Background
Although the majority of data generated from breast cancer research has come from studies using Caucasian women (CW) as subjects, it is becoming increasingly clear that the incidence, mortality, and length of survival after treatment for breast cancer vary greatly among different ethnic groups. Although overall incidence of breast cancer in the United States is higher for CW (125.4/100,000) than for African American women (AAW) (116.4/100,000) [
1], breast cancer incidence is higher in young AAW compared to CW such that 30-40% of AAW with breast cancer are under age 50 when diagnosed compared to just 20% of CW [
2]. In addition, the five-year survival rate for AAW (77%) is significantly lower than for CW (90%) [
3] across all ages and tumor stages and subtypes, and the age-adjusted mortality rate for AAW (32.4/100,000) is the highest rate for any ethnic group studied [
1].
Triple negative breast cancer (TNBC) is defined as tumors that do not express the estrogen or progesterone receptors or HER2. TNBC is an aggressive tumor phenotype, characterized by diagnosis at a younger age, high-tumor grade, larger mean tumor size, and higher rates of mortality compared to other tumor subtypes [
4]. Several clinical trials are underway testing targeted agents, such as PARP, angiogenesis and EGFR inhibitors; however, to date cytotoxic therapy remains the standard treatment for patients with TNBC. TNBC is diagnosed significantly more frequently in premenopausal AAW (39%) compared to either postmenopausal AAW (14%) or in non-African Americans of any age (16%) [
5]. This higher prevalence in young AAW coupled with higher mortality rates and lack of available targeted treatments provides an explanation, at least in part, for the less favorable outcomes of AAW with breast cancer [
6].
A number of epidemiological risk factors have been associated with TNBC including reproductive factors such as younger ages at menarche and at first full-term pregnancy (FFTP), higher parity, and shorter (or lack of) duration of breastfeeding, as well as anthropometric factors such as higher body mass index (BMI) and waist-to-hip ratio [
7]. In addition, gene expression differences have been detected in primary breast tumors between AAW and CW [
8,
9], although these studies were not limited to TNBC but included a range of tumor subtypes. Identification of both epidemiological and molecular factors that differ between AAW and CW with TNBC is critical to developing more effective risk reduction strategies as well as treatment options for AAW. To this end, differences in both a range of epidemiological factors including obesity, estrogen exposure, breastfeeding, diet and physical activity, and co-morbidities, as well as gene expression profiles were evaluated between AAW and CW with TNBC.
Discussion
To decrease survival disparities between AAW and CW with breast cancer, the source of outcome differences must be identified. Higher mortality rates have been detected for AAW in both the general population and the military when breast cancer was considered as a single disease [
3,
18], however, breast cancer is heterogeneous, with an array of phenotypic and molecular differences. Given the higher frequency of TNBC in AAW, higher mortality rates in AAW compared to CW with TNBC may explain outcome disparities between populations.
Data generated here do not support TNBC as a more aggressive disease in AAW. Mortality rates and length of disease-free survival did not differ significantly between populations. These results are supported by data from the Carolina Breast Cancer Study (CBCS) that demonstrated that while AAW had overall higher breast cancer mortality rates, when only patients with TNBC were considered, mortality rates did not differ significantly [
19]. In addition, a recent study conducted at a single institution with similar treatment and follow-up between populations also failed to find differences in disease-free or overall survival between AAW and CW with TNBC [
20]. Together, these data do not support TNBC as a clinically more aggressive tumor type in AAW compared to CW.
In conjunction with the inability to detect outcome differences between groups, TNBC tumors from AAW and CW were molecularly similar, with PCA failing to separate gene expression patterns by population. One gene,
CRYBB2P1, was expressed at significantly higher levels in tumors from AAW compared to CW.
CRYBB2P1 has significant sequence similarity to
crystallin, beta B2, a member of the crystallin gene family that encodes the major structural components of the vertebrate eye lens, however, CRYBB2P1 has been designated a pseudogene, and to date, the possible function of
CRYBB2P1 transcripts are unknown [
21]. Higher expression of the probe for
CRYBB2P1 has been detected in a number of tissues from African Americans, including breast (of mixed subtypes), prostate and colorectal tumors, disease-free breast and prostate tissues [
8,
9,
22,
23] as well as blood endothelial cells [
24]. Given the differential expression of this pseudogene in a variety of tissues, both malignant and non-malignant, additional studies must be performed to determine whether
CRYBB2P1 plays a causative role in tumorigenesis or reflects population stratification.
Although outcome disparities were not detected in this population, diagnosis of TNBC was significantly higher in AAW (28%) compared to CW (12%). Thus, identification of risk factors, both modifiable and non-modifiable, leading to the higher frequency of TNBC in AAW may reduce survival disparities by preventing the development of TNBC. For example, a SNP on chromosome 5p15 near the TERT locus was associated with TNBC in a mixed population of patients of African and European ancestries [
25]; data from the Black Women’s Health Study (BWHS) confirmed this association and found that SNP rs8170 in the BABAM1 gene, was associated with increased risk of TNBC in an African American population [
26]. A higher prevalence of the causative allele from these SNPs in women of African ancestry may explain the higher incidence of TNBC in AAW.
Modifiable risk factors that differed between populations in our study include caffeine and alcohol consumption, obesity and breastfeeding. In a study evaluating coffee and black tea consumption, a protective effect for coffee was found in pre-menopausal women, although this study was comprised of 98% Caucasian women [
27]. In contrast, results from the BWHS failed to find an association between caffeine consumption and breast cancer risk, either overall or by menopausal or hormone receptor status [
28]. Evaluation of alcohol consumption found a decreased risk of TNBC in alcohol consumers compared to non-drinkers and a significantly lower risk in those who consumed ≥7 drinks/week [
29]. Thus, the possible protective advantages conferred by caffeine and alcohol consumption may not be realized by AAW, although more research is needed to definitively determine the benefits of caffeine and alcohol use in patients with TNBC.
A number of studies have evaluated the role of obesity on development of TNBC with mixed results. A pooled analysis of data from the Breast Cancer Association Consortium, which is comprised of 92% patients of European ancestry, did not detect an association between obesity and TNBC in case–control analysis of young women, although case-case analysis did find an association between obesity and TNBC in young women [
30]. In contrast, associations between obesity and TNBC have been reported for patients not using hormone replacement therapy [
31], and an elevated waist-hip ratio was associated with increased risk of basal-like breast cancers [
32]. A recent meta-analysis found a significant association between obesity and TNBC in both case-case and case–control analyses, especially in pre-menopausal women [
33]. With nearly half of our African American TNBC population having a BMI ≥30, this high incidence of obesity may contribute to the higher frequency of TNBC in AAW.
Breastfeeding, or lack thereof, has also been associated with increased risk of developing TNBC. Case-case analysis found that patients in the CBCS with TNBC breastfed for shorter durations than those with luminal A tumors, and case-controls analysis found an inverse relationship between breastfeeding and risk of TNBC [
32]. A number of other studies have found an inverse association between breastfeeding and TNBC [
34‐
37]. In our study, although the cumulative number of months spent breastfeeding did not differ significantly between parous AAW and CW, only 33% of parous AAW with TNBC ever breastfed, compared to 63% of CW. In contrast, significantly different rates of breastfeeding were not detected in 115 AAW and 596 CW with ER+/HER2- tumors enrolled in the CBCP, thus failure to breastfeed in parous women may be a risk factor specifically for the development of TNBC.
Limitations of this study include possible selection bias and provision of equal-access health-care. Despite having no protocols to specifically recruit any ethnic group into the program, the CBCP has been effective in enrolling AAW, who encompass 16% of female patients with invasive breast cancer. Data regarding the number of patients who declined enrollment were not available, thus whether participation in the CBCP differs between AAW and CW could not be determined. Factors associated with refusal to participate in clinical trials include mistrust of the medical community, lack of compliance with research protocols, and increased co-morbidities [
38], thus, patients who agreed to participate in the CBCP may be healthier, more educated, and more compliant with short- and long-term treatments than those who did not. In addition, patients in the CBCP were provided with standardized health-care through the Department of Defense, which included screening mammograms, clinical breast exam, breast surgical procedures and chemo- and radiation therapies, regardless of ability to pay. Our study and that from Washington University [
20] failed to find survival differences between AAW and CW with TNBC who received similar clinical care, suggesting that TNBC is not inherently a different disease in AAW, but reflect disparities in access to quality health-care.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
LAS generated the microarray data and reviewed the manuscript, JM validated expression levels of CRYBB2P1 and reviewed the manuscript, KM performed statistical analysis and reviewed the manuscript, CDS provided patient samples and clinical interpretation of the data, REE designed the study and wrote the manuscript. All authors read and approved the final manuscript.