Rapid Progression of Metastatic Pancreatic Adenocarcinoma During Platinum‐Based Therapy in a Patient Harboring a Pathogenic BRCA2 Germline Variant

Abstract

Familial pancreatic adenocarcinoma (PDAC) is most commonly related to inheritance of a pathogenic BRCA variant (J Med Genet 2005;42:711–719). The National Comprehensive Cancer Network recommends germline testing for patients diagnosed with PDAC and recommends platinum‐based chemotherapy as the preferred initial systemic therapy for patients harboring a pathogenic BRCA germline variant with PDAC (https://www.nccn.org/guidelines/guidelines-detail?category=1&id=1455).

PDACs related to pathogenic BRCA germline variants typically demonstrate BRCA loss of heterozygosity (LOH), which results in ineffective DNA damage repair due to a lack of normal BRCA gene product activity. By causing DNA damage, platinum‐based therapies have been shown to be highly effective therapies (Cancer Cell 2010;18:499–509, Gen Med 2015;17:569). In contrast, platinum‐based therapies would be predicted to be significantly less effective for PDACs in patients with pathogenic BRCA germline variants who have cancers that lack BRCA LOH.

Poly (ADP‐ribose) polymerase 1 (PARP) is also key to effective DNA repair. The Food and Drug Administration has approved PARP inhibitors for patients carrying germline pathogenic BRCA variants and metastatic breast cancer or ovarian cancer (Ann Oncol 2019;30:558–566, J Clin Oncol 2015;33:244–250). PARP inhibitors would again be expected to be far less effective in patients who carry pathogenic BRCA germline variants with breast and ovarian cancers (those that lack BRCA LOH) than in those with BRCA‐related breast and ovarian cancers (which typically demonstrate BRCA LOH), because PARP is involved in DNA repair.

Here, we present a patient harboring a pathogenic BRCA germline variant whose PDAC grew rapidly during platinum‐based therapy and lacked BRCA LOH and therefore was not likely BRCA related. Given the molecular fingerprint of BRCA‐related PDAC in patients with pathogenic BRCA germline variants and the mechanism of action of platinum‐based therapies and PARP inhibitors, this case underscores the importance of future studies aimed at determining whether the lack of BRCA LOH in PDACs in pathogenic BRCA germline variant carriers is a biomarker of less responsiveness to platinum‐based chemotherapy and PARP inhibitors.

Key Points

• Platinum‐based therapy or Poly (ADP‐ribose) polymerase 1 (PARP) inhibitor therapies are highly effective systemic therapy options for most patients with pancreatic adenocarcinoma who carry a germline pathogenic BRCA variant.

• In the case presented here, a patient carrying a germline pathogenic BRCA variant saw rapid progression of his pancreatic adenocarcinoma while on platinum‐based therapy. Next‐generation sequencing confirmed that his pancreatic cancer was likely not related to BRCA loss of heterozygosity (LOH).

• Studies are needed to determine, in patients who harbor germline pathogenic BRCA variants, whether similar cancers (i.e., those that lack BRCA LOH) are less responsive to platinum‐based or PARP inhibitor therapies than are those more common BRCA‐related cancers (i.e., those that demonstrate LOH).

Short abstract

Most patients with pancreatic adenocarcinoma who carry a germline pathogenic BRCA variant respond well to platinum‐based therapy or PARP inhibitor therapies. The case presented here suggests that more studies are needed to determine whether sporadic cancers are less responsive to such therapies than BRCA‐related cancers.

Patient Story

K.B. was a 62‐year‐old male who presented in October 2020 with abdominal pain. His father had prostate cancer. A computed tomography (CT) scan at presentation showed a 3.9 × 2.9 cm pancreatic tail mass and multiple apparent liver metastases. Endoscopic ultrasound showed a 2.8 × 2.8 cm pancreatic tail mass and a biopsy was interpreted as showing “adenocarcinoma with squamous features.”

During October and November 2020, the patient was treated with FOLFIRINOX chemotherapy (5‐fluorouracil/irinotecan/oxaliplatin/leucovorin). The CA19‐9 increased from 18.30 (nl 0.00–35.00 U/mL) in October 2020 to 65.81 in early December 2020. A repeat CT scan in December 2020 demonstrated that the liver metastases had increased in size (the largest lesion increasing from 7.4 × 4.2 cm to 12.3 × 8.3 cm) compared with October 2020, as well as the development of periportal lymphadenopathy and ascites.

Molecular Tumor Board

Multigene panel germline testing demonstrated BRCA2 c.2830 > T (p.Lys944*) and RECQL4 c.2869 > 7 (p.Gln957*) pathogenic variants (Invitae Corporation, San Francisco, CA), confirming that K.B. was a carrier of a pathogenic BRCA germline variant and was at risk for developing BRCA‐related cancers, including pancreatic ductal adenocarcinoma (PDAC).

Tumor next‐generation sequencing (NGS) demonstrated microsatellite status–stable; tumor mutational burden‐4 mutations/Mbase; pathogenic variants—BRCA2 K944*, KRAS G12V, TP53 M246R, and U2AF1 S34F; BRCA2 K944* mutation allelic frequency (MAF) was 47% (Foundation Medicine, Cambridge, MA). This MAF result was consistent with K.B.’s PDAC being less likely a BRCA‐related PDAC, which would typically—although not necessarily— demonstrate loss of heterozygosity (LOH).

Discussion of Targeted Therapy Considerations in Patients Carrying a Pathogenic BRCA Germline Variant Who Have Apparent, Non– BRCA ‐Related Cancers

Roughly 5% of hereditary PDACs are related to inheritance of a pathogenic BRCA variant [1]. Lowery et al. reported an 83% response rate (RR) of PDACs to platinum‐based therapy in patients carrying pathogenic BRCA germline variants [2], and O'Reilly et al. reported a 74% RR in a similar group of patients [3]. The Food and Drug Administration (FDA) approval of Poly (ADP‐ribose) polymerase 1 (PARP) inhibitor therapy for patients with PDAC who carry pathogenic BRCA germline variants is largely based on the roughly doubling of progression‐free survival compared with non‐PARP inhibitor therapy given to these patients [4, 5].

The classic molecular mechanism of transformation—the “two‐hit” hypothesis—is through BRCA LOH with somatic deletion of the wild‐type allele and has been considered “essential for carcinogenesis” [6]. However, a second non–BRCA‐related hit or inactivation of BRCA unrelated to a mutation in some cases is key to carcinogenesis in patients carrying a pathogenic BRCA germline variant. Cancers might be BRCA related—rather than sporadic—when transformation is through mechanisms (as described below) other than those resulting in BRCA LOH.

Cancer cells lacking expression of normal BRCA protein expression have deficient homologous recombination (HRD) responses to DNA damage. Platinum agents have been shown to be highly effective therapy in patients with hereditary PDACs presumably because these drugs cause further DNA damage and resulting cytotoxicity [7]. Cancer cells lacking normal BRCA protein expression also rely on expressed PARP to correct DNA damage [7, 8]. The National Comprehensive Cancer Network (NCCN) recommends platinum‐based chemotherapy as initial systemic therapy for patients with PDACs and pathogenic BRCA germline variants. PARP enzymes are also involved in DNA repair pathways, and PARP inhibitors are NCCN endorsed and FDA approved for patients with PDACs who harbor pathogenic BRCA germline variants [4, 9].

Based on the above presumed mechanisms of action of platinum‐based therapy and PARP inhibitor therapy, which rely on lack of normal BRCA expression for efficacy, PDACs without loss of normal BRCA protein expression (i.e., typically those tumors lacking BRCA LOH) would be expected to be far less likely to respond to either platinum agents or PARP inhibitors. For ovarian cancers in patients with pathogenic BRCA1 or BRCA2 germline variants, overall survival with platinum‐based therapy is decreased for those patients with pathogenic BRCA germline variants and ovarian cancer that demonstrate no BRCA LOH compared with those with BRCA LOH [8]. We were unable to identify reports of platinum‐based therapies or PARP inhibitor therapy in patients with pathogenic BRCA germline variants who develop sporadic pancreatic cancers that lack LOH. However, the response rate to standard platinum‐based therapy in patients with PDAC is roughly 32%, more than 90% of whom would be presumed to have cancers that lack BRCA LOH [10].

The BRCA MAF seen in K.B.’s PDAC was 47%. Given the presumed mechanism of action of platinum‐based therapy in patients with PDAC who harbor a pathogenic BRCA germline variant, it seems reasonable to conclude that the lack of LOH seen by NGS in K.B.’s cancer is one possible explanation for why his cancer progressed so rapidly on platinum‐based therapy. It also seems reasonable to consider that many of the 20%–25% of PDACs in patients with pathogenic BRCA germline variants that did not respond to platinum‐based chemotherapy lacked BRCA LOH [2, 3, 4, 5, 7].

However, it is important to emphasize the uncertainty regarding the reported 47% MAF. With tumor NGS, DNA from normal tissue is sequenced as well. The reported tumor MAF might therefore include DNA from normal tissue and 47% could represent “dilution” of the mutated BRCA, such that the actual MAF for tumor DNA could be significantly higher, possibly in a range consistent with BRCA LOH.

BRCA LOH commonly leads to HRD. HRD results in “cellular dependence on alternative error‐prone DNA pathways and the presence of HRD may expose specific therapeutic vulnerabilities” [11]. However, somatic events such as epigenetic silencing and other mechanisms “can serve as the ‘second‐hit’ in these genes” [11, 12]. For example, a distinct pattern of substitution mutations—Signature 3—has been identified in tumors with biallelic inactivation of BRCA1/2. Signature 3 has also been identified in patients with germline nonsense and frameshift variants in PALB2 and associated with epigenetic silencing of RAD51 and BRCA1 by promotor methylation. Polak and others concluded that “current predictors of PARP‐inhibitor or cisplatin sensitivity have notable clinical success in a limited number of patients, indicating the need for additional biomarkers such as HRD or Signature 3 to identify other patients who could also benefit from these treatments” [11].

PARP inhibitor therapy is also FDA approved for treating metastatic castrate‐resistant prostate cancer. de Bono et al. demonstrated a 77% response rate (i.e., greater than 50% reduction in prostatic‐specific antigen) in patients with castrate‐resistant prostate cancers with a tumor BRCA mutation [13]. The authors did not identify whether these responding tumors had tumor BRCA LOH [11]. However, it appears that prostate cancers with only sporadic BRCA2 mutations are fairly rare. Robinson et al. demonstrated that 90% of prostate cancers with tumors that demonstrated BRCA2 mutations showed BRCA2 LOH, suggesting to the authors that these were tumors related to inheritance of pathogenic BRCA2 germline variants [14]. As in pancreatic cancer, it therefore remains unclear whether BRCA is a biomarker for PARP inhibitor efficacy in castrate‐resistant prostate cancer [12, 13, 14].

The apparent lack of BRCA LOH described for this patient might explain the lack of response seen with platinum‐based therapy. However, given the uncertainty of the MAF reported and the possibility of other events resulting in HRD, further study is needed before concluding that a reported lack of BRCA LOH precludes responsiveness to therapies that target HRD.

NGS with MAF of tumor tissue or from cell‐free DNA is now widely available. Future studies of platinum‐based therapies and PARP inhibitor efficacy in patients with PDAC who carry pathogenic BRCA germline variants should include subgroup analysis to help determine whether lack of BRCA LOH in these tumors or lack of BRCA LOH in DNA shed from these tumors is a biomarker for efficacy of these therapies.

Patient Update

Tragically, K.B. died 3 months after his diagnosis of PDAC.

Glossary of Genomic Terms and Nomenclature

Biallelic inactivation; deficient homologous recombination; germline nonsense and frameshift variants; loss of heterozygosity; microsatellite; pathogenic germline variant; tumor mutation burden.

Author Contributions

Conception/design: Steven Sorscher, Shakti Ramkissoon

Provision of study material or patients: Steven Sorscher, Shakti Ramkissoon

Collection and/or assembly of data: Steven Sorscher, Shakti Ramkissoon

Data analysis and interpretation: Steven Sorscher, Shakti Ramkissoon

Manuscript writing: Steven Sorscher, Shakti Ramkissoon

Final approval of manuscript: Steven Sorscher, Shakti Ramkissoon

Disclosures

Shakti Ramkissoon: Foundation Medicine (E, OI). Steven Sorscher indicated no financial relationships.

(C/A) Consulting/advisory relationship; (RF) Research funding; (E) Employment; (ET) Expert testimony; (H) Honoraria received; (OI) Ownership interests; (IP) Intellectual property rights/inventor/patent holder; (SAB) Scientific advisory board