunohistochemical staining was performed as in previous studies.7 Slides prepared from paraffin blocks were deparaffinized and heated in a 10-mmol/L concentration of sodium citrate (pH 6.0) in a microwave oven for 30 minutes. After cooling, immunohistochemical staining was performed with the use of an automated immunohistochemical stainer (DAKO Autostainer, DAKO, Carpinteria, CA) in accord with the manufacturer’s instructions. Primary antibodies were DAKO N-series mouse antihuman estrogen receptor 1D5 and progesterone receptor PR1A6. Prediluted antibodies were used. Detection was performed using the DAKO Envision detection kit, which uses a secondary mouse anti-immunoglobulin conjugated to a horseradish peroxidase–linked polymer. Color development was accomplished with diaminobenzidine as the chromogen. Positive and negative controls were run with each set of tumors and checked to verify ER and PR staining. Tumors were scored as positive for ER or PR if 20% or more of the nuclei were stained by the antibody. Tumors were scored as negative if less than 5% of nuclei stained and as borderline if 5% to 19% of the nuclei were stained. These are the cutoff © American Society of Clinical Pathologists

points used at the University of Utah Health Science Center for clinical evaluation of ER and PR in breast cancers. Slides were immunostained with vimentin8 on adjacent sections to verify antigen preservation. If vimentin staining was patchy, ER and PR results were evaluated in the areas of the tissue in which vimentin positivity was seen. All slides were read by the same pathologist (W.S.S.), and tumors read as borderline or positive were read by a second pathologist (J.A.H.) who also reviewed cases the first pathologist found difficult to interpret because of excessive nonspecific (cytoplasmic) staining or because of percentages of positive nuclei slightly less than borderline. The population is described in terms of use of HRT, reproductive history, and tumor characteristics. Tumors in the proximal colon are those located in the cecum through the transverse colon; distal tumors are those in the splenic flexure through the sigmoid colon. Surveillance Epidemiology and End Results summary stage codes (ie, local, regional, and distant) were used to define tumor stage.

Results The population is described in ❚Table 1❚. Almost 40% of the population was 70 years or older at the time of diagnosis. Tumors were located equally in the proximal and distal segments of the colon. More than half of the women had 4 or more pregnancies, and approximately half of the women reported having taken HRT at some time. All tumors evaluated were ER-negative. One tumor was PR-positive; the woman had taken HRT for 6 years. Five tumors were considered PR-negative by one pathologist and borderline positive by the other, while 2 other tumors were considered borderline PR positive by one pathologist and negative by the other. Of these 7 tumors, 2 of the patients had taken HRT and 5 had not. The proportion of PR-positive tumors was not significantly different among those who took HRT and those who did not (P = .28, Fisher exact test). Normal tissue was present for 89 (57.0%) of the 156 tumors evaluated; the normal tissue was also categorized as ER- and PR-negative, although borderline PR positivity was detected from 2 of these normal tissue samples. The corresponding tumors from these samples were classified as borderline PRpositive and PR-positive.

Discussion Data from the present population-based study suggest that neither normal nor tumor tissue from the colon are positive for ERs and that the prevalence of positive PR in colonic tissue is rare. These results agree with the findings of some Am J Clin Pathol 2000;113:364-368

365

Slattery et al / ESTROGEN AND PROGESTERONE RECEPTORS IN COLON TUMORS

❚Table 1❚ Characteristics of 156 Utah Women Diagnosed With Colon Cancer From September 1991 Through September 1994* N (%) Age (y) 30-39 40-49 50-59 60-69 70-79 Education High school or less Some college College graduate Tumor site Proximal Distal Unknown Stage of tumor at diagnosis (n = 127) Local Regional Distant Number of pregnancies None 1-3 4-5 >5 Menopausal status (n = 154) Premenopausal Postmenopausal Use of hormone replacement therapy in postmenopausal women (n = 140) Ever Never

6 (3.8) 9 (5.8) 27 (17.3) 53 (34.0) 61 (39.1) 89 (57.0) 45 (28.8) 22 (14.1) 72 (46.2) 74 (47.4) 10 (6.4) 44 (34.6) 63 (49.6) 20 (15.7) 7 (4.5) 55 (35.2) 53 (34.0) 41 (26.3) 13 (8.4) 141 (91.6) 70 (50.0) 70 (50.0)

* For some characteristics, data for fewer than 156 women were available; numbers are given with those variables.

studies that essentially show no ER or PR positivity,9-12 although other studies suggest that 20% to 54% of colon tumors are positive for ER4,13-16 and as many as 42.8% of colon tumors are positive for PR. 13 Studies also have suggested that people with ER receptors in normal tissue have improved survival after diagnosis.15 While studies of parity and HRT show stronger associations for proximal tumors,3 studies that have evaluated ER and PR show that the percentage of ER- or PR-positive tumors are not different by tumor location or other tumor characteristics.4,15 Differences in findings could stem from several sources, including source of tissue, methods of detecting ER and PR, and level of staining at which ERs and PRs are declared positive. Many studies have relied on fresh tissue,4,10,13,14,16,17 while in the present study, tissue was obtained from paraffin blocks. A limitation of the present study is our lack of control on how colon tissue was handled at the time of fixation. Since cases, and therefore tissue, were collected from 15 different hospitals, we could not implement a standard method of fixation. This could contribute to decreased levels of ER and PR positivity. To deal with this potential problem, vimentin antibody stains were done along with ER and PR antibodies to assess antigen preservation after 366

Am J Clin Pathol 2000;113:364-368

fixation.8 Although vimentin staining was variable, we were able to detect at least focal vimentin staining in almost all sections, and, therefore, we believe that our lack of ER or PR detection was not a result of antigen degradation during fixation or embedding. While vimentin staining has been reported to be an effective means to evaluate ER and PR antigen preservation, we cannot exclude the possibility that the antigens could be more easily degradable than vimentin under our processing conditions. However, it should be noted that our results using paraffin-embedded tissue are identical to those obtained by Cameron and colleagues, 11 who found no ER-positive tumors using frozen colonic tissue and the same categories to determine positivity. Different techniques to detect ER and PR also have been used by different studies. For instance, many of the early studies that report detection of ER or PR in colon tumors used dextran-coated charcoal-competitive proteinbinding assays.12,16 Others have used the ER-D5 antibody and detected reactivity only in the cytoplasm.17 The antibodies used in the present study are used on a clinical basis to detect ER and PR in breast tissue18 and have been shown to be effective for detecting ER- and PR-positive tumors using paraffin-fixed blocks.19 It is possible that using other © American Society of Clinical Pathologists

Anatomic Pathology / ORIGINAL ARTICLE

antibodies or techniques might yield different results, but it should be noted that the antibodies used in our study are used on a clinical basis to guide treatment decisions. Unlike breast cancer studies, most studies of colon tissue have reported low levels of ER or PR positivity, usually less than 10 fmol/mg protein.4,13,20 These levels have been referred to in some articles as very low.11,14 Others have reported actual levels whereby tissue was considered positive for ER with levels of less than 3 fmol/mg of cytosolic protein,10 1 fmol/mg of cytosolic protein,15 or the mean level of positive was 1.3 fmol/mg of protein for ER and 3.9 fmol/mg of protein for PR. It is possible that such low levels may be undetectable by immunohistochemistry or could be associated with nuclear staining less than 5%, our threshold for a negative stain. Optimal overall breast cancer survival has been shown to be predicted with a 20-fmol threshold,21 although this level may not be most relevant to colon cancer. Variation in the literature in terms of the prevalence of ERor PR-positive tumors could stem from individual variability among researchers in how they classify positivity. 10 However, in the present study, we detected no tumors or normal tissue with more than 5% positivity for ER and very few with greater than 5% positivity for PR. The literature is fairly consistent that if colonic tissue or colonic tumor tissue is ER- or PR-positive, the level of positivity is very low. Issa et al5 reported methylation of the cytosine phosphate guanine (CpG) island of the ER gene in colon tumors and also in colonic tissue with increasing age. Methylation of the CpG island results in diminished or absent gene expression and deregulated cell growth. Consistent with the finding by Issa et al,5 in the present study, primarily of older women, we did not detect significant levels of ER or PR in normal colonic tissue or in tumor tissue. While detection of ER or PR in colonic tumors would have implications for biologic pathways to disease, its implications for treatment or prognostic factors are less clear. Use of tamoxifen and other related drugs is a viable treatment for women diagnosed with breast cancer when ER-positive tumors are present. Use of similar drugs as a treatment for colon cancer is questionable since studies have shown that women using tamoxifen actually may be at increased risk of developing colon cancer.22,23 In this group of women diagnosed with colon cancer, we detected no ER-positive tumors and only 1 PR-positive tumor. Use of HRT did not seem to influence ER or PR positivity. Although our results agree with those of some studies and are slightly different than those of others, there is not strong support from any of the studies that colon tumors have a high degree of ER or PR positivity. From the Departments of 1Family and Preventive Medicine and 2Pathology, University of Utah, Salt Lake City.


This study was supported through grants CA48998 (Dr Slattery) and National Cancer Institute (Bethesda, MD) contract NO1-CN-6700 (Utah Cancer Registry). Address reprint requests to Dr Slattery: Health Research Center, 391 Chipeta Way, Suite G, Salt Lake City, UT 84108. Acknowledgments: We acknowledge the contributions of Sandie Edwards, MS, Leslie Palmer, and Khe-Ni Ma, MS, to this project.

References 1. Wingo PA, Ries LAG, Rosenberg HM, et al. Cancer incidence and mortality, 1973-1995. Cancer. 1998;82:11971207. 2. Potter JD, Slattery ML, Bostick RM, et al. Colon cancer: a review of the epidemiology. Epidemiol Rev. 1993;15:499-545. 3. Hebert-Croteau N. A meta-analysis of hormone replacement therapy and colon cancer in women. Cancer Epidemiol Biomarkers Prev. 1998;653-660. 4. Meggouh F, Lointier P, Saez S. Sex steroid and 1,25dihydroxyvitamin D3 receptors in human colorectal adenocarcinoma and normal mucosa. Cancer Res. 1991;51:1227-1233. 5. Issa JPJ, Ottaviano YL, Celano P, et al. Methylation of the estrogen receptor CpG island links aging and neoplasia in human colon. Nat Genet. 1994;7:536-540. 6. Edwards SL, Slattery ML, Mori M, et al. Objective system for interviewer performance evaluation for use in epidemiologic studies. Am J Epidemiol. 1994;140:1020-1028. 7. Lynch BJ, Holden JA, Buys SA, et al. Pathobiologic characteristics of hereditary breast cancer. Hum Pathol. 1998;29:1140-1144. 8. Battifora H. Assessment of antigen damage in immunohistochemistry: the vimentin internal control. Am J Clin Pathol. 1991;96:669-671. 9. Dawson PM, Shousha S, Blair SD, et al. Oestrogen receptors in colorectal carcinoma. J Clin Pathol. 1990;43:149-151. 10. Sciascia C, Olivero G, Comandone A, et al. Estrogen receptors in colorectal adenocarcinomas and in other large bowel diseases. Int J Biol Markers. 1990;5:38-42. 11. Cameron BL, Butler JA, Rutgers J, et al. Immunohistochemical determination of the estrogen receptor content of gastrointestinal adenocarcinomas. Am Surg. 1992;58:758-760. 12. Deamant FD, Pombo MT, Battifora H. Estrogen receptor immunohistochemistry as a predictor of site of origin in metastatic breast cancer. Appl Immunohistochem. 1993;1: 188-192. 13. Geelhoed GW, Alford C, Lippman ME. Biologic implications of steroid hormone receptors in cancers of the colon. South Med J. 1985;78:252-254. 14. Bracali G, Caracino AM, Rossodivita F, et al. Estrogen and progesterone receptors in human colorectal tumour cells (study of 70 cases). Int J Biol Markers. 1988;3:41-48. 15. DiLeo A, Messa C, Russo F, et al. Prognostic value of cytosolic estrogen receptors in human colorectal carcinoma and surrounding mucosa: preliminary results. Dig Dis Sci. 1994;39:2038-2042. 16. McClendon JE, Appleby D, Claudon DB, et al. Colonic neoplasms: tissue estrogen receptor and carcinoembryonic antigen. Arch Surg. 1977;112:240-241.

Am J Clin Pathol 2000;113:364-368

367

Slattery et al / ESTROGEN AND PROGESTERONE RECEPTORS IN COLON TUMORS

17. Alford TC, Do HY, Geelhoed GW, et al. Steroid hormone receptors in human colon cancers. Cancer. 1979;43:980-984. 18. Takeda H, Yamakawa M, Takahashi T, et al. An immunohistochemical study with an estrogen receptor–related protein (ER-D5) in human colorectal cancer. Cancer. 1992;69: 907-912. 19. Veronese SM, Barbareschi M, Morelli L, et al. Predictive value of ER1D5 antibody immunostaining in breast cancer. Appl Immunohistochem. 1995;3:85-90. 20. Singh S, Sheppard MC, Langman MJS. Sex differences in the incidence of colorectal cancer: an exploration of oestrogen and progesterone receptors. Gut. 1993;34:611-615.

368

Am J Clin Pathol 2000;113:364-368

21. Battifora H, Mehta P, Ahn C, et al. Estrogen receptor immunohistochemical assay in paraffin-embedded tissue: a better gold standard? Appl Immunohistochem. 1993;1:39-45. 22. Persson I, Yuen J, Bergkvist L, et al. Cancer incidence and mortality in women receiving estrogen and estrogen-progestin replacement therapy: long-term follow-up of a Swedish cohort. Int J Cancer. 1996;67:327-332. 23. Rutqvist LE, Johansson H, Signomklao T, et al. Adjuvant tamoxifen therapy for early stage breast cancer and second primary malignancies. J Natl Cancer Inst. 1995;87:645-651.
