How Inherited BRCA Mutations Shape Breast Cancer's Genomic Behavior and Point Toward a New Treatment Paradigm

Germline mutations in BRCA1 and BRCA2 have long been recognized as breast cancer risk genes, but whether they continue to shape a tumor's behavior after it develops—and specifically how it responds and eventually resists treatment—has been far less clear. A new study published in Nature by researchers at Memorial Sloan Kettering Cancer Center (MSK) and collaborators addresses this question directly, using clinical genomic data from more than 5800 patients with breast cancer to examine how inherited BRCA1 and BRCA2 mutations interact with the tumor's later somatic mutations.¹

The study found that BRCA1 and BRCA2 carriers, despite both being HRD homologous recombination-deficient (HRD) genes historically analyzed together, diverge sharply in how their tumors evolve. Germline BRCA2 mutations in particular were strongly associated with acquired loss of RB1, a key tumor suppressor and well-established driver of resistance to CDK4/6 inhibitors—a cornerstone therapy for hormone receptor-positive, HER2-negative breast cancer. The researchers traced this vulnerability to 2 compounding factors: BRCA2 carriers are more likely to already carry only 1 functional copy of RB1 before treatment even starts (because the 2 genes sit close together on chromosome 13), and the ongoing genomic instability caused by BRCA2 loss makes it easier for the tumor to knock out that remaining copy under the selective pressure of CDK4/6 inhibitor treatment.

Consistent with this, the team found that patients with germline BRCA2 mutations had significantly shorter progression-free survival on CDK4/6 inhibitors plus endocrine therapy compared with noncarriers, a finding replicated in an independent, real-world clinical dataset. Conversely, PARP inhibitors, which directly exploit HRD biology, produced better outcomes in these same patients than CDK4/6 inhibitors did, even when given in later treatment lines. Preclinical models, including patient-derived xenografts, further confirmed that RB1 loss driven by HRD biology was the mechanism behind CDK4/6 inhibitor resistance in BRCA2-mutated tumors, while sensitivity to PARP inhibitors remained intact. These findings form the scientific basis for EvoPAR-Breast01 (NCT06380751), an ongoing international phase 3 trial testing whether moving a PARP inhibitor to the frontline setting, ahead of CDK4/6 inhibitors, can improve outcomes for this genomically defined group of patients.²

Pedram Razavi, MD, PhD, a medical oncologist at MSK and senior author of the study, spoke with Targeted Oncology^TM about how the research came together, what the data revealed about the distinct biology of BRCA1 vs BRCA2 tumors, and what questions the team hopes to answer next.

Targeted Oncology: What got you involved in this line of research?

Pedram Razavi, MD, PhD: We've been doing genomic analysis, and we started doing genomic analysis clinically at MSK since 2013. That's when we started offering clinical testing to many of our patients—almost all of our patients with metastatic disease and some of our patients with early-stage disease. This idea came up very early on, when we started doing this: Can we bring in high-quality clinical data from our patients and use this clinical genomic platform as a discovery tool to understand the biology of breast cancer, understand mechanisms of response and resistance to therapy?

This program has been pretty successful so far. We have almost 10,000 patients—more than 10,000 tumors with breast cancer sequenced at MSK—and the analysis so far has resulted in a lot of novel mechanisms of resistance to endocrine therapy, to CDK4/6 inhibitors, anti-HER2 therapy, and antibody–drug conjugates. In this particular study, we aimed to answer a different question. In all of our previous analyses, we've been focusing on somatic mutations: the tumor-derived variants that are in the tumor. Here we asked: do the inherited variants, the ones patients are born with, have any effect on the genomic profile of the tumor? And if so, does this interaction change the outcome of patients on certain therapies that we offer?

What were your findings?

When we started doing the analysis, we started with looking at hyper-mutating genes—genes known to be associated with increased risk of cancer—because we thought these are the alterations with potentially the biggest effect on the somatic profile of the tumor. It was very striking, because previous studies had been done on relatively small numbers of patients and put BRCA1 and BRCA2 together in almost all of those analyses. Here, for the first time, because of the size of the cohort we have, we were able to differentiate between BRCA1 and BRCA2. And as soon as we did that, we found a very distinct difference in how BRCA1 mutations affect the somatic profile of the tumor compared to BRCA2—something that was completely unknown before this.

The striking part was that TP53 mutations, which are common in BRCA-associated tumors, are actually very much enriched in BRCA1 but not BRCA2, and RB1 loss-of-function mutations—RB1 being a major regulator in the cell cycle pathway—were depleted in BRCA1 but very significantly enriched in BRCA2. That was the hint to us that the biology of the disease isn't just based on which pathway is affected by the genetic variant, but how it affects the pathway, and the other interactions these genetic factors can have with how the tumor behaves.

Then we moved on and asked: if RB1 alterations are associated with BRCA2, and RB1 is such a major mechanism of resistance to CDK4/6 inhibitors, are inherited BRCA2 mutations also associated with outcomes and response to CDK4/6 inhibitors? There was clearly a signal there: BRCA2 mutant tumors were resistant to CDK4/6 inhibitors, both first line and later lines of treatment. We were able to replicate and validate that in a large independent cohort outside MSK as well, so we were confident about these results.

Then we looked into the biology of it. RB1 loss-of-function mutations are very rare in breast cancer overall—they happen in less than 1% of breast cancers—but in BRCA2 tumors, it was seen in almost 20% of tumors. But what was really interesting was that the majority of these RB1 loss-of-function mutations were acquired; they weren't seen prior to exposure to CDK4/6 inhibitors. That gave us a hint that this interaction between BRCA and RB1 can change the behavior of cancer under the selective pressure of CDK4/6 inhibitors.

There was a hint here: BRCA2 and RB1 are both on chromosome 13q, and patients born with a defective germline BRCA2 variant often go through a process during tumorigenesis called loss of heterozygosity—they lose the wild-type copy of BRCA2 to develop complete loss of BRCA2. Because RB1 and BRCA2 are relatively close together, RB1 is often lost in that same process. So, a lot of these BRCA2 germline tumors already had only 1 copy of RB, and it immediately clicked to us that we're essentially lowering the evolutionary barrier—instead of needing 2 hits to develop RB1 loss of function, the tumor now just needs 1 hit, because it already lost 1 copy through that process.

We showed that RB1 loss of heterozygosity and hemizygosity—meaning only 1 copy of RB remaining—was strongly associated with acquired RB1 loss-of-function mutations on CDK4/6 inhibition, because almost exclusively, these acquired mutations were seen in tumors that, in a pretreatment sample, already had only 1 copy of RB. We also showed that mutational processes—specifically HRD as part of BRCA loss—contribute to this too, because these tumors are more likely to develop loss-of-function mutations. So, there's a mutagenesis process and a copy-number process, and both come together to make these tumors very prone to developing CDK4/6 inhibitor resistance through loss of RB1 function.

It seems this isn't unique to RB1. Some other tumor suppressors behave this way too, but this RB loss was unique to CDK4/6 inhibitors specifically. Selective pressure from endocrine therapy alone didn't result in acquisition of RB1 mutations. The tumor needed to go through that specific pressure to develop the phenotype. Breast cancers don't really 'like' RB1 loss-of-function mutations—it's not a very fit phenotype for breast cancer. So only under this particular selective pressure were the tumors pushed and cornered into developing RB1 loss of function.

Do you see these findings having implications for how we approach genomic testing during the course of breast cancer treatment?

At this point, especially with these acquired resistance mechanisms, we've mostly been reactive. We wait until the tumor develops a mechanism of resistance, and then we react, we go after that mechanism if it's actionable. We think this is a relatively new paradigm for us. If we can predict with high accuracy what the potential mechanism of resistance will be for a given tumor, we can start being proactive.

There are many different paths to resistance. We and others have identified many mechanisms of resistance to CDK4/6 inhibitors. In the beginning it seemed like all tumors could develop all of these, but now we're fine-tuning our cohorts and sub-cohorts to understand who the patients are who can potentially develop an RB1 loss-of-function mutation, and this can be expanded to other mechanisms as well.

More importantly in this scenario, we found that HRD tumors—BRCA-carrier patients—when they receive a PARP inhibitor first, seem to do better, even better than they did in the clinical setting where we usually give PARP inhibitors far later than CDK4/6 inhibitors, as second, third, or later lines. We showed that progression-free survival, as well as responses to PARP inhibitors, in the majority of these cases were superior to the first-line response to CDK4/6 inhibitors, despite being a later line of treatment. That gave us a hint that we could bring PARP inhibitors to the frontline setting.

Another hint came from extensive lab work we did in collaboration with Sarat Chandarlapaty, MD, PhD’s lab, my colleague [at MSK]. We developed animal models and patient-derived xenografts [PDXs] from our BRCA-carrier patients. We had 1 patient for whom we had 2 tumors: 1 BRCA carrier with intact HRD, and another that, because the patient had received PARP inhibition, developed a reversion mutation that corrected the BRCA2 reading frame and was no longer HRD. We showed that resistance only happened in the tumor that did not have the reversion mutation, and around 30% to 40% of patients on PARP inhibition develop reversion mutations.

If we bring PARP inhibition to the frontline setting, not only might we improve outcomes in the first-line setting, but because a lot of these tumors also correct their HRD status through reversion mutations or other mechanisms, they may go on to have a better outcome on the next-line CDK4/6 inhibitor as well. That's exactly what we showed in the animal model and overall across our multiple PDXs: PARP inhibition had a better outcome compared with CDK4/6 inhibitors in this setting.

That's the basis for the clinical trial currently ongoing: EvoPAR-BC1, a phase 3 randomized trial of saruparib plus a next-generation endocrine therapy vs standard-of-care endocrine therapy plus a CDK4/6 inhibitor, in patients with germline or somatic BRCA1, BRCA2, or PALB2 mutations. We're exactly testing this hypothesis: can we bring PARP inhibitors to the frontline setting, and can we improve overall outcomes for patients, both progression-free survival and overall survival?

What questions do you still have? What are you still trying to figure out?

We think this is the very beginning of a new field. If we can, with high accuracy, predict the behavior of cancer under therapeutic pressure, it has major implications both in the metastatic setting and in the curative, early-stage setting. If we know a particular tumor is prone to develop a specific resistance mechanism, do we need to wait until that resistance happens, the resistant clone expands, and develops much more clonal diversity as it grows—making it much harder to treat?

If we know the mechanism of resistance ahead of time, we can potentially bring this type of approach to a much earlier setting—the early-stage setting, or in the metastatic setting, try to block the evolutionary paths available to the tumor, and by doing that, we might be able to cure a subset of metastatic patients. We don't know that yet. But it seems like a tumor may only have the capability of developing 2 or 3 mechanisms of resistance to a particular therapy.

We're putting together a whole program now to go after other mechanisms of resistance in this way. I don't think we can predict everything at this point, but there are a lot of these highly effective phenotypes that tumors are prone to developing that could potentially be predicted. For example, ESR1 mutations don't happen in every patient. Even in the very late setting, only 40% to 50% of patients develop ESR1 mutations after exposure to aromatase inhibitors and antiestrogen therapies. Why can't we predict which patients those will be, and why the rest don't develop these mutations? There are a lot of similar questions we can ask and hopefully answer.

Watch the interview with Dr Razavi.

REFERENCES

1. Safonov A, Lee M, Brown DN, et al. Homologous recombination deficiency and hemizygosity drive resistance in breast cancer. Nature. 2026 Apr;652(8110):752-762. doi: 10.1038/s41586-026-10197-0. Epub 2026 Mar 4. PMID: 41781623; PMCID: PMC13083263.

2. Saruparib (AZD5305) Plus Camizestrant Compared With CDK4/​6 Inhibitor Plus Endocrine Therapy or Plus Camizestrant in HR-Positive, HER2-Negative (IHC 0, 1+, 2+/​ ISH Non-amplified), BRCA1, BRCA2, or PALB2m Advanced Breast Cancer (EvoPAR-BR01). ClinicalTrials.gov. Updated June 29, 2026. Accessed July 23, 2026. https://clinicaltrials.gov/study/NCT06380751