Fax: (206) 667-5815
Racial differences in the expression of cell cycle–regulatory proteins in breast carcinoma
Article first published online: 28 APR 2004
Copyright © 2004 American Cancer Society
Volume 100, Issue 12, pages 2533–2542, 15 June 2004
How to Cite
Porter, P. L., Lund, M. J., Lin, M. G., Yuan, X., Liff, J. M., Flagg, E. W., Coates, R. J. and Eley, J. W. (2004), Racial differences in the expression of cell cycle–regulatory proteins in breast carcinoma. Cancer, 100: 2533–2542. doi: 10.1002/cncr.20279
This article is a U.S. Government work and, as such, is in the public domain in the United States of America.
The opinions expressed herein do not necessarily reflect the views of the Centers for Disease Control and Prevention, the National Cancer Institute, or the U.S. Government.
- Issue published online: 2 JUN 2004
- Article first published online: 28 APR 2004
- Manuscript Revised: 15 MAR 2004
- Manuscript Accepted: 15 MAR 2004
- Manuscript Received: 30 JAN 2004
- National Cancer Institute. Grant Numbers: RO1 CA64292, RO1 CA71735
- Avon Foundation
- breast carcinoma;
- cell cycle;
- young women
African-American (AA) women are more likely to be diagnosed with an advanced stage of breast carcinoma than are white women. After adjustment for disease stage, many studies indicate that tumors in AA women are more likely than tumors in white women are to exhibit a high level of cell proliferation and features of poor prognosis. The purpose of the current study was to compare tumor characteristics and cell cycle alterations in AA women and white women that might affect the aggressiveness of breast carcinoma.
The study included 124 AA and 397 white women, ages 20–54 years. These women were enrolled in a case–control study in Atlanta, Georgia, between 1990 and 1992. Breast tumor specimens obtained from these women were centrally reviewed for histologic characteristics and evaluated for expression of estrogen and progesterone receptors (ER/PR), c-ErbB-2, Ki-67, p53, cyclin E, cyclin D1, p27, p16, pRb, and p21 by immunohistochemistry. Logistic regression models were used to assess the age- and stage-adjusted associations of various tumor characteristics with race.
The odds of a breast carcinoma diagnosis at a younger age and at a later stage were higher for AA women than for white women. After adjustment for disease stage and age at diagnosis, AA women also were found to have increased odds of having a higher-grade tumor, a higher mitotic index, marked tumor necrosis, ductal histology, loss of ER and PR, overexpression of cyclin E, p16, and p53 and low expression of cyclin D1 at diagnosis.
The observed differences between tumor specimens obtained from AA women and tumor specimens obtained from white women, independent of stage and age at diagnosis, indicated that race may be a determinant, or a surrogate for other determinants, of aggressive breast carcinoma and specific cell cycle defects. Cancer 2004. Published 2004 American Cancer Society.
It is well established that African-American (AA) women are more likely to be diagnosed with advanced-stage breast carcinoma than are white women.1–4 Advanced stage at diagnosis is a major factor contributing to overall decreased survival for AA women with breast carcinoma. However, differences in survival between AA women and white women persist even within stage and treatment groups.3, 4
Many studies have explored the reasons for the increased likelihood of presentation at a later stage among AA women; most have found differences between white and AA women, such as differences in socioeconomic status, cultural beliefs, and access to care, that affect stage at diagnosis.3, 5–7 Less is known about the differences in tumor biology that may contribute to differences in stage at diagnosis or, ultimately, to differences in survival between white women and AA women who present with similar stages of disease.
Most studies, including analyses of data from cancer registries and large breast carcinoma studies with representative inclusion of AA women, have reported marked differences in tumor characteristics between AA women and white women.1, 8 AA women are more likely than white women are to present with high-grade, steroid receptor–negative breast carcinoma, independent of the association between race and disease stage at presentation.4, 9 Higher proliferative activity in the breast tumors of AA women compared with white women has been identified by both flow cytometric S phase measurement and by mitotic counting.2, 4, 9, 10
Increased tumor cell proliferation may contribute to diagnosis at later stages and, possibly, to poorer prognosis for AA women compared with white women. Controlled cell proliferation depends on the orderly expression of regulatory proteins for progression of the cell through the checkpoints of the cell cycle.11, 12 The checkpoint controlling the G1-to-S phase transition of the cell cycle is of particular interest with regard to oncogenesis and is considered not only the primary event controlling progression through another mitotic cycle but also the checkpoint for maintenance of chromosomal integrity. Regulation of the G1 checkpoint is dependent on the sequential expression of cylins D, E, and A and their interactions with the cyclin-dependent kinase inhibitor proteins p27, p21, and p16, which result in inactivation of the retinoblastoma tumor suppressor protein and transcription of genes required for DNA synthesis in S phase.
Data from human breast carcinoma studies show that abnormalities in regulators of the G1/S phase transition are common in breast carcinoma and that these defects are related to poor clinical outcome.13–15 In the current population-based study, we compared the breast tumor characteristics of young AA and white women, including expression of several key G1 cell cycle–regulatory proteins. Histopathologic and biologic tumor characteristics were evaluated by centralized review and assayed in a single laboratory. Analyses were conducted to assess stage- and age-adjusted differences in tumor characteristics by race.
MATERIALS AND METHODS
The patients with breast carcinoma included in the current study were previously enrolled in a population-based case–control study of breast carcinoma in young women conducted between 1990 and 1995.16 The participants were women ages 20–54 years who were residents of a 3-county area in metropolitan Atlanta (Cobb, Fulton, and Dekalb Counties) when they were newly diagnosed with invasive breast carcinoma. Cases were identified through a rapid ascertainment of hospital admission, surgery, and pathology records. Completeness of ascertainment was assessed by periodic checks against the database of the Georgia Center for Cancer Statistics (GCCS), part of the National Cancer Institute (NCI)-funded Surveillance, Epidemiology, and End Results (SEER) program.
The 950 women diagnosed with unilateral invasive disease in the original case–control study who identified themselves as AA or white were eligible for the current study. Data collection for the current study was conducted between 1996 and 2000, after institutional review board approval was obtained. Interviewed participants in the original study were contacted for consent to review medical records and obtain tumor specimens. Pathologic materials from Atlanta hospital pathology laboratories were obtained for 521 women (124 AA and 397 white women) who consented or who had died before the initiation of pathology specimen retrieval in 1998. These 521 women (54.8% of the entire cohort) make up the pathology cohort evaluated in the current study. Requests to hospital and commercial pathology laboratories in the Atlanta area for pathology materials resulted in retrieval of microscopic slides for centralized pathology review for 521 women. In addition, archival tissue specimens appropriate for laboratory analysis were obtained for 479 of the eligible study subjects.
The GCCS registry was the source of data on age, race, SEER stage, histologic grade obtained through SEER, and marital status for the entire group of 950 women who were eligible for the study.17, 18 Follow-up data on breast carcinoma–related mortality was obtained via medical record abstraction and also from the GCCS, which routinely matches all cancer cases with the State of Georgia death tape. Women who no longer resided in Georgia were followed using the National Death Index. Follow-up continued through February 2002. TNM classification and stage groupings were defined by integrating data on tumor size, lymph node status, and distant metastasis abstracted from hospital medical records and from the GCCS registry into the American Joint Committee on Staging (AJCC) staging system that was in place at the time of ascertainment.19 All other tumor characteristics were obtained during the study by centralized pathologic review and laboratory testing.
Pathologic Review and Immunohistochemical Testing
All tumor tissue specimens were reviewed by a pathology expert (P.L.P.), who was blinded to race, other personal characteristics, and clinical outcome, to determine histologic subtype according to the World Health Organization classification of malignant breast tumors and Nottingham tumor grade.20, 21 Mitotic rate was categorized by counting mitotic figures in 10 high-power microscopic fields (HPF). For example, 0–9 mitoses per 10 HPF was considered low, 10–19 mitoses was considered intermediate, and ≥ 20 mitoses was considered high. Tumor necrosis was graded as low (rare necrotic cells), intermediate (areas of necrosis at intermediate magnification), or high (areas of necrosis at low-power magnification).
Representative tumor blocks were selected for testing, and the expression levels of the following proteins were assessed by immunohistochemical (IHC) methods: estrogen receptor (ER; ER1D5; Immunotech, Westbrook, ME),22–24 progesterone receptor (PR; 1A6; Novocastra, Newcastle-Upon-Tyne, United Kingdom),25 p53 tumor suppression gene protein (pAb 1801; Oncogene Science, Manhassett, NY),26–29 Ki-67 proliferation-related antigen (MIB-1; Immunotech),30, 31 c-ErbB-2 (HER-2/neu) oncogene protein (AO485/DAKO15; Dako, Carpinteria, CA),32 and the cell cycle–regulatory proteins cyclin E (cyclin E polyclonal; Dr. James M. Roberts, Fred Hutchinson Cancer Research Center, Seattle, WA),15, 33 cyclin D1 (5D4; Immunotech),34, 35 p16 (Ab-2[ZJ11]; Neomarkers Fremont, CA), p21 (EA10; Calbiochem, San Diego, CA),36 pRb (G3-245; PharMingen, San Diego, CA),37 and p27 (Ab1-DCS-72.F6; Neomarkers).15, 38 Assays were performed using standard IHC techniques, including antigen retrieval when appropriate, on tumor tissue sections using the antibodies specified above.30, 31, 39–41
IHC assays for each antibody were scored according to staining intensity and/or the percentage of tumor cells that were positive, and scores subsequently were collapsed into positive/high and negative/low categories. For ER and PR, any nuclear staining was considered indicative of positivity. For p53, nuclear staining of ≥ 10% of tumor cells was considered indicative of positivity. Antibody staining for c-ErbB-2 was performed using the AO485 antibody from Dako and was initiated before the acceptance of the HercepTest kit (Dako) as the Food and Drug Administration–approved technique for the evaluation of c-ErbB-2 expression.42 The clinical use of this marker to indicate therapy was not the objective of the current study, and as such, we assigned tumor specimens to the positive category if they exhibited ≥ 2+ positive membranous staining relative to normal breast epithelium specimens. Expression levels of cyclin E and p27 were assigned scores ranging from 1 (negative) to 7 (highest intensity), with subsequent grouping as previously described.15
The remaining cell cycle–regulatory proteins (Ki-67, p21, p16, pRb, and cyclin D1) were graded according to the percentage of tumor nuclei with positive staining in four HPF. The average number of tumor cells evaluated for each case was > 1000. The results were categorized into tertiles (excluding zero values) and were considered to be low if there was no positive staining or if the percentage of positive cells was in the lowest tertile and high if the percentage of positive cells was in the middle or highest tertile. All components of the assessment of IHC assays were executed without knowledge of race, other personal characteristics, or clinical outcome.
To assess the extent to which the demographic and SEER-defined clinical and tumor characteristics of the 521 women in the pathology cohort differed from those of the eligible population of 950 women, we compared these two groups with respect to age, race, vital status in 1998, marital status, SEER stage, and histologic grade obtained through SEER (Table 1). This analysis addressed the ability of the study to generalize from the pathology cohort to the eligible population of women with breast carcinoma in the three-county Atlanta area between 1990 and 1992, despite our inability to obtain specimens for all eligible women.
|Characteristic||No. of patients (%)a|
|Total eligible population||Pathology sample|
|All patients||950 (100)||521 (54.8b)|
|Age at diagnosis (yrs)|
|21–34||88 (9.3)||52 (10.0)|
|35–39||136 (14.3)||83 (15.9)|
|40–44||248 (26.1)||126 (24.2)|
|45–49||255 (26.8)||137 (26.3)|
|50–54||223 (23.5)||123 (23.6)|
|White||660 (69.5)||397 (76.2)|
|African American||288 (30.3)||123 (23.6)|
|Unknown||2 (0.2)||1 (0.2)|
|Alive||722 (76.0)||399 (76.6)|
|Dead||228 (24.0)||122 (23.4)|
|Single||157 (16.5)||89 (17.1)|
|Married||583 (61.4)||319 (61.2)|
|Separated/widowed/divorced||185 (19.5)||101 (19.4)|
|Unknown||25 (2.6)||12 (2.3)|
|Local||520 (54.7)||294 (56.4)|
|Regional||380 (40.0)||202 (38.8)|
|Distant||35 (3.7)||19 (3.7)|
|Unknown||15 (1.6)||6 (1.2)|
|1||59 (6.2)||34 (6.5)|
|2||241 (25.4)||144 (27.6)|
|3 or 4||388 (40.8)||218 (41.8)|
|Unknown||262 (27.6)||125 (24.0)|
To determine whether our inability to obtain tumor specimens for 421 of the eligible women affected the relations we observed, we used logistic regression to estimate odds ratios (ORs) and 95% confidence intervals (CIs) and to thereby describe the correlations between race and tumor characteristics in the eligible population compared with the pathology cohort. This comparison indicated that AA women accounted for 30% of the eligible population but only 24% of the pathology cohort. In addition, AA women in the pathology cohort tended to be younger (16% of AA women in the eligible population were ages 21–34 years, compared with 20% in the pathology cohort) and more likely to have died of disease (64% vs. 55%) than were AA women in the overall eligible population. These differences between the eligible patient cohort and the pathology cohort were not observed among white women. Therefore, to reduce the possibility of overestimating the magnitude of tumor differences based on race, we weighted the pathology cohort analyses based on the probabilities that women in the eligible population were included in the pathology cohort. Women were categorized into 20 groups, based on race, age at diagnosis (5 groups), and vital status at the time of contact to obtain tumor consent (1998). For each group, the statistical weighting procedure used the inverse of the probability of inclusion of women in that group.
To assess the effect of weighting, we used logistic regression analysis to estimate both weighted and unweighted ORs for SEER stage at diagnosis. In the eligible population, the odds of having distant disease were 2.3 times greater for AA women than for white women. In an unweighted analysis of the pathology cohort, the odds of having distant disease were 3.3 times greater for AA women compared with white women. However, statistical weighting effectively equalized the odds of distant disease for AA women in the pathology cohort (OR, 2.4) relative to the odds observed in the eligible population. Weighted analyses, therefore, were less likely to overestimate the odds related to race and age, and thus we report weighted percentages and ORs.
To determine, within the pathology cohort, whether AA women were more likely than white women to have tumors with poorer prognostic biologic characteristics or different histologic types, we compared weighted percentage distributions of the characteristics of AA women with those of white women and used P values to determine the statistical significance of the observed differences (Table 2). In addition, we used logistic regression analysis to estimate crude ORs and their associated 95% CIs for racial differences.
|Characteristic||No. of patientsa(%)||P valueb||Unadjusted ORa(95% CI)||Age and stage-adjusted OR (95% CI)a,c|
|I||24 (23.5)||186 (46.2)||0.00||REF|
|IIA||40 (34.1)||101 (24.8)||2.7 (1.5–4.8)|
|IIB||32 (22.5)||61 (15.7)||2.8 (1.5–5.3)|
|III or IV||27 (19.9)||48 (13.3)||3.0 (1.5–5.7)|
|Tumor size (cm)|
|T1a,b (0–1.0)||13 (11.6)||85 (21.4)||0.00||REF|
|T1c (1.1–2.0)||25 (23.7)||155 (38.1)||1.2 (0.6–2.4)|
|T2 (2.1–5.0)||61 (46.2)||114 (28.5)||3.0 (1.5–5.9)|
|T3 (> 5.0)||8 (6.0)||24 (6.7)||1.7 (0.6–4.6)|
|T4d||16 (11.8)||17 (4.8)||4.5 (1.8–11.4)|
|Lymph node status|
|Negative||53 (47.6)||239 (58.8)||0.01||REF|
|1–3 positive||34 (25.9)||78 (19.8)||1.6 (1.0–2.7)|
|4–9 positive||13 (10.1)||38 (9.5)||1.3 (0.6–2.7)|
|10+ positive or N2,3 disease||16 (10.8)||34 (8.6)||1.6 (0.8–3.1)|
|No||114 (93.3)||389 (97.8)||0.01||REF|
|Yes||9 (6.2)||7 (2.0)||3.3 (1.2–9.2)|
|Unknown||1 (0.8)||1 (0.2)||—|
|Low||8 (7.9)||92 (23.2)||0.00||REF||REF|
|Intermediate||40 (32.4)||174 (43.8)||2.2 (1.0–5.0)||2.0 (0.8–4.6)|
|High||76 (59.7)||131 (32.9)||5.3 (2.4–11.8)||4.0 (1.7–9.2)|
|Low||43 (36.6)||239 (60.0)||0.00||REF||REF|
|Intermediate||28 (22.6)||70 (17.6)||2.1 (1.2–3.7)||1.6 (0.8–3.0)|
|High||53 (40.8)||88 (22.4)||3.0 (1.8–4.8)||2.2 (1.3–3.9)|
|None||18 (15.7)||108 (27.5)||0.00||REF||REF|
|Low||50 (38.1)||209 (52.5)||1.3 (0.7–2.3)||1.0 (0.5–2.0)|
|Intermediate||15 (11.2)||28 (7.3)||2.7 (1.2–6.1)||1.9 (0.7–4.9)|
|High||39 (32.9)||49 (12.0)||4.8 (2.5–9.4)||3.5 (1.7–7.6)|
|Positive||35 (30.5)||211 (53.1)||0.00||REF||REF|
|Negative||81 (62.8)||149 (37.4)||2.9 (1.9–4.7)||2.5 (1.5–4.1)|
|Positive||42 (35.4)||238 (60.0)||0.00||REF||REF|
|Negative||74 (57.9)||123 (30.7)||3.2 (2.0–5.0)||2.9 (1.8–4.6)|
|Low||91 (74.9)||343 (86.3)||0.00||REF||REF|
|High||25 (18.4)||17 (4.1)||5.1 (2.6–10.0)||4.3 (2.0–9.2)|
|Low||72 (59.6)||298 (75.1)||0.00||REF||REF|
|High||44 (33.7)||60 (14.9)||2.9 (1.8–4.6)||2.5 (1.5–4.2)|
|Low||73 (59.5)||284 (62.7)||0.00||REF||REF|
|High||42 (32.7)||76 (27.3)||2.0 (1.3–3.2)||1.7 (1.0–2.9)|
|Low||86 (69.2)||215 (53.7)||0.01||REF||REF|
|High||28 (22.9)||137 (34.4)||0.5 (0.3–0.9)||0.5 (0.3–0.8)|
|Low||44 (37.4)||169 (42.3)||0.11||REF||REF|
|High||71 (54.7)||190 (47.9)||1.3 (0.8–2.0)||0.9 (0.5–1.5)|
|Low||71 (58.7)||249 (62.7)||0.11||REF||REF|
|High||45 (34.7)||109 (27.3)||1.4 (0.9–2.1)||1.4 (0.9–2.3)|
|High||39 (31.6)||157 (40.0)||0.07||REF||REF|
|Low||77 (61.7)||203 (50.5)||1.6 (1.0–2.4)||1.5 (0.9–2.4)|
|Low||97 (79.6)||311 (78.1)||0.74||REF||REF|
|High||19 (13.8)||50 (12.6)||1.1 (0.6–2.0)||0.9 (0.4–1.7)|
To determine whether any observed racial differences in prognostic tumor characteristics were due to differences between groups in terms of age at diagnosis or stage distribution, we estimated ORs that were adjusted for these characteristics. Analyses were performed using SAS software (Version 8.2; SAS Institute, Cary, NC) or, for weighted analyses, STATA software (Version 7.0; StataCorp, College Station, TX). The statistical significance threshold was set at alpha = 0.05 (two-tailed) for all analyses.
Differences in Stage and Tumor Characteristics by Race
The characteristics of the 521 women with tumor specimens available for pathology review differed by race (Table 2). Most strikingly, 13% of white women had AJCC Stage III or higher disease, whereas nearly 20% of AA women had this level of advanced disease. The weighted odds that AA women had AJCC Stage III/IV rather than Stage I disease were 3 times (OR = 3.0) the corresponding odds for white women.
In addition to racial differences in stage at diagnosis, many tumor characteristics also were distributed differently among AA and white women (Table 2). For AA women, the odds of having high grade (OR, 5.3), highly mitotic (OR, 3.0), and highly necrotic tumors (OR, 4.8) were higher than the corresponding odds for white women. Steroid receptor status also differed according to race. For example, approximately 63% of AA women had ER-negative tumor specimens, and 58% had PR-negative tumor specimens. In contrast, 37% and 31% of white women had ER-negative and PR-negative tumor specimens, respectively. The proliferation rate measured by Ki-67 IHC analysis did not exhibit the same association with race that was observed when the mitotic rate was measured. However, the odds of abnormal expression of specific regulators of the cell cycle were greater in tumor specimens obtained from AA women. For example, compared with specimens obtained from white women, tumor specimens obtained from AA women had higher levels of cyclin E (OR, 5.1), p16 (OR, 2.9), and p53 (OR, 2.0), as well as lower levels of cyclin D1 (OR, 0.5). There was a trend toward higher levels of p21 and lower levels of p27 among tumor specimens obtained from AA women compared with those obtained from white women, but racial differences in the expression of these two cell cycle–inhibitory proteins did not reach statistical significance. Racial differences in c-ErbB-2 expression were not observed.
Statistical adjustment for age and disease stage at diagnosis reduced the magnitude of the racial differences in terms of tumor-related prognostic factors (Table 2). For example, the unadjusted OR for high histologic grade was 5.3, whereas the adjusted OR was 4.0. This change in ORs reflected the known association of many of the tumor characteristics with age and disease stage and the relations between race and age and race and stage. However, adjustment did not eliminate the racial differences in the distributions of the tumor characteristics. Even after adjustment for age and disease stage, AA women had greater odds of having higher-grade (OR, 4.0), highly mitotic (OR, 2.2), highly necrotic (OR, 3.5), ER-negative (OR, 2.5), and PR-negative (OR, 2.9) tumor specimens compared with white women. The odds of having high levels of cyclin E (OR, 4.3), p16 (OR, 2.5), and p53 (OR, 1.7) also remained greater for AA women than for white women. The odds of having low cyclin D1 levels in tumor specimens for AA women were unaffected by adjustment for age and stage (unadjusted OR, 0.5; age- and stage-adjusted OR, 0.5).
Ductal carcinoma was the most common type of tumor for both AA and white women (Table 3). The number of women with other less common subtypes of breast carcinoma was small, but there were some racial differences in the distribution of subtypes. A greater percentage of white women were diagnosed with lobular malignancies than were AA women (6.5% vs. 0.8%). Using ductal carcinoma as a reference group in regression analyses (data not shown), the weighted odds of having lobular tumors were more than 8 times higher for white women compared with AA women (OR, 0.1; 95% CI, 0.0–0.8).
|Histology||No. of patients (%)|
|Ductal carcinoma NOS||108 (87.1)||333 (83.9)|
|Tubular||1 (0.8)||15 (3.8)|
|Mucinous||3 (2.4)||5 (1.2)|
|Medullary/atypical medullary||8 (6.5)||9 (2.3)|
|Lobular carcinoma||1 (0.8)||26 (6.5)|
|Mixed ductal and lobular||2 (1.6)||7 (1.8)|
|Inflammatory||1 (0.8)||0 (0.0)|
|Other||0 (0.0)||2 (0.5)|
In the current study, we compared the clinical and breast tumor characteristics of AA women and white women age ≤ 54 years who had previously been enrolled in a case–control study of breast carcinoma in Atlanta. As has been well established by other population studies and analyses, the odds of being diagnosed at an advanced clinical stage were greater for AA women than for white women.1–3
Our data support other studies that have examined racial differences in breast carcinoma at presentation and found that the tumor specimens obtained from AA women, compared with those obtained from white women, were more likely to exhibit features of aggressive behavior, such as high S phase fraction, loss of steroid receptors, high histologic grade, nuclear atypia, and the presence of tumor necrosis.2–4, 9, 43–47 However, not all of these other studies controlled for the effect of stage on tumor characteristics. In the NCI Black/White Cancer Survival Study (BWCSS), which adjusted for age, stage, and geographic region, tumor specimens obtained from AA women were significantly more likely than those obtained from white women to have high nuclear atypia, high mitotic activity, higher histologic grade, and a greater degree of tumor necrosis.9 In addition, AA women who presented with Stage II, lymph node–positive disease or with Stage III/IV disease had tumor specimens with higher nuclear and histologic grades than did white women with similar stages of disease.5, 45 Elmore et al.43 found that tumors obtained from AA women in Connecticut were larger and more likely to exhibit marked necrosis compared with tumors obtained from white women after adjustment for age, income, method of detection, and insurance.43
To our knowledge, the current study is the largest to evaluate racial differences in ER and PR expression in breast carcinoma using data consistently assayed in a single laboratory. Analyses of SEER breast carcinoma data have shown an increased likelihood of ER-negative disease in AA women compared with white women among women of all ages (39% for AA women vs. 22% for white women) and among women age < 35 years (39% for AA women vs. 31% for white women).1, 2 However, these analyses did not adjust for stage at diagnosis and may have overestimated racial differences in steroid receptor expression.
Although some studies have failed to find significant differences in steroid receptor status, many are underpowered to address this issue because of frequent missing data on ER and PR status. Elmore et al.43 found no statistical difference in ER expression between AA women and white women identified through the Connecticut tumor registry but did find that AA women were less likely than white women were to have PR-positive tumors after adjustment for age (OR, 0.48; 95% CI, 0.25–0.93). However, ER and PR data were only available for approximately one-half of the 400 eligible participants in their study. Ownby et al.46 failed to find a significant racial difference in ER status (determined biochemically) among a large cohort of women with breast carcinoma in Detroit, Michigan, although that cohort (n = 1078) included only 73 AA women. In the BWCSS, which used ER status assessments from multiple laboratories, ER-positive tumors were less common in AA women, although the difference did not possess statistical significance.9
Very limited data regarding interracial differences in other prognostic tumor markers are available. The largest study to address molecular markers of prognosis by race is that of Elledge et al.4 In their study, no difference in either c-ErbB-2 or p53 status was identified. In another hospital-based study of 204 patients with breast carcinoma, including 32 AA women, radioimmunoassay measurements also failed to reveal an association between race and the expression of c-ErbB-2.10
We report the first comprehensive evaluation, to our knowledge, of the racial distributions of specific abnormalities in G1/S phase cell cycle–regulatory proteins in tumor specimens obtained from AA and white women who were enrolled in a population-based study. Abnormalities in the expression of G1 regulatory proteins that were evaluated in the current study are common in breast carcinoma, and there is growing evidence that specific abnormalities in the expression of cell cycle components are related to breast carcinoma behavior or response to therapy.15, 48–55 Keyomarsi et al.13 reported a striking 13-fold increase in mortality rate for women whose tumor specimens exhibited high levels of full-length cyclin E as measured by Western blot. We and others have found similar, albeit reduced, associations between high levels of IHC-detected cyclin E and poor outcome for women with breast carcinoma.15, 56, 57 In most studies, cyclin D1 is strongly associated with positive ER status and variably associated with survival for women with breast carcinoma. The largest study to address the prognostic value of cyclin D1 (n = 345) found that high levels of cyclin D1 were associated with improved survival.58, 59
The developing data concerning tumor cell expression of the two primary G1 cyclins suggest that tumor specimens with increased levels of either cyclin E or D1 may reflect disruption of alternate pathways, both of which result in unrestrained cell proliferation.56, 60. In a multiparameter analysis of 120 breast carcinoma specimens, Landberg56 found that high levels of cyclin D1 expression in tumor specimens were associated with ER-positive tumors in older women, whereas high levels of cyclin E expression were associated with aneuploid, ER-negative tumors in younger women. Among 100 women with lymph node–negative breast carcinoma, women with tumor specimens exhibiting low cyclin E levels and high cyclin D levels were significantly less likely to die of disease at 6 years than were women with tumor specimens with high levels of cyclin E and low levels of cyclin D1 (0% vs. 50%).48 Expression of the cell cycle–inhibitory protein p16 also differed between AA women and white women in the current study. The tumor suppressor gene p16 is a key component of the cyclin D/retinoblastoma (Rb) pathway and is frequently altered by mutation, methylation, or loss in human malignancy.52, 61–63 Loss of p16 function due to gene promoter hypermethylation has been associated with poor prognosis in breast carcinoma.64 Overexpression of p16 is an important indicator of Rb pathway dysfunction as well. For example, high levels of p16/p16 protein and mRNA have been associated with an aggressive breast carcinoma phenotype, poor clinical outcome, and poor outcome in patients treated with specific chemotherapy regimens, suggesting that altered expression can affect response to therapy.14, 53, 65, 66 It remains to be determined whether the higher expression of p16 in tumor specimens obtained from AA women, as was reported in the current study, is related to differences in survival.
The use of centralized pathologic review and testing in the current, relatively large study adds to the strength of our findings. However, there are limitations as well. Most notably, we were unable to analyze tumor specimens from all women in the cohort. The population-based design allowed us to assess how well the tested group reflected the entire cohort, and we identified differences in race, age, and vital status in the tested group. Weighted analyses uniformly, and we believe appropriately, decreased the magnitude of the odds associated with racial differences in tumor characteristics. For SEER stage, a variable for which we had complete data on all eligible participants, weighting effectively equalized the odds ratios for the pathology cohort to those for the eligible population. It is possible that additional bias in the composition of the pathology cohort could account for the differences that we identified. However, the magnitude of the odds associated with specific tumor characteristics is greater than the magnitude that would be expected to result from bias alone, and the differences and similarities that we identified in the distribution of prognostic factors are consistent with those previously reported in the literature.
Differences in survival between AA and white women with breast carcinoma are undoubtedly influenced by racial differences in terms of a range of socioeconomic and cultural factors that affect screening, diagnosis, treatment services, and survival.67–69 However, the current study adds to growing evidence that, for reasons that have not yet been determined, compared with white women, AA women tend to have more aggressive breast carcinoma phenotypes, even after controlling for disease stage at diagnosis. One possible explanation for this association between race and disease characteristics is that variation in the distribution of breast carcinoma risk factors, including reproductive experiences. These factors may affect the selective pressure on breast epithelial cells during early tumor development and result in tumors with alterations in specific pathways. Variability in the distribution of risk factors by race has been proposed as a reason for racial differences in the incidence of breast carcinoma.70–73. AA women age < 50 years have a higher incidence of breast carcinoma than do white women, whereas after age 50, the incidence rates are higher for white women.73, 74 One of the factors suggested as a reason for the higher incidence of breast carcinoma among AA women age < 45 years is the paradoxic high short-term risk conferred by early pregnancy.72 In support of this hypothesis, breast tumors have been reported to exhibit increased proliferation in the short term following pregnancy.75 Recent studies relating risk factors to tumor characteristics have also linked cyclin D1 overexpression with oral contraceptive use and breastfeeding for ≥ 1 year with a decreased chance of c-ErbB-2-positive breast carcinoma.76, 77 Further analysis of data from this cohort of young AA and white women with breast carcinoma will allow us to evaluate the associations of various risk factors with breast tumor characteristics.
In summary, the current study revealed differences in histologic appearance, steroid receptor status, and cell cycle regulator expression between tumors obtained from AA women and those obtained from white women. These differences may ultimately contribute to differences in breast carcinoma survival. Future research aimed at determining the predictors and consequences of cell cycle disruption in cancer cells must include sufficient numbers of AA women to allow us to understand the effect of cell cycle defects on survival and treatment response for women of both races.
The authors thank Stephanie Stafford and Ann Yoder for providing database support and Kelly Wirtala and Judy Schmidt for their technical contributions.
- 17YoungJ, RoffersS, RiesL, FritzA, HurlbutA, editors. SEER summary staging manual—2000: codes and coding instructions. Bethesda: National Cancer Institute, 2001.
- 18FritzA, RiesL, editors. The SEER program code manual (3rd edition). Bethesda: National Cancer Institute, 1998.
- 19FlemingID, CooperJS, HensenDE, et al., editors. AJCC cancer staging manual (5th edition). Philadelphia: Lippincott-Raven, 1998.
- 21TavassoliFA, DevilleP, editors. Pathology and genetics of tumours of the breast and female genital organs. Oxford: Oxford University Press, 2003.
- 39Strategies for improving the immunohistochemical staining of various intranuclear prognostic markers in formalin-paraffin sections: androgen receptor, estrogen receptor, progesterone receptor, p53 protein, proliferating cell nuclear antigen and Ki-67 antigen revealed by antigen retrieval techniques. Hum Pathol. 1994; 25: 263–270., , , , , .
- 69African-American ethnicity, socioeconomic status, and breast cancer survival. A meta-analysis of 14 studies involving over 10,000 African-American and 40,000 white American patients with carcinoma of the breast. Cancer. 2002; 94: 2844–2854., , , et al.
- 73RiesL, KosaryC, HankeyB, MillerB, CleggL, EdwardsB, editors. SEER cancer statistics review, 1973-1996. Bethesda: National Cancer Institute, 1999.