Cancer susceptibility genes and potential implication about systemic therapy for early stage breast cancer

The development of leukemias following chemotherapy and/or radiation therapy for breast cancer has been recognized for four decades. Some of these leukemias may not be secondary to cytotoxic therapy (therapy-related leukemias, or TRLs) but rather, at least in some cases, may arise independently from a common genetic background (1–3). In this context, it is intriguing that diverse epidemiologic studies have observed a familial link between breast cancer and acute myelogenous leukemia (AML), substantiating the notion that the two diseases may share specific inherited mutations in one or more cancer susceptibility genes that are integral to their pathogenesis (1, 3–6). Moreover, the risk for TRL is increased in women with positive family history for breast and/or ovarian cancers (3–6).

In this issue of Cancer, Churpek et al (7) conducted a retrospective tumor registry study that brings long-awaited genetic detail to these clinical and epidemiological observations. They used 42-gene BROCA targeted genomic capture and next generation sequencing (NGS) technologies to analyze the genomes of hematopoietic cells (either leukemia or post-leukemia treatment remission samples) and germline epithelial cells (buccal mucosa, skin fibroblast) obtained from women with leukemia who had received cytotoxic therapy for a previous diagnosis of early stage breast cancer (ESBC). The investigators found deleterious germline mutations in breast cancer susceptibility genes in roughly 20% of them. The aberrations occurred in genes involved in pivotal DNA damage response and repair pathways, namely BRCA1, BRCA2, p53, CHEK2 and PALB2 (partner and localizer of BRCA2). These data confirm and expand data from previous smaller studies regarding the link between breast cancer germline gene mutations and development of TRL (4,5,8). Interestingly, BRCA2 and PALB2 are integral components of the Fanconi pathway and the breast cancer that occurs in PALB2 mutation carriers may overlap with that of BRCA2 carriers (8). In particular, defects in both alleles of BRCA2 (identical to FANCD2) and PALB2 (identical to FANCN) result in Fanconi Anemia (FA), an inherited chromosomal instability disorder characterized by bone marrow failure, cellular hypersensitivity to DNA cross-linking agents, and a striking 800-fold increased risk for developing myelodysplasia (MDS) and AML (9). There are currently 18 complementation groups that have been defined, and all FA proteins cooperate in a pathway that leads to recognition and repair of DNA damage (9). Moreover, FA proteins coordinate other proteins to determine the net response to DNA damage, including ATM, ATR, Bloom Syndrome and CHK1 (9). Thus, it seems plausible to speculate that inherited mutations or acquired defects such as epigenetic silencing of one or more FA genes could have a formative role in both breast cancer and AML pathogenesis.

Of interest, while the vast majority of the leukemias in the Churpek report (7) were AMLs, 8% were acute lymphoblastic leukemias (ALLs), a finding that recapitulates a recent observation by our group (3). These ALLs appear to be linked selectively to germline mutations in p53 and, as detected in the large study by Stengel et al (10), are accompanied by complex karyotypes.

The authors raise the issue of what is a “true TRL” and what is potentially an independent leukemia, using as an example a patient who developed t(3;21)-associated AML 18 years after breast cancer chemotherapy. In fact, in our recent analysis of marrow malignancies arising after adjuvant therapy for ESBC, we found that the risk for developing a marrow neoplasm did not plateau at 5 years but instead remained steady well into years 6 through 10 (3). In addition, t(3;21) often occurs in the setting of environmental exposures (e.g., benzene) and is associated with MDS/AML, which in some instances may take decades to fully evolve. Thus, it can be difficult to determine what constitutes TRL and what may be a second malignancy that is unrelated to a previous toxin (2).

Despite these questions, the information yielded by germline testing in specific subsets of patients following a diagnosis of ESBC may potentially identify those for whom close monitoring of hematologic parameters might be considered, and could potentially lead to early detection and intervention for marrow malignancies, especially MDS. In this context, it is interesting that the somatic mutations (detected by 70-gene Oncoplex targeted genomic capture and NGS) in 8 of the 9 AML populations tested included the epigenetic regulatory genes DNMT3A, TET2 and ASXL1 that are linked to MDS/AML as well as other AML-associated genes (e.g., FLT3, NRAS, NF1, WT1). Provocatively, mutations in ASXL1 may be linked to a therapy-related etiology (11) and thereby be able to help assess whether or not an AML arising in the setting of breast cancer is indeed a TRL (2).

In addition to defining specific inherited and somatic genomic aberrations, this retrospective study by Churpek et al (7) highlights some important issues. We think that prospective collection of paired malignant and normal cells samples with comprehensive clinical annotation, and in particular a careful family history, is critical to refining our understanding of genetic risk for specific types of cancers. The availability of the 40-year repository at the University of Chicago provides a unique opportunity for these studies. Nonetheless, the 88 breast cancer survivors in this report incompletely represent the true universe of breast cancer patients treated with cytotoxic therapy and at risk for a subsequent TRL. As we suggested in our recent report of the National Comprehensive Cancer Network (NCCN) experience (3), existing familial cancer registries that are prospectively following breast cancer patients and their families are uniquely positioned to ascertain the true frequency of subsequent leukemias and their associations with the therapies received and the known germline genetic alterations.

A clearer understanding of these associations could inform clinical practice and potentially improve clinical outcomes. It may be premature to advocate for broad population-based germline BRCA1/BRCA2 germline mutation screening, due to concerns about the prevalence of variants of unknown significance when testing is done in otherwise unselected healthy individuals at low risk for harboring deleterious germline mutations. However, it is worth discussing scenarios that might lead to broadening the criteria and timing for germline mutation testing for selected individuals diagnosed with ESBC.

At this present time and as exemplified in the most recent NCCN guideline (v1.2015), germline testing is viewed as informative for patient management regarding decisions about preventive options like prophylactic surgeries and for immediate relatives as part of genetic counseling. However, we can envision a time when genetic testing might also be used to guide systemic therapy decisions. As an example, the TNT trial recently reported in abstract form (12) showed that, among patients with locally advanced or metastatic triple negative breast cancer, a subset of 43 patients ultimately found to carry a germline BRCA1 or BRCA2 mutation had a much higher response rate and longer progression-free survival when treated with carboplatin instead of docetaxel, and ongoing trials evaluating the survival benefit offered by platinum agents in selected ESBC (such as BRCA1 or BRCA2 germline mutation carriers) may confirm the clinical utility of such patient selection strategies.

In the meantime, we expect that the ongoing discovery of germline mutations in additional genes affecting DNA repair pathways and TRL susceptibility may further refine our understanding of TRL pathogenesis and risk. As a case in point, the presence of t(9;11) AML is clearly a treatment-related event, occurring in 10–20% of the TRLs observed in women following chemotherapy for ESBC. In fact, the findings in the current study are similar to our series (3), where translocations in the mixed lineage leukemia (MLL) gene were observed in seven (26%) of 27 patients with prior exposure to adjuvant chemotherapy (five patients with AML and two patients with ALL. In contrast, no MLL translocations were observed among the 14 patients with available cytogenetics who received no prior chemotherapy (3). It is therefore tempting to speculate, as did the authors (7), that the genetic lesion(s) predisposing to development of t(9;11) AML reside in genes responsible for nonhomologous end joining repair that determines the fidelity of recombination events, a finding that could shed light on the pathogenesis of diverse translocation-associated leukemias.

Hence, it is certainly conceivable that germline mutation testing may in the future be used to help refine locoregional and systemic decisions in ESBC in view of improved characterization of the risk of TRL, both in terms of the selection of chemotherapy regimen and possibly even to help decide if to administer adjuvant chemotherapy at all, especially in patients with lower risk tumors where the potential improvement in survival benefit might be small. For instance, recent data from the Suppression of Ovarian Function Trial (SOFT) in premenopausal women with estrogen receptor (ER)-positive ESBC showed that an improvement in outcomes was observed in those who were felt to be at sufficient risk for recurrence to have received adjuvant chemotherapy and had subsequently remained premenopausal (13). Available evidence from other studies like IBCSG Trial VIII indicate that ovarian function fully recovers after temporary suppression with an LH-RH agonist, in contrast to the permanent ablation effects often observed with chemotherapy (14). In aggregate, these findings indirectly suggest that optimal dual endocrine therapy with an LHRH agonist without chemotherapy might be a feasible strategy for many premenopausal patients with ER-positive ESBC, as it exploits the related benefits from the temporary suppression of ovarian function, reduces long-term health-related complications from earlier onset of menopause, and may reduce the risk of TLRs. In other words, while the small risk of recurrence for many patients with ESBC still justifies consideration of adjuvant systemic therapy, the risk of serious side effects associated with some treatment modalities like chemotherapy is not zero.

In conclusion, the data by Churpek et al (7) expand on a growing body of knowledge for understanding an individual’s genetic risk, not only for breast cancer but also for second cancers such as leukemia, whether they are a direct result of cytotoxic therapy or an independent event. As we better understand the long-term effects of the therapies we have to offer, these data will become increasingly critical to guide strategies for cancer risk reduction and also decisions about optimal locoregional and systemic therapies in patients with ESBC, so as to maximize benefit and minimize adversity especially as most nowadays are expected to survive their breast cancer diagnosis.