Pancreatic Cancer: BRCA Targeted Therapy and Beyond

Simple Summary Pancreatic cancer is associated with poor outcomes for several reasons, including diagnosis at an advanced stage, the absence of effective screening for the diagnosis, and resistance to treatments. Pancreatic cancers associated with BRCA1/2 mutations have emerged as a distinct subgroup with sensitivity to other treatments and, in some cases, durable responses. Furthermore, beyond BRCA1/2 mutations, there is increasing recognition that other gene mutations may behave in a similar manner. The focus of this review is to discuss recent developments in the management of BRCA-associated pancreatic cancer, emerging therapeutic strategies, and future directions for this subgroup of patients. Abstract Pancreatic ductal adenocarcinoma (PDAC) is projected to become the second leading cause of cancer-related death in the US by 2030, despite accounting for only 5% of all cancer diagnoses. Germline gBRCA1/2-mutated PDAC represents a key subgroup with a favorable prognosis, due at least in part to additional approved and guideline-endorsed therapeutic options compared with an unselected PDAC cohort. The relatively recent incorporation of PARP inhibition into the treatment paradigm for such patients has resulted in renewed optimism for a biomarker-based approach to the management of this disease. However, gBRCA1/2 represents a small subgroup of patients with PDAC, and efforts to extend the indication for PARPi beyond BRCA1/2 mutations to patients with PDAC and other genomic alterations associated with deficient DNA damage repair (DDR) are ongoing, with several clinical trials underway. In addition, despite an array of approved therapeutic options for patients with BRCA1/2-associated PDAC, both primary and acquired resistance to platinum-based chemotherapies and PARPi presents a significant challenge in improving long-term outcomes. Herein, we review the current treatment landscape of PDAC for patients with BRCA1/2 and other DDR gene mutations, experimental approaches under investigation or in development, and future directions.


Introduction
By 2030, pancreatic ductal adenocarcinoma (PDAC) is projected to become the secondleading cause of cancer-related mortality in the US [1]. The 5-year survival rate for PDAC is increasing and is now 12%, up from 5-6% in the previous decade [2]. Approximately 80% of new diagnoses are in the setting of advanced or metastatic disease [3], and in those patients who do undergo curative-intent surgery, disease recurrence is observed in up to 80% [4]. Until recently, precision medicine approaches to the management of PDAC have demonstrated modest incremental benefits; however, recent developments have given cause for renewed hope for specific subsets of patients.
The Pancreas Olaparib Ongoing (POLO) trial was the first phase III trial to establish a biomarker-based approach to the management of PDAC for patients with germline gBRCA1/2 mutations [5,6], highlighting BRCA-associated PDAC as a distinct biological group with the potential for a personalized approach to treatment. Stadler et al. reported a pan-cancer analysis of almost 12,000 patients at Memorial Sloan Kettering Cancer Center (MSK) across over 50 tumor types who underwent comprehensive germline testing [7]. The OncoKB Precision Oncology Knowledge Base [8] was used to define the actionability of identified gene alterations, and 19.6% of PDAC cases were identified to harbor a pathogenic or likely pathogenic germline variant. Earlier, Lowery and colleagues carried out comprehensive germline testing on 615 patients with exocrine pancreatic neoplasms and identified pathogenic germline variants in 19.8% [9]. In both studies, pathogenic or likely pathogenic variants in BRCA1/2 and other genes associated with homologous recombination accounted for a substantial proportion of the variants identified. With the National Comprehensive Cancer Network (NCCN) [10] and American Society of Clinical Oncology (ASCO) guidelines [11] recommending both germline and somatic testing for PDAC, there is potential for the incorporation of precision medicine approaches in a greater number of patients with PDAC than before. Herein, we review recent advances and future directions in the approach to the management of patients with PDAC, with a focus on BRCA1/2 and other key HR gene mutations.

BRCA and DNA Damage Repair Pathways
Based on the type and frequency of structural variations identified by whole exome sequencing of a large cohort of PDAC tumor samples, Waddell et al. proposed four distinct PDAC subtypes: stable, locally rearranged, scattered, and unstable [12]. The unstable subtype is characterized by recurrent structural variation events owing to germline and somatic mutations of genes known to be involved in DNA damage repair (DDR), including BRCA1, BRCA2, PALB2, RAD51C, RAD51D, and ATM. These genes are relevant for the normal function of homologous recombination (HR) repair. Germline and somatic alterations in these key genes are now recognized to potentially lead to a HR deficient (HRD) phenotype within a tumor [13], which results in deficient double-stranded DNA break repair [14,15]. The roles of the BRCA1 and BRCA2 genes in encoding proteins required for HR are well described [16]. More recently, the function of PALB2 as a "binding protein" for BRCA2, localizing BRCA2 to sites of DNA breaks [17], has also been well supported. Together, BRCA1/2 and PALB2 are recognized as "core" genes that, when mutated, confer HRD in most settings of PDAC. Other "non-core" DDR variants and their role in inducing HR are less well defined but of growing interest.

Prevalence of BRCA and HRD in Pancreatic Cancer
The initial analysis of the Know Your Tumor registry trial identified mutations in the HR-DDR (DNA damage repair) pathway in 25% of 820 patients with PDAC, subdivided into three groups: Group 1: BRCA1/2 and PALB2; Group 2: ATM/ATR/ATRX; and Group 3: BAP1, BARD1, BRIP1, CHEK1/2, RAD50/51/51B, or FANCA/C/D2/E/F/G/L [18]. While gBRCA1/2 and gPALB2 mutations are identified in 5-6% of an unselected PDAC population [14], rates of gBRCA mutations are known to be enriched in certain populations, including those of Ashkenazi Jewish heritage (5-16%), patients with a family history of pancreatic cancer (5-19%), and those with a family history of ovarian or breast cancer (5-10%) [19][20][21][22]. A retrospective analysis of the first 2206 patients screened for enrollment in the POLO trial demonstrated an overall prevalence of BRCA1/2 mutations of 7.2% (5.8% after exclusion of populations known to be enriched with people of Ashkenazi Jewish heritage) and identified substantial geographic as well as some racial variability in the prevalence of gBRCA mutations amongst those screened, acknowledging a degree of selection bias given that almost 20% of patients enrolled in this trial had a prior documented gBRCA mutation [23].
While most patients with PDAC and an associated HRD phenotype have a pathogenic germline variant, a minority (2-4%) harbor a somatic-only HRD alteration. The widespread use of next-generation sequencing (NGS), as well as improvements in the sequencing assays utilized, has permitted more comprehensive identification of alterations in DDR genes. Waddell et al.'s whole genome sequencing and copy number variation analysis on 100 patients with PDAC identified two with a somatic BRCA1 mutation and three with a somatic BRCA2 alteration [12]. Work from our group demonstrated somatic HRD mutations in 4% of our cohort with PDAC and observed similarly favorable outcomes relative to those with germline HRD who received first-line platinum-based chemotherapy [14].

BRCAness Phenotype
While pathogenic variants in the DDR pathway have the potential to result in HRD and impaired double-stranded DNA break repair, as exemplified by BRCA1/2 mutations, this does not appear to be the case for all DDR pathogenic variants. The concept of BRCAness or HRD phenotype, defined as double-strand break repair deficiency in the absence of a BRCA1/2 variant [24,25], is of growing interest across tumor types, including PDAC. Tumors exhibiting BRCAness appear to respond favorably to DNA-damaging therapies such as platinum-based chemotherapy and poly-ADP ribose polymerase inhibitors (PARPi). There is an increasing recognition that, beyond focused germline and somatic profiling of HR-associated genes, evaluating genomic signatures through validated HRD scores may represent a more inclusive method of determining tumors that may respond to DDRtargeted treatments. These methods are more established in the context of breast and ovarian cancers [26,27]. Candidate variants of particular interest in PDAC and whether they confer a BRCAness phenotype include PALB2, RAD51, RAD50, CHEK2, ATM, BRIP, and BLM variants.
Several methods for determining HRD status are in various stages of development. DNA "scars", referring to gross chromosomal abnormalities and mutational signatures, are specific genomic features that are characteristic of HRD [28]. The three key types of chromosomal aberrations associated with HRD are telomeric allelic imbalance (TAI), large-scale state transitions (LSTs), and loss of heterozygosity (LOH) [29][30][31]. In triplenegative breast cancer, a composite score of TAI, LST, and LOH is a robust biomarker for identification of HRD status, demonstrating improved pathologic complete response rates to platinum-based chemotherapy in those with a high combined score, regardless of BRCA1/2 status [26]. This strategy of HRD detection is the basis of Myriad's MyChoice HRD assay, which, in addition to testing for pathogenic BRCA1/2 mutations, also evaluates Genomic Instability Scores (GIS) based on TAI, LST, and LOH and is now an FDA-approved companion test to determine eligibility for olaparib and niraparib [32,33].
Beyond genomic scars, stereotyped mutational patterns based on the accumulation of single-base substitutions (SBS), insertion deletions (indels), and rearrangements are observed in tumors with HRD [28,34]. Signature 3, a SBS pattern associated with large deletions and microhomology, is observed regularly in BRCA1/2 and PALB2-mutated PDAC [12,35] and has been demonstrated to predict response to platinum-based chemotherapy in patients with PDAC [15]. Other tests based on NGS have been developed, most notably HRDetect, which employs a weighted model of mutational signatures based on whole genome sequencing to detect HRD in breast, ovarian, and pancreatic cancer [36]. A recent meta-analysis of over 60 studies and >21,000 patients with PDAC identified the prevalence of HRD to be 14.5-16.5% based on targeted NGS but 24-44% by whole genome or whole exome sequencing [37], highlighting the need to clarify the definition and optimal method of detection of HRD in PDAC.
In one large study of 391 patients with PDAC, including 49 with gBRCA1/2 or gPALB2 pathogenic variants, HRD classifiers, including (i) GIS using Myriad's MyChoice assay, (ii) Signature 3, (iii) HRDetect, and (iv) structural variant burden, were applied [15]. In this study, GIS scores had a sensitivity of 91% and a specificity of 83% for the identification of HRD, and improved clinical outcomes with platinum chemotherapy were associated with higher GIS scores. Most notably, HRDetect appeared to outperform both Sig3 and GIS scores in its ability to identify HRD PDAC, with high HRDetect scores present in up to 10% of patients who did not have a germline HR gene mutation. A HRDetect score >0.7 predicted gBRCA1/PALB2 deficiency with a sensitivity of 98% and specificity of 100%. In addition, high HRDetect scores predicted improved survival in patients receiving platinum-based chemotherapy. A notable feature of this study was that pathogenic variants in CHEK2 and ATM (N = 2 and N = 6, respectively) did not result in HRD by any of the above criteria. Identifying which, if any, of the DDR gene mutations result in a HRD phenotype and would thus present similar therapeutic actionability as BRCA1/2 mutations is of considerable interest. As an example, our group at MSK recently evaluated 46 patients with PDAC and germline or somatic ATM variants [38] and determined that while ATM-mutated PDAC is associated with a favorable overall survival (OS) relative to genomically unselected PDAC, pathogenic variants in ATM did not appear to confer a HRD signature [38]. Several studies are underway evaluating other candidate HRD genes and whether they confer a HRD phenotype, as well as additional HRD/BRCAness detection methods, including integrating DNA-and RNA-based HRD detection methods together [39]. The selected tests for evaluating HRD status are summarized in Table 1. Table 1. Select Tests to Evaluate for HRD (not specific to pancreatic cancer).

Test Description
Telomeric Allelic Imbalance (TAI) [28] Allelic imbalance at the telomere of the chromosome is due to the propensity for inappropriate end joining in HRD, identified by single nucleotide polymorphism (SNP) genotyping.
Large Scale Transitions (LST) [29] Chromosomal breaks larger than 10 Mb, which arise in HRD cells secondary to erroneous recombination between segments of the chromosome, are identified by single nucleotide polymorphism (SNP) genotyping.
Loss of Heterozygosity (LOH) [30] Uniparental disomy is owing to inaccurate repair of sister chromatids during S/G2 phase, resulting in the loss of entire genes and the surrounding chromosomal region, as identified by single nucleotide polymorphism (SNP) genotyping.
Genomic Instability Score eg by Myriad Genetics MyChoice Assay TAI + LST + LOH

Signature 3 (Sig 3) [35]
A single base substitution mutational pattern, associated with microhomology and large deletions, was identified by whole exome sequencing.
HRDetect [36] A weighted model incorporating a weighted score of base substitution/rearrangement signatures, microhomology-mediated deletions, and an HRD score based on genomic scars identified by whole exome sequencing.

Platinum-Based Chemotherapy
Platinum chemotherapies can crosslink purine bases in DNA, interrupting DNA transcription and replication, which leads to the accumulation of DNA damage due to a limited DNA damage repair capacity. The body of evidence supporting the use of platinum-based therapy in patients harboring BRCA1/2 mutations is ever-growing, including in PDAC, with several studies demonstrating superior outcomes with platinum-based chemotherapy in patients with BRCA1/2 mutations compared with unselected PDAC cohorts. Data suggests that the pleiotropic effects of mutant BRCA1/2 are tumor-lineage dependent and that the therapeutic relevance of gBRCA1/2 mutations is of most relevance in "BRCA-associated cancer types", namely pancreatic, breast, ovarian, and prostate cancer [40].
In the Know Your Tumor Program, for patients with advanced PDAC who received platinum-based chemotherapy, the mOS was 1.27 years for HR-DDR mutated patients (N = 53) versus 1.45 years for patients with proficient HR-DDR status (N = 258), representing a statistically significant and clinically meaningful difference (p = 0.000072; HR 0.44; 95% CR 0.29 to 0.66). Data from our group at MSK demonstrated that in 50 patients with advanced PDAC and a germline or somatic HR-associated mutation treated with systemic therapy, a median progression-free survival (mPFS) advantage was observed for those in receipt of a platinum regimen versus a non-platinum regimen (12.6 vs. 4.4 months, HR 0.44, 96% CI 0.29-0.67, p < 0.01) [14]. This analysis also demonstrated that patients with biallelic HR mutations had higher genomic instability and derived further benefit from front-line platinum therapy, with a significantly improved mPFS of 13.3 months (95% CI, 0.26-0.70) in platinum-treated patients versus 3.8 months (95% CI, 2.79-not reached (NR); p < 0.0001) in non-platinum-treated patients. Momtaz et al. reported on a large cohort of patients with PDAC and germline or somatic BRCA1/2 mutations [41]. Of 81 patients with stage IV disease, the mOS for patients who received upfront platinum-based therapy (N = 65) was 23 months (95% CI, [19][20][21][22][23][24][25][26], versus 29 months (95% CI, 19 months to NR) for those who did not receive frontline platinum-based therapy (N = 14). Of the 14 patients who did not receive platinum-based therapy in the first line, 10 did receive it in the second line, and favorable responses were noted (a PR was noted in 7 patients and SD in 1 patient as the best response), with a median duration of therapy of 11 months (range, . Notably, amongst those with metastatic disease who did receive frontline platinum-based therapy, the mOS for patients with biallelic status (N = 39) was 26 months (95% CI, 20-52 months) and 8.66 months (95% CI, 6.2-NR) for those with monoallelic status (N = 4). Beyond the advanced PDAC setting, several studies have also demonstrated an advantage associated with the receipt of platinum-based therapy in the neoadjuvant setting for patients with BRCA1/2-mutated PDAC [42,43].
To enhance response and potentially delay resistance, O'Reilly et al. evaluated the role of veliparib in combination with cisplatin and gemcitabine in untreated metastatic and locally advanced PDAC with BRCA1/2 or PALB2 mutations [44]. Patients were randomized to receive cisplatin and gemcitabine, with or without veliparib. Delayed emergence of resistance and improved survival were not observed with the addition of veliparib; however, high response rates in both groups were observed (74.1% in the experimental arm and 65.2% in the control arm, p = 0.55), as well as a favorable OS signal from first-line cisplatin-based therapy in both arms (15.5 months and 16.4 months, respectively, p = 0.73), establishing cisplatin and gemcitabine as a standard approach in gBRCA1/2-and PALB2mutated PDAC and an alternative to FOLFIRINOX.

Poly-ADP Ribose Polymerase Inhibitors
The PARP enzymes have a key role in base excision repair of single-stranded DNA breaks [45], and inhibition of these enzymes results in the accumulation of single-stranded breaks (SSBs), leading to replication fork collapse, and the generation of double-stranded breaks (DSBs). As discussed earlier, DSBs rely on HRR for repair and to avoid cell cycle arrest and apoptosis [46]. Thus, in patients with mutations in HRR genes, PARPi can be employed to create a state of synthetic lethality in tumors [46]. Select recently-completed and ongoing trials evaluating the role of PARPi for BRCA-mutated and HRD PDAC are summarized in Tables 2 and 3. In a multicenter phase 2 study of patients with breast, prostate, ovarian, and PDAC tumors with gBRCA1/2 mutations, the use of oral PARPi olaparib demonstrated an overall response rate of 26.2% (95% CA 21.3-31.6) across all patients [47]. Specifically, in the PDAC cohort, which included 23 patients, all with advanced disease, an overall response rate of 21.7% was observed, including one complete response and four partial responses.    Olaparib was prospectively evaluated in the phase III Pancreas Olaparib Ongoing (POLO) trial, which enrolled 154 patients with gBRCA1/2 mutations and advanced or metastatic PDAC [5,6]. Patients were required to have demonstrated disease response or stability following a minimum of 16 weeks of platinum-based chemotherapy. The primary end point of PFS was met, with a mPFS of 7.4 months in the olaparib arm compared with 3.8 months in the placebo arm (HR 0.53, p = 0.004). Following this data, FDA approval for olaparib was granted in December 2019. The secondary endpoint of OS, however, was not met (median 19.0 months versus 19.2 months; HR 0.83, 95% CI, 0.56-1.22, p = 0.3487). OS at 36 months was 33.9% for olaparib and 17.8% for placebo, and mPFS2, defined as median time from randomization to second disease progression, was 16.9 months for olaparib versus 9.3 months for placebo (hazard ratio, 0.66; p = 0.0613), indicating a trend toward benefit for olaparib, though not reaching statistical significance. Notably, quality of life scores between the arms were equivalent. POLO was the first trial to establish a biomarker-based approach to the management of PDAC in patients with gBRCA mutations; however, the lack of OS advantage and the placebo control arm have led to questions regarding the magnitude and importance of the observed PFS benefit.
Reiss et al. conducted a phase 2 trial evaluating PARPi rucaparib in the post-platinum maintenance setting in patients with PDAC harboring germline or somatic BRCA1/2 or PALB2 alterations [48]. They reported a mPFS of 13.1 months (95% CI, 4.4-21.8) and a mOS of 23.5 months (95% CI, [20][21][22][23][24][25][26][27]. An ORR of 41.7% was observed, which included 3 complete responses and 12 partial responses. Rucaparib was also evaluated in the phase 2 RUCA-PANC study, which included patients with germline or somatic BRCA1/2 mutations [49]. While two-thirds of patients had received prior platinum, platinum sensitivity and prior platinum were not mandated. The study was terminated early based on a modest ORR of 15.8%, and notably, none of the patients who had platinum-resistant disease at the time of enrollment had an objective response. Indeed, sensitivity to platinum has emerged as a key predictive biomarker for sensitivity to PARPi. In a phase 2 study of veliparib in 16 patients with metastatic or locally advanced PDAC and gBRCA1/2 or PALB2 mutations, Lowery et al. reported no objective response, although stable disease for >4 months was noted in four patients [50]. Most patients enrolled in this study had platinum-resistant PDAC, which most likely explains the limited signal observed. Beyond gBRCA mutations, Javle and colleagues described the clinical outcomes of olaparib monotherapy in patients with PDAC and alterations in other DDR genes, including ATM, PALB2, ARID1A, sBRCA, PTEN, RAD51, CCNE, and FANCB. They identified that patients with platinum-sensitive PDAC had an improved mPFS (4.1 months, 95% CI, 3.6-7.8) over those with platinum-resistant PDAC (2.2 months, 95% CI, from 1.8 to not reached, p = 0.01) [51]. This benefit was also observed in the analysis of OS (10.5 vs. 5.4 months for platinum-sensitive versus platinum-resistant disease, respectively, p = 0.03). Based on this study, in addition to others, it is apparent that in PDAC, the role of PARPi is maximized in the platinum-sensitive maintenance setting.
Beyond their role in advanced disease, the utility of PARPi in the early-stage setting is under investigation in clinical trials. The APOLLO study (NCT04858334) is a randomized phase II study of adjuvant olaparib versus placebo in patients with pancreatic cancer and germline or somatic BRCA1/2 or PALB2 mutations in whom all curative intent standard treatment has been completed [52].

Resistance to PARP Inhibition
Despite the rationale and enthusiasm for the use of PARPi in patients with BRCA1/2 mutations (and other DDR gene alterations), primary resistance is observed in a notable proportion of patients, including those with platinum-sensitive PDAC. In the POLO trial, despite all patients enrolled having confirmed platinum sensitivity in the context of gBRCA1/2 mutations, approximately one fifth of patients had evidence of progression of disease on first assessment [5,6]. Similarly, in the phase II study by Reiss et al. evaluating maintenance rucaparib in the post-platinum maintenance setting for patients with germline or somatic BRCA1/2 or PALB2 pathogenic variants, 16% of patients experienced progression of disease within the first 8 weeks of treatment [48]. This builds upon the data from Golan and colleagues, in which patients with PDAC and gBRCA1/2 mutations had HRDetect performed on their pancreatic cancer tumor samples, and 12% did not have a HRD signature [15], which may account for primary resistance to PARPi in a proportion of patients.
Secondary acquired resistance to PARPi is common, and several mechanisms are postulated [53]. In several studies, hyperactivation of the ATR/CHK1 pathway, resulting in maintained genomic stability, has been suggested as a mechanism of PARPi resistance in BRCA-driven tumors [54,55], and forms the basis of several efforts to overcome resistance. This pathway is discussed in greater detail below. In addition, the emergence of reversion mutations, resulting in the restoration of HRR following treatment with PARPi, has been described. In a study by Pettitt et al., data from the literature pertaining to 308 reversion mutations in the setting of PARPi or platinum resistance identified in 91 patients were reported and suggest that for BRCA2 mutations, the position of the mutation may affect the risk of reversion [56]. In addition, they concluded that many reversions are predicted to encode highly immunogenic neopeptides, which may provide a potential opportunity for avoiding resistance, for example, through the integration of immunotherapy. Specific to PDAC, preclinical models of ATM-deficient pancreatic cancer cells identified an association between the upregulation of alternative-end joining via upregulation of drug efflux transporters and detoxication enzymes and resistance to PARP inhibition [57]. However, the mechanisms of PARPi resistance in pancreatic cancer appear complex and, as of yet, are less well defined [58]. Ongoing efforts to deepen the responses to PARPi as well as to delay resistance are underway, with several combination strategies under investigation in PDAC ( Table 2).

PARP Inhibition and Chemotherapy
As discussed earlier, our group evaluated veliparib in combination with cisplatin and gemcitabine in untreated metastatic and locally advanced PDAC with BRCA1/2 or PALB2 mutations in a randomized phase 2 study [43] and demonstrated that while a favorable overall survival signal from first-line cisplatin and gemcitabine chemotherapy (+/− veliparib) was observed in both arms (15.5 months and 16.4 months, respectively, p = 0.73), improved survival and delayed resistance were not observed with the addition of veliparib in this trial. The SWOG S1513 trial was a randomized phase 2 study that evaluated veliparib with or without FOLFIRI in the second line for patients with stage IV PDAC [59]. Accrual was halted early, with 123 patients enrolled, due to the lack of benefit from the addition of veliparib. For the entire group, mOS was not improved with the addition of veliparib (5.4 vs. 6.5 months, HR 1.23, p = 0.28). Notably, for patients with HR-DDR defects (N = 22, 19%) compared to those without HR-DDR gene defects, mPFS and mOS were 7.3 vs. 2.5 months (p = 0.05) and 10.1 vs. 5.9 months (p = 0.17), respectively, with FOLFIRI alone, and 2.0 vs. 2.1 months (p = 0.62) and 7.4 vs. 5.1 months (p = 0.10), respectively, with veliparib plus mFOLFIRI. Both arms performed favorably compared with the overall group, indicating a potential role for irinotecan-based therapy in patients with such defects. In addition, an ongoing non-randomized phase 2 study is evaluating 5-FU and liposomal irinotecan in combination with rucaparib in patients with metastatic PDAC with BRCA1/2, PALB2, or other DDR-HRD genomic alterations who have not received systemic therapy in the metastatic setting (NCT03337087).

PARPi and Immune Checkpoint Inhibitors
In pre-clinical studies, PARPi have been demonstrated to enhance cancer-associated immunosuppression and upregulate PD-L1 expression [60]. In addition, the combination of PARPi and immune checkpoint blockade has demonstrated an accumulation of tumor neoantigens and activation of interferon pathways, resulting in a synergistic antitumoral effect [61,62] in pre-clinical studies. The phase 1b/2 PARPVAX study of niraparib plus nivolumab and niraparib plus ipilimumab in patients with platinum-sensitive PDAC was evaluated and compared to a historical reference [63]. Subgroup analysis of patients with BRCA1/2 or PALB2 variants demonstrated a mPFS of 3.7 months and a mOS of 12.2 months in the niraparib and nivolumab arm, versus a prolonged mPFS of 10.4 months and a mOS of 38 months in the niraparib plus ipilimumab arm, suggesting that CTLA-4 inhibitors may potentially have an enhanced benefit in combination with PARPi compared with PD-1 inhibitors, and this signal warrants further investigation.
The non-randomized phase 2 POLAR study (NCT04666740) is evaluating the combination of maintenance pembrolizumab and olaparib in patients with metastatic pancreatic cancer and HR deficiency (or exceptional response to platinum) post-platinum-based chemotherapy. The SWOG S2001 trial (NCT04548752) is a randomized study evaluating olaparib with/without pembrolizumab, also in the post-platinum maintenance setting, in patients with metastatic pancreatic cancer and gBRCA1/2 mutations. Several studies are evaluating PARPi and immunotherapy +/− other agents, for example, PD-1 inhibitor dostarlimab and PARPi niraparib in patients with advanced BRCA-mutated tumors, including PDAC (NCT04673448, NCT04493060), dostarlimab, niraparib, and radiotherapy (NCI04409002), PD-L1 inhibitor avelumab, PARPi talazoparib, and MEK inhibitor binimetinib in patients with locally advanced or metastatic RAS-mutated solid tumors (NCT03637491, closed), Beyond BRCA1/2 variants, the MAZEPPA trial is evaluating PARPi of olaparib or kinase inhibitor selumetinib plus PD-1 inhibitor durvalumab in patients with metastatic PDAC and BRCAness (NCT04348045).

PARPi and Anti-Angiogenic Agents
Agents inhibiting vascular epidermal growth factor (VEGF) families result in a hypoxic state and subsequent downregulation of HRR gene expression [64]. Series suggest that this hypoxic tumor microenvironment results in the functional inactivation of RAD51 and BRCA without a genetic alteration in these genes, with a resulting BRCAness phenotype [65,66]. Some promise has been observed in combining PARPi and VEGF inhibitors in both ovarian cancer and prostate cancer. A randomized phase 2 study of olaparib with or without the pan-VEGF inhibitor cediranib demonstrated a significant radiographic PFS advantage over olaparib alone in men with metastatic castrate-resistant prostate cancer [67]. In ovarian cancer, the phase 3 PAOLA-1/ENGOT-ov25 trial of olaparib and bevacizumab in the postplatinum maintenance setting demonstrated a significant PFS advantage in those with BRCA and non-BRCA HRD mutations [32]. Olaparib combined with the anti-VEGF agent cediranib is under investigation in advanced solid tumors, including PDAC, in an ongoing phase 2 study (NCT02498613).

PARPi and Other Agents
ATR kinase-mediated DDR pathways are thought to promote tumor survival during PARP inhibition, and in pre-clinical studies, the ATR-inhibitor AZD6738 (ceralasertib) in combination with olaparib has demonstrated synergistic tumor inhibition [68]. A phase 2 study is currently underway investigating the role of AZD6738 with or without olaparib in patients with advanced solid tumors, including PDAC, who harbor ARID1A expression or ATM loss/mutations (NCT03682289). In addition, pre-clinical data from one study demonstrated that cells resistant to PARP inhibition showed evidence of elevated RAS/MAP kinase signaling, and a signal for partial reversal with the addition of MEK inhibition was observed [69]. On this basis, PARPi and MEK inhibition are being investigated both in the neoadjuvant (NCT04005690) and advanced settings (NCT03637491).

Immune Checkpoint Blockade in BRCA and HRD Pancreatic Cancer
PDAC is considered a prototypical immunogenically "cold" tumor, owing to an inherently immunosuppressive, hypoxic tumor microenvironment as well as a dense surrounding stroma. Outside of the small proportion of patients with microsatellite-instable PDAC (approximately 1%), a role for immune checkpoint blockade has been challenging to support, with poor responses observed in a phase 2 study from our group [70]. However, greater genomic instability in patients with PDAC with biallelic loss of BRCA1 and BRCA2 has been demonstrated across several series [14,71], suggesting a possible role for the incorporation of immune checkpoint blockade in this subset of patients. Terraro et al. reported the results of a series of 12 patients with chemotherapy-refractory pancreaticobiliary cancers and pathogenic germline variants in HRD genes (specifically BRCA1, RAD51C, ATM, BRCA2, and RAD51D) who received dual checkpoint blockade with ipilimumab and nivolumab. Of the ten patients with PDAC (out of a total of 12; 10 PDAC, 1 cholangiocarcinoma, and 1 ampullary carcinoma), two had a complete response to therapy, one had a partial response, and two had stable disease [72], indicating a potential role for immune checkpoint blockade as a therapeutic strategy in BRCA-mutated and other HRD PDAC.

Targeting DNA Replication Stress
While much of the focus of novel therapeutic strategies for PDAC has been on targeting HR, defective DDR may result in "replication stress", defined as the perturbation of errorfree DNA replication and the slowing of DNA synthesis, resulting in genomic instability and oncogenic transformation [73]. The ATR kinase has an important role in the cellular response to replication stress [74]. Replication protein A (RPA) binds single-stranded DNA (ssDNA) and recruits ATR, which results in the downstream activation of ATRinteracting proteins and subsequently ATR and CHEK1 through phosphorylation [74]. ATR-CHEK1 activation results in s-G2 cell cycle arrest and activation of HR repair, as well as replication fork stabilization [73,74]. CHEK1 is negatively regulated by WEE1 and MYT1 via phosphorylation and results in the activation of WEE1 and degradation of CDC25A [75,76]. Attempts at targeting replication stress have led to the development of several compounds, many of which are in preclinical and clinical development, including ATR inhibitors (ATRi), CHEK1 inhibitors, and WEE1 inhibitors.
ATRi elimusertib was evaluated as a monotherapy in a phase 1 trial of 21 patients with advanced solid tumors (NCT03188965). An ORR of 19% (4 of 21) was observed, all 4 occurring in patients harboring alterations in or loss of ATM [77]. Intravenous ATRi berzosertib was evaluated as a monotherapy and in combination with carboplatin in 40 patients with advanced solid tumors (NCT02157792). One patient with metastatic colorectal cancer harboring ATM loss and an ARID1A mutation had a complete response; another patient with platinum-refractory and PARPi-resistant ovarian cancer harboring a gBRCA1 mutation had a partial response [78]. CHEK1/2i prexasertib has been evaluated as a monotherapy in two studies. In a phase 1 study (NCT01115790) of 45 patients with advanced solid tumors, 2 patients had a partial response, and 15 patients had stable disease as the best overall response, with a more favorable signal in patients with squamous histology. Thus, an expansion cohort, including patients with squamous histology only, was carried out and demonstrated partial responses in 15% of patients with anal cancer and 5% of patients with head and neck tumors [79]. In the second study (NCT02203513), a phase 2 study, 42 patients with ovarian cancer were enrolled. Of 24 patients without BRCA1/2 mutations, an ORR of 33% was observed, and among 18 patients harboring BRCA1/2 mutations, an ORR of 11% was observed [80,81]. WEE1i adavosertib was evaluated in combination with chemotherapy (gemcitabine, paclitaxel, carboplatin, or pegylated liposomal doxorubicin) in a phase II study in 94 patients with primary platinum-resistant gynecologic malignancies (NCT02272790) [82]. A promising ORR of 32% was observed. ZN-C3 has been evaluated in a phase 1 study (NCT04158336) of patients with high-grade serous endometrial carcinoma and resulted in a complete response in one patient and two partial responses in eleven patients in total [83]. Several studies combining chemotherapy with inhibitors of the pathway are underway, including gemcitabine, carboplatin, and berzosertib in advanced ovarian cancer (NCT02627443), and gemcitabine, nab-paclitaxel, and adavosertib in PDAC (not specific to patients with HRD) (NCT02194829-active, not recruiting).

Other Promising Strategies in BRCA and HRD
Inhibitors of polymerase theta (POLθ, which is encoded by POLQ) are gaining interest as an additional therapeutic strategy in HRD malignancies. POLθ is the key enzyme in microhomology-mediated end-joining (MMEJ), and several studies have demonstrated the potential of POLθ inhibitors in evasion of MMEJ and inducing synthetic lethality [84,85]. In a key recent study in BRCA-mutated breast cancer and PDAC cell lines, either novobiocin, an inhibitor of the POLθ ATPase domain, or ART558, an inhibitor of the POLθ polymerase domain, demonstrated activation of the cGAS/STING pathway. Activation of this pathway drives the expression of type I interferon response elements, including PD-L1, and increases CD8+ T-cell tumor infiltration and activation, as well as the activation of antigen-presenting dendritic cells [86]. In addition, the antitumor activity of novobiocin was augmented with the addition of PD-1 blockade in a BRCA2-deficient mouse model in this study. The first human study of novobiocin will begin accruing patients with tumors with alterations in DNA repair genes in 2023 (NCT05687110). Current and emergent targeted strategies are summarized in Figure 1.

Conclusions and Future Directions
With modest progress in survival outcomes for patients with PDAC over the last two decades, the emergence of a personalized, biomarker-based approach to the management of PDAC is promising. Patients with BRCA1/2 mutations represent a biologically distinct subgroup with a favorable predictive profile compared with an unselected PDAC population. In addition, patients with HRD status are emerging as a group that may also benefit from the incorporation of a similar personalized approach. However, despite developments in the management of patients with BRCA1/2-and HRD-associated PDAC, resistance is a critical limitation. Future efforts should focus on the continued development of accurate HRD detection methods. In addition, well-designed, adequately powered, randomized clinical trials of novel combination approaches to deepen responses and overcome resistance are imperative. Lastly, the incorporation of translational, interdisciplinary science and, most importantly, universal patient access to somatic and germline testing as

Conclusions and Future Directions
With modest progress in survival outcomes for patients with PDAC over the last two decades, the emergence of a personalized, biomarker-based approach to the management of PDAC is promising. Patients with BRCA1/2 mutations represent a biologically distinct subgroup with a favorable predictive profile compared with an unselected PDAC population. In addition, patients with HRD status are emerging as a group that may also benefit from the incorporation of a similar personalized approach. However, despite developments in the management of patients with BRCA1/2and HRD-associated PDAC, resistance is a critical limitation. Future efforts should focus on the continued development of accurate HRD detection methods. In addition, well-designed, adequately powered, randomized clinical trials of novel combination approaches to deepen responses and overcome resistance are imperative. Lastly, the incorporation of translational, interdisciplinary science and, most importantly, universal patient access to somatic and germline testing as recommended by international guidelines should be prioritized.