Oncogenic Drivers and Therapeutic Vulnerabilities in KRAS Wild-Type Pancreatic Cancer

Abstract

Purpose: Approximately 8% to 10% of pancreatic ductal adenocarcinomas (PDAC) do not harbor mutations in KRAS. Understanding the unique molecular and clinical features of this subset of pancreatic cancer is important to guide patient stratification for clinical trials of molecularly targeted agents.

Experimental design: We analyzed a single-institution cohort of 795 exocrine pancreatic cancer cases (including 785 PDAC cases) with a targeted multigene sequencing panel and identified 73 patients (9.2%) with KRAS wild-type (WT) pancreatic cancer.

Results: Overall, 43.8% (32/73) of KRAS WT cases had evidence of an alternative driver of the MAPK pathway, including BRAF mutations and in-frame deletions and receptor tyrosine kinase fusions. Conversely, 56.2% of cases did not harbor a clear MAPK driver alteration, but 29.3% of these MAPK-negative KRAS WT cases (12/41) demonstrated activating alterations in other oncogenic drivers, such as GNAS, MYC, PIK3CA, and CTNNB1. We demonstrate potent efficacy of pan-RAF and MEK inhibition in patient-derived organoid models carrying BRAF in-frame deletions. Moreover, we demonstrate durable clinical benefit of targeted therapy in a patient harboring a KRAS WT tumor with a ROS1 fusion. Clinically, patients with KRAS WT tumors were significantly younger in age of onset (median age: 62.6 vs. 65.7 years; P = 0.037). SMAD4 mutations were associated with a particularly poor prognosis in KRAS WT cases.

Conclusions: This study defines the genomic underpinnings of KRAS WT pancreatic cancer and highlights potential therapeutic avenues for future investigation in molecularly directed clinical trials. See related commentary by Kato et al., p. 4527.

Figure 1.. KRAS WT pancreatic cancers harbor frequent BRAF alterations

(A) Mutational landscape of KRAS WT pancreatic cancer displays frequent oncogenic activating alterations in BRAF and GNAS. Inactivating alterations in TP53, CDKN2A, and SMAD4 appear disproportionately fewer compared to the entire pancreatic cancer cohort (TP53: 36% WT versus 73% Cohort; CDKN2A: 18% WT versus 33% Cohort; SMAD4: 11% WT versus 21% Cohort) (see Suppl. Fig. 2A). Signature ‘N/A’ indicates mutational signatures were not assessable on OncoPanel as some versions of the assay predate the use of the mutational signature algorithm. The alteration % reflects only samples with coverage of the relevant gene. Genes not covered in earlier OncoPanel versions on this plot include PBRM1 (v1) and MTAP (v1/v2). x-axis: Deidentified patient IDs. The two cases with red boxes represent BRAF in-frame deleted organoids used for drug testing in Figure 3. In-frame indels in BRAF refer to in-frame deletions (N486_P490del, n = 9) and duplications (T599dup, n = 1). PDAC = Pancreatic Adenocarcinoma, PACC = Pancreatic Acinar Cell Carcinoma, PASC = Pancreatic Adenosquamous Carcinoma, TMB = Tumor Mutational Burden (B) Gene-level comparison between KRAS WT (n = 73) and KRAS mutant (n = 722) cases reveals significant enrichment in BRAF, GNAS, and ARID2 alterations in the WT cohort as well as enrichment in TP53 alterations in the mutant cohort. Statistical significance was assessed using Fisher’s Exact Test with the Benjamini-Hochberg Procedure for p-value multiple hypothesis corrections. (C) Top, Most BRAF alterations found in KRAS WT pancreatic cancer are functionally Class II i.e., display constitutive dimer-dependent kinase activity. Bottom, Pathogenic BRAF alterations reside within the BRAF tyrosine kinase domain and in-frame deletions in the β3-αC loop are the most frequent (n = 9). (D) Class II are the most common functional class of BRAF alterations in KRAS WT pancreatic cancer (72%) unlike colorectal cancer and melanoma in which Class I BRAF alterations are predominant (77% and 75%, respectively). Percentages reflect proportion of BRAF alterations of a given class out of the total number of BRAF mutations. Data for melanoma (n = 242) and colorectal cancer (n = 304) are from the PROFILE cohort in cBioPortal.

Figure 2.. Alternative MAPK alterations including RTK fusions are frequent drivers in KRAS WT pancreatic cancer

(A) Exon/intron structure of predicted RTK fusion transcripts resulting from rearrangements in KRAS WT tumors detected via DNA-based targeted sequencing using OncoPanel (n = 5) or via RNA-based fusion assay ArcherDx (n = 3). Shown are relevant RTK functional domains as encoded across exons. (B) Immunohistochemistry with a pan-TRK antibody reveals diffuse membranous TRK staining supporting the presence of the detected TJAP1::NTRK1 fusion. (C) Left, ROS1 break-apart fluorescence in situ hybridization (FISH). In this system, separation of the normally fused red/green signal is indicative of a genomic alteration. The loss of red signal suggests deletion of the 5’ end of ROS1 (Exons 1–32) which was also captured by sequencing (see Fig. 4D). Right, ROS1 IHC displays ectopic ROS1 expression confirming the presence of the detected SLC4A4::ROS1 rearrangement. (D) Alternative (i.e., non-KRAS) MAPK alterations are identified in 32 out of 73 (44%) KRAS WT tumors with BRAF alterations and RTK fusions being the most frequent. Amplifications in other MAPK drivers including ERBB2, EGFR, KRAS and MET were also identified. Alternative MAPK drivers tend to be mutually exclusive pointing to a likely driver role. The alteration % reflects only samples with coverage of the relevant gene. Genes not covered in earlier OncoPanel versions on this plot include NRG1, RAC1, and RASA1 (v1/v2). PDAC = Pancreatic Adenocarcinoma, PACC = Pancreatic Acinar Cell Carcinoma, PASC = Pancreatic Adenosquamous Carcinoma, TMB = Tumor Mutational Burden (E) Pathway-level comparison of KRAS WT and mutant pancreatic cancer shows enrichment in alternative MAPK genes in the WT cohort. Statistical significance was assessed using Fisher’s Exact Test with the Benjamini-Hochberg Procedure for p-value multiple hypothesis corrections.

Figure 3.. Dual pan-RAF and MEK inhibition demonstrates synergistic inhibition of pancreatic cancer patient-derived organoids harboring BRAF in-frame deletions

(A) Two independent pancreatic cancer patient-derived organoids (PDOs) harboring BRAF in-frame deletions (p.N486_P490del) are resistant to the growth inhibitory effects of the BRAF^V600 inhibitor (dabrafenib) but remain sensitive to pan-RAF (LY3009120), MEK (trametinib), and ERK (BVD-523) inhibitors. Shown are growth rate-corrected dose sensitivity curves. (B) Combination of pan-RAF inhibitor (LY3009120) with MEK (trametinib) or ERK (BVD-523) inhibitors shows synergistic activity in the PANFR0172_T2 PDO line whereas no synergy is displayed by combining a pan-RAF (LY3009120) inhibitor with a BRAF^V600 inhibitor (dabrafenib). Growth inhibition with various combinations of drug concentrations is displayed as a sigmoidal curve (Top) or heatmap (Middle). (Bottom) Calculation of synergy between different compounds using the Lowe additivity model.

Figure 4.. Durable response to sequential ROS1 targeted therapy in a KRAS WT PDAC harboring a SLC4A4-ROS1 fusion

(A) Treatment history of index patient with a SLC4A4-ROS1 fusion who received prolonged benefit from targeted therapy with upfront crizotinib followed by cabozantinib. (B) Representative images with target lesions in the liver and lung (yellow arrow) showing initial response to crizotinib (Day 61) followed by disease progression at Day 234. Target lesions display overall stability between Days 234 to 361 while the patient was receiving cabozantinib. (C) Longitudinal monitoring of 5 target lesions used to determine disease response while on ROS1-directed therapy. (D) Genome-level copy number profiles from initial diagnostic (Top) and crizotinib-resistant (Day 234, Middle) samples both show chromosome 6q deletion (arrowhead) leading to loss of the 5’ end of ROS1, indicative of a likely genomic rearrangement involving ROS1. (Bottom) Magnified view of the ROS1 Exon 1–32 deletion.

Figure 5.. Impact of genomic alterations on pancreatic cancer survival

(A) Increased number of core pancreatic cancer mutations (KRAS, TP53, CDKN2A, SMAD4) is associated with significantly worse overall survival in the entire pancreatic cancer cohort by Kaplan-Meier analysis. P-value calculated by the log-rank test. (B) KRAS WT pancreatic cancer is associated with a significantly improved overall survival compared to KRAS mutant pancreatic cancer by Kaplan-Meier analysis. P-value calculated by the log-rank test. (C-D) Association of clinical and genomic features with overall survival in patients with KRAS WT (C) and KRAS mutant (D) pancreatic cancers using Cox proportional-hazards regression. P-values calculated by the Cox proportional-hazards model, with significant p-values are annotated with a star. (E) Precision therapy landscape for KRAS WT pancreatic cancer. Summary of targetable genomic alterations detected in KRAS WT pancreatic cancer and potential therapeutic options for genomic subgroups. i = Inhibitor