Can Breast Cancer Prevention Strategies Be Tailored to Biologic Subtype and Unique Reproductive Windows?

The advent of molecular profiling of breast cancers has led to the identification of at least 5 intrinsic molecular subtypes, cementing the concept that breast cancer is a heterogeneous disease arising from multiple, yet unknown, etiologies. The clinical importance of these molecularly defined subtypes lies in the fact that breast cancer survival rates differ by intrinsic subtype. Clinical proxies for the intrinsic subtypes have been delineated using routine tumor markers, including estrogen receptor, progesterone receptor, HER2, and tumor grade that incorporates proliferation rates, tumor differentiation, and nuclear dysmorphia. These clinically assigned “intrinsic-like” subtypes have permitted broad use of molecular subtyping data without requisite tumor RNA sequencing and are fundamental to patient care. For example, patients with luminal A-like cancers in general have good prognoses, and overall, these patients receive less chemotherapy compared with those diagnosed with triple-negative breast cancer (TNBC), a subtype that progresses to metastasis more frequently even after combined surgery, radiation, and chemotherapy treatments.

Although the understanding that breast cancer is a heterogeneous disease with distinct outcomes is fully recognized within treatment-research and clinical care specialties, the incorporation of these concepts for breast cancer prevention is still emerging. The potential for tailored prevention strategies is promising, because targeting the most lethal subtypes of breast cancer will have high impact. In this issue of the Journal, Jung et al. (1) report on associations between reproductive events, including age at menarche, childbearing, breastfeeding history, and menopause, and the risk of developing specific subtypes of breast cancer. The results from this large meta-analysis help point the way forward to subtype-specific prevention approaches informed by a better understanding of reproductive risk factors.

The Jung et al. (1) study, impressive in size and scope, is a meta-analysis of 31 European-centric, population-based case control and cohort studies that have been pooled by the Breast Cancer Association Consortium. Data were harmonized from 23 353 cases and 71 072 controls across 16 countries and 4 continents, making it the largest study of its kind to date. Further, the study builds on a storied tradition, with a history spanning 300 years of research. The link between reproductive factors and breast cancer risk was first reported in the early 1700s by Dr Bernardino Ramazzini, an Italian physician studying disease prevention from an occupational health perspective. Dr Ramazzini found that nuns had significantly higher breast cancer rates compared with women in other professions and hypothesized that nulliparity was a risk factor for breast cancer (2). Over the ensuing 300 years, we have come to understand that full-term pregnancy reduces breast cancer risk (the protective effect of pregnancy); this protection is modified by age at first pregnancy (the dual effect of pregnancy—as pregnancy is protective if you are young at time of first childbirth, but can be promotional if you are older); and proximity to recent childbirth is a risk factor for breast cancer regardless of a women’s age (the crossover effect, because close proximity to childbirth increases risk, with crossover to protection occurring with increasing time since last childbirth). At the same time, age at menarche, age at menopause, and breastfeeding history have been identified as additional reproductive factors that affect breast cancer risk.

In spite of this long, robust, and fruitful history of scientific inquiry, many solid epidemiologic studies investigating reproductive factors have yielded conflicting results, and to date, a unifying theory for how reproductive biology affects breast cancer risk remains elusive. One notable limitation of many of these epidemiologic studies is the binning of all breast cancer cases into a single homogenous disease due to the lack of available intrinsic subtype tumor data in many longitudinal studies and the relatively small numbers of TNBC cases. Jung et al. (1) overcame these barriers by assembling data from many studies that had sufficient clinical data to determine tumor subtype. They found intrinsic-subtype specificity to several of the reproductive risk factors assessed.

Two observations are particularly worth highlighting in the study by Jung et al. (1), because they have high translational potential with respect to breast cancer prevention and control efforts. The first observation is that women with TNBC have a distinct reproductive risk profile compared with all other subtypes. Somewhat surprisingly, the other subtypes, including the HER2 subtypes, share similar reproductive risk profiles. These observations reinforce the idea that TNBC is a distinct disease with unique etiology. Identified TNBC-specific risk factors include a persistent increased risk with parity; an interesting relationship between older age at first pregnancy and reduced risk of TNBC; and the observation that breastfeeding mitigates risk for TNBC, whereas breastfeeding is ineffective or only modestly protective for all other subtypes. Research focused on understanding the unique etiology of TNBC and how parity and breastfeeding interface with risk for TNBC has the potential to substantially reduce the burden of breast cancer mortality. For example, the long tail of increased risk of TNBC associated with childbirth suggests a disease initiation process, whereas the short duration increased risk for the luminal subtypes is consistent with recent childbirth promoting preexisting disease. Rodent models that distinguish between tumor initiation and progression may provide insights needed to effectively reduce incidence of TNBC. In addition, the findings regarding breastfeeding and reduced risk of TNBC offer additional evidence of beneficial health effects of breastfeeding for the mother as well as the baby; legislative policies and public health programs that support the feasibility of breastfeeding for new mothers may result in a reduced incidence of TNBC.

It is worth noting that the association of parity with an increased risk of TNBC, which conflicted with the commonly held belief that parity reduces risk of breast cancer, was first observed in the Carolina Breast Cancer Study (3), a case-control study designed to overrepresent young women and include approximately equal numbers of Black women and White women. The unexpected association was next shown in data from the Black Women’s Health Study, a prospective cohort of 59 000 US Black women, and in the African American Breast Cancer Epidemiology and Risk (AMBER) consortium, which pooled data from those 2 studies and 2 other studies of breast cancer in Black women (4,5). Previous subtype-specific analyses of women from predominantly White studies had failed to demonstrate a positive association of parity with TNBC (6), likely due to the lower proportion of TNBC cases in those studies and the higher prevalence of breastfeeding. Based on the current article by Jung et al. (1), with the largest sample size to date and over 3000 TNBC cases, breastfeeding appears to mitigate the negative effects of parity on TNBC in White women as well and provides persuasive confirmation of the earlier results obtained from studies focused on Black women. Collectively, these prior studies highlight the importance of conducting epidemiologic research in multiple populations with differing patterns of disease incidence and exposure prevalence.

The second dominant and potentially actionable observation from the research by Jung et al. (1) is the finding that risk for all breast cancer subtypes is increased within 5-10 years of a completed pregnancy, especially for a second or later pregnancy. For example, relative to nulliparous women, the odds ratio for biparous women associated with having given birth 5-9 years previously was 1.62 (95% confidence interval = 1.23 to 2.13) for luminal A-like subtypes and 3.28 (95% confidence interval = 2.33 to 4.63) for TNBC. Thus, proximity to last childbirth appears to increase breast cancer risk to levels equivalent to or higher than first-degree family history of breast cancer. These data fortify and extend previous observations suggesting that a recent childbirth is an independent risk factor for breast cancer, at least within this large European ancestry study, and argue for continued study of this risk window as a target for breast cancer prevention. Potential impact is suggested by reports that approximately 45% of all early-onset breast cancer cases are diagnosed within 5-10 years of childbirth (7,8). In some respects, one could argue that Dr Bernardino Ramazzini’s research has come full circle. For many years, there was a focus on understanding the protective effect of pregnancy with the goal of putting this biologic mechanism into a pill to prevent breast cancer (9,10). Research in the 21st century suggests that a focus on a possible “postpartum” pill, one designed to alleviate the tumor-promoting potential of recent childbirth, may be a worthy goal for breast cancer prevention (11,12).

Funding

NCI R01CA169175 and DoD BC191620 (PS). NCI UO1CA164974 and Susan G. Komen Foundation SAC180086 (JRP).

Notes

Role of the funder: The funders had no role in the writing of this editorial or the decision to submit it for publication.

Disclosures: The authors have no disclosures.

Author contributions: Writing—original draft: PS, Writing—review and editing: PS, JRP.

Data Availability

No data were generated or presented in this editorial.