Gene fusions involving receptor tyrosine kinases (RTKs) define an important class of genomic alterations with many successful targeted therapies now approved for ALK, ROS1, RET and NTRK gene fusions. Fusions involving the ERBB family of RTKs have been sporadically reported, but their frequency has not yet been comprehensively analyzed and functional characterization is lacking on many types of ERBB fusions. We analyzed tumor samples submitted to Caris Life Sciences (n=64,354), as well as the TCGA (n=10,967), MSK IMPACT (n=10,945) and AACR GENIE (n=96,324) databases for evidence of EGFR, ERBB2 and ERBB4 gene fusions. We also expressed several novel fusions in cancer cell lines and analyzed their response to EGFR and HER2 tyrosine kinase inhibitors (TKIs). In total, we identified 1,251 ERBB family fusions, representing an incidence of approximately 0.7% across all cancer types. EGFR, ERBB2, and ERBB4 fusions were most frequently found in glioblastoma, breast cancer and ovarian cancer, respectively. We modeled two novel types of EGFR and ERBB2 fusions, one with a tethered kinase domain and the other with a tethered adapter protein. Specifically, we expressed EGFR-ERBB4, EGFR-SHC1, ERBB2-GRB7 and ERBB2-SHC1, in cancer cell lines and demonstrated that they are oncogenic, regulate downstream signaling and are sensitive to small molecule inhibition with EGFR and HER2 TKIs. We found that ERBB fusions are recurrent mutations that occur across multiple cancer types. We also establish that adapter-tethered and kinase-tethered fusions are oncogenic and can be inhibited with EGFR or HER2 inhibitors. We further propose a nomenclature system to categorize these fusions into several functional classes.
<p>Association of B7-H3 expression and common hallmark mRNA signatures. GSEA with functional oncogenic pathways on several tumor types with high and low B7-H3 expression. Normalized enrichment scores are displayed. FDR < 0.05 based on values between −3 and 3. Boxes marked with “X” indicate that the enrichment of the pathway did not meet thresholds of significance in the tumor type.</p>
Abstract Purpose: Estrogen receptor 1 ( ESR1 ) mutations and fusions typically arise in patients with hormone receptor-positive breast cancer after aromatase inhibitor therapy, whereby ESR1 is constitutively activated in a ligand-independent manner. These variants can impact treatment response. Herein, we characterize ESR1 variants among molecularly profiled breast cancers. Methods: DNA next-generation sequencing (592-gene panel) data from 9860 breast cancer samples were retrospectively reviewed. Gene fusions were detected using the ArcherDx fusion assay or whole transcriptome sequencing (N = 344 and N = 4305, respectively). Statistical analyses included Chi-square and Fisher’s exact tests. Results: An ESR1 ligand binding domain (LBD) mutation was detected in 8.6% of tumors evaluated and a pathogenic ESR1 fusion was detected in 1.6%. The majority of ESR1 LBD mutations/fusions were from estrogen receptor (ER)-positive samples (20.1% and 4.9%, respectively). The most common ESR1 LBD mutations included D538G (3.3%), Y537S (2.3%), and E380Q (1.1%) mutations. Among biopsy sites, ESR1 LBD mutations were most observed in liver metastases. Pathogenic ESR1 fusions were identified in 76 samples (1.6%) with 40 unique fusion partners. ESR1 variant (mutation/fusion) samples more frequently expressed androgen receptor (78.0% vs 58.6, p<0.0001) and less frequently immune checkpoint proteins than ESR1 wild-type (PD-1 20.0% vs 53.4, p<0.05; immune cell PD-L1 10.0% vs 30.2, p<0.0001). Conclusion: We have described one of the largest series of ESR1 fusions reported. ESR1 LBD mutations were commonly identified in ER-positive disease. Limited data exists regarding the clinical impact of ESR1 fusions, which could be an area for future therapeutic exploration.
Abstract ARID genes encode subunits of SWI/SNF chromatin remodeling complexes and are frequently mutated in human cancers. We investigated the correlation between ARID mutations, molecular features, and clinical outcomes in melanoma patients. Cutaneous melanoma samples (n = 1577) were analyzed by next-generation sequencing. Samples were stratified by pathogenic/likely pathogenic mutation in ARID genes ( ARID1A/2/1B/5B ). PD-L1 expression was assessed using IHC (SP142; positive (+): ≥ 1%). Tumor mutation burden (TMB)-high was defined as ≥ 10 mutations/Mb. Transcriptomic signatures predictive of response to immune checkpoint inhibitors—interferon gamma and T-cell inflamed score were calculated. Real-world overall survival (OS) information was obtained from insurance claims data, with Kaplan–Meier estimates calculated from time of tissue collection until last date of contact. Mann–Whitney U, Chi-square, and Fisher exact tests were applied where appropriate, with p values adjusted for multiple comparisons. ARID2 mutations were more prevalent in cutaneous melanoma compared to ARID1A (11.0%: n = 451 vs 2.8%: n = 113), with concurrent ARID1A / ARID2 mutation in 1.1% (n = 46) of samples. ARID mutations were associated with a high prevalence of RAS pathway mutations— NF1 ( ARID1A , 52.6%; ARID2 , 48.5%; ARID1A/2 , 63.6%; and ARID-WT, 13.3%; p < 0.0001) and KRAS ( ARID1A , 3.5%; ARID2 , 3.1%; ARID1A/2 , 6.5%; and ARID-WT, 1.0%; p = 0.018)), although BRAF mutations were less common in ARID-mutated cohorts ( ARID1A , 31.9%; ARID2 , 35.6%; ARID1A/2 , 26.1%; and ARID-WT, 50.4%; p < 0.0001). TMB-high was more common in ARID-mutated samples ( ARID1A , 80.9%; ARID2 , 89.9%; ARID1A/2 , 100%; and ARID-WT, 49.4%; p < 0.0001), while PD-L1 positivity was similar across subgroups ( ARID1A , 43.8%; ARID2 , 51.1%; ARID1A/2 , 52.5%; and ARID-WT, 44.9%; p = 0.109). Patients with ARID1A mutations had a higher prevalence of dMMR/MSI-H compared to those with ARID-WT (2.7% vs 0.2%, p = 0.030). Median IFN-γ and T-cell signatures were higher in ARID2 -mutated samples compared to ARID-WT (IFN-γ: − 0.15 vs − 0.21, p = 0.0066; T-cell: 23.5 vs − 18.5, p = 0.041). ARID2- mutated patients had improved survival compared to ARID-WT; (HR: 1.22 (95% CI 1.0–1.5), p = 0.022). No additional OS benefit was observed with anti-PD-1 therapy for ARID2 mutation compared to ARID -WT. Melanoma patients with ARID mutations exhibited higher prevalence of markers associated with ICI response, including TMB-H, and immune-related signatures. Our data also suggests improved survival outcome in patients with ARID2 mutations, irrespective of anti-PD1 therapy.
Abstract Metastatic colorectal cancer (CRC) is a deadly disease with a 13% survival rate. CRC is often treated with chemotherapeutic agents 5-fluorouracil, irinotecan, and oxaliplatin, but predictive biomarkers are limited. We hypothesized that tumor expression of genes related to extent of drug exposure, stratified by p53 status, is associated with clinical outcomes on these regimens. To this end, we evaluated the prognostic and predictive significance of genes within p53-independent, p53-dependent, pan-drug, and drug-specific gene signatures established in CRC cells treated with 5-FU, irinotecan, or oxaliplatin. CRC patient samples were analyzed by DNA/RNA next-generation sequencing at Caris Life Sciences, and real-world survival outcomes were inferred from insurance claims data and Kaplan-Meier estimates. Samples with benign or no TP53 mutations detected (wt) (n=2983) or TP53 loss-of-function (LOF) mutations (n=6229) were stratified (high/low) by median expression of signature genes for comparison. Overall prognostic and predictive values in response to 5-FU, irinotecan, or oxaliplatin, regardless of line of therapy or concurrent therapies, or first-line FOLFOX were evaluated to estimate the contribution of each drug. From signatures established in CRC cell lines, a prognostic effect was observed for genes in p53 wt (n=19) and LOF (n=22) subgroups, with a similar survival effect observed for several genes (n=16). Both prognostic and non-prognostic gene expression had a significant effect on survival outcomes following specific drug treatments. Predictive genes included BTG2, a p53-dependent gene upregulated by 5-FU, irinotecan, and oxaliplatin in CRC cells that predicted better outcomes when expressed at high levels in p53 wt patients. Several genes in drug-specific signatures were predictive in a drug-specific manner including RABGAP1L and RAD51B (irinotecan), and SAT1 (oxaliplatin), suggesting their particular relevance for patients receiving these drugs. For some genes such as RABGAP1L and RAD51B, drug treatment downregulated gene expression but high expression predicted better survival in patients, suggesting benefit of combination therapy to limit such effects. From a real-world cohort of p53 wt and LOF CRC tumors, we demonstrate the prognostic and predictive potential of candidate gene expression in response to commonly used chemotherapeutics. Accompanying data on drug regulation of predictive genes provides valuable information in evaluating clinical impact of pan-drug or drug-specific mechanisms. Future directions include associations with clinically relevant features (microsatellite instability, tumor sidedness, and age), establishment of multi-gene signatures to improve clinical utility, and evaluation of other regimens including FOLFIRI and FOLFOXIRI. Citation Format: Lindsey Carlsen, Andrew Elliott, Marzia Capelletti, Micheal Hall, Philip A. Philip, Heinz-Josef Lenz, Howard Safran, Khaldoun Almhanna, Rimini Breakstone, Alexander G. Raufi, Emil Lou, John L. Marshall, W. Michael Korn, Wafik S. El-Deiry. Prognostic and predictive drug-induced gene signatures for colorectal cancer patients personalized based on p53 status and treatment with FOLFOX, 5-FU, oxaliplatin, or irinotecan [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 1231.
161 Background: Presentation of neoantigens, short peptides derived from tumor-specific somatic mutations, by HLA molecules triggers the activation of anti-tumor immune response. Recognition of neoantigens by autologous T cells promotes sensitivity to immune checkpoint inhibitors (ICI) in mismatch repair deficient (MMRd)/microsatellite instability high (MSI-H) tumors. Neoantigen-targeted reactivity has also been reported in microsatellite stable (MSS) tumors, and neoantigens are emerging targets for tumor vaccines and adoptive cell therapy. Here, we aimed to comprehensively assess the spectrum of immunogenic neoantigens in CRC to identify new subpopulations for these therapeutic strategies. Methods: 7,053 CRC tested at Caris Life Sciences (Phoenix, AZ) with NextGen Sequencing on DNA (720-gene panel) and RNA (whole transcriptome) were analyzed. MSI status was determined by immunohistochemistry of MMR proteins and/or NGS. Immune epitope prediction was performed on translated peptide sequences harboring detected mutations using the NetMHCpan v4.0 method in the Immune Epitope Database, with HLA genotyping performed using arcasHLA. Immune/stromal cell abundance in the tumor microenvironment (TME) was estimated from RNA expression profiles using MCP-Counter. Gene expression profiles were analyzed for a transcriptional signature predictive of response to immunotherapy (T cell-inflamed signature, TIS). Results: 6,600,005 unique peptide:allele interactions were tested, of which 368,468 (5.6%) were predicted neoantigens with patient-specific HLA alleles, including 16,184 (0.25%) with high affinity. MMRd/MSI-H tumors (6.9%) had higher neoantigen load at all affinity levels compared to MSS, with 10,243 (63.3%) and 5,621 (34.7%) of the predicted high affinity neoantigens exclusively associated with MSS and MMRd/MSI-H, respectively. 32 recurrent high affinity neoantigens were identified in > 10 samples, most frequently derived from APC mutations in MSS tumors (78% of recurrent neoantigens), and from RNF43 (25%), ASXL1 (20%), and KMT2D (15%) in MMRd/MSI-H. TIS scores positively correlated with the abundance of various immune cell populations in the TME, notably cytotoxic lymphocytes (r = 0.46). Among the recurrent high affinity neoantigens, those with the highest mean TIS scores (> 90 th percentile) resulted from mutation of MSH3 (K383fs, 5.1%) and KMT2D (multiple variants, 2.8%) in MMRd/MSI-H CRC, while those in MSS CRC derived from SOX9 (multiple variants, 0.2%) and APC (multiple variants, 0.7%) but were not associated with increased mean TIS scores (~55 th percentile). Conclusions: This is one of the largest studies to investigate the landscape of immunogenic neoantigens in CRC. We were able to identify candidate recurrent peptides with high HLA binding affinity and an association with a positive TIS signature supporting the role of neoantigens as potential cancer immunotherapy targets.
