Tumor Genome Research Articles

Abstract 5377: Investigating human leukocyte antigen (HLA) class 1 leukemogenic pathways in acute myelogenous leukemia

Abstract Background: Aging, smoking and germline predisposition are known to accelerate leukemogenesis. However, the the ability for myeloid clonal pathogenic variants (MCPV) to elicit an antigen-driven host antileukemia immune response is not well characterized. A better understanding on how cellular immunity selectively discriminates leukemia-related “neopeptides” could uncover mechanisms supporting tumor genome evolution. Methods: 122 patients with acute myelogenous leukemia (AML) were included and 111 had available protein coding sequences. A neopeptide library was created using protein FASTA coding extracted for individual myeloid clonal mutations via National Center for Biotechnology Information (NCBI). Patient class 1 HLA typing was obtained through Tempus, and Genomic Testing Cooperative laboratories. Prediction of number and strength of AML neopeptide bindings to MHC class I molecules were obtained via NetMHCPan-1. Analysis was performed with K-means algorithm assigning age at diagnosis, detected MCPV, number of neopeptides binding (as continuous variable), overall survival (OS) and event-free survival (EFS) as variables. Unsupervised data input allowed 3 cluster partitions for differential survivals allowing HLA class 1 peptide binding aggregation with potential roles in leukemogenesis and patient outcome. Descriptive statistics were performed with SAS. Results: Sixty-six different MCPV were detected. Of 111 patients, 70 (63%) with available MCPV produced at least 1 MHC I neopeptide binding. Frequently observed variants were P53 (13%), FLT3 (9.2%), WT1 (7%), DNMT3A (6.5%), RAS (5%), ASXL1 (4.3%), RUNX1 (4.2%), NPM1 (4.2%) and IDH1/2 (2.6%). Cluster 2 demonstrated a lack of favorable and higher proportion of adverse ELN 2022 subgroups. Cluster 3 exhibited a higher of P53 [p=0.03] and MDS-related mutations. OS was 79, 239, and 1551 days for cluster 2,1 and 3, respectively [p=0.0001]. EFS was 134, 363 and 451 days for cluster 2,1 and 3, respectively [p=0.0001]. Interestingly, Cluster 2 and 3 assembled the highest probability for peptide binding [1 v 2 v 3 = 1.76% v 36.6% v 62%, p=0.01]. Lastly, albeit in low frequencies, higher diversity of DNA repairing enzymes was seen in cluster 2. Conclusions: Different MCPV arising in AML genome are variably immunogenic and leukemia-related neopeptide binding’s ability is differentially expressed in our 3 clusters. Frequent neopeptide bindings in clusters 2 and 3 may reflect higher AML immunogenicity resulting in genomic instability. It is possible that less “edited” genomes in cluster 1 preserve higher chemosensitivity resulting in better OS, while highly edited genomes in cluster 2 and 3 are less chemo-sensitive and circumvent host immunosurveillance. Further expansion of cluster 2 is necessary to better understand the role of DNA repairing enzymes in younger patients’ outcomes, especially when genomic instability is observed. Citation Format: Paul Sackstein, Sarah Mudra, Lacey Williams, Garrett Diltz, Rachel Zemel, Sravanti Teegavarapu, Martha Mims, Catherine Lai, Kimberley Doucette, Anne Renteria, Olga Timofeeva, Gustavo A. Rivero. Investigating human leukocyte antigen (HLA) class 1 leukemogenic pathways in acute myelogenous leukemia [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 5377.

Abstract 5610: Investigating the role of BRCA2 in the development of pediatric rhabdomyosarcoma

Abstract Introduction: In a large-scale exome sequencing study of pediatric rhabdomyosarcoma (RMS) patients, we identified a statistical enrichment of patients with germline variants in BRCA2 (Li, et al. JNCI, 2020), which prompted this study to assess the role of heterozygous BRCA2 variants in the development of pediatric RMS in both precursor cell and mouse models. Methods: We engineered BRCA2 loss of function (LoF) variants in lines of immortalized mesenchymal stem cells, a potential precursor cell to RMS, using CRISPR and Prime Editing (PE). CRISPR was used to induce frameshift alleles in Exons 10 and 11 and with PE, we created two RMS patient specific LOF variants in Exons 5 and 17. We generated a genetically engineered mouse model (GEMM) with heterozygous excision of Exon 11 of Brca2. This allele is now being crossed into two different previously established GEMMs of RMS: a myoblast-specific (MyoD1-Cre) conditional loss of p53, and a gain-of-function mutant p53, LSL-p53R172H (Nakahata, et al. Dis. Model. Mech., 2022), to assess changes in the pace of tumorigenesis and metastatic potential. Through an X01 agreement with the Gabriella Miller Kids First Program, we have submitted paired tumor-normal samples for additional whole-genome sequencing and RNA sequencing (tumor only) on >500 patients with RMS, including those with germline variants in BRCA2. Results: Using CRISPR, we generated cell lines heterozygous for the frameshift variants c.1345dup, c.5680del, and c.5680dup. With Prime Editing, we created heterozygous lines mimicking two specific variants detected in the RMS cohort, c.462_463del and c.7857G>A. Generation of a C57BL6 GEMM with a heterozygous removal of Exon 11 in Brca2 has been successful. No homozygous mice were obtained following 7 heterozygous crosses (50 pups), a departure from Mendelian ratios (p=0.0001) consistent with embryonic lethality. Heterozygous Brca2 mice derived from crosses into two RMS models are being monitored for tumor development. We have completed tumor-normal whole-genome and tumor RNA sequencing for 609 RMS patients, including 228 subjects not previously analyzed. We will directly assess whether there is evidence for mutation or loss of expression of the remaining BRCA2 allele in tumors and whether these tumor genomes exhibit signatures of DNA repair deficiency. Conclusions: The goal of this project is to determine the molecular basis for the enrichment in germline BRCA2 LOF variants in pediatric malignancies, including RMS. Ongoing experiments using the engineered heterozygous BRCA2 in vitro cell line models will assess genomic stability, homologous recombination proficiency, and DNA replication fork stability. Assessment of the effect of heterozygous Brca2 variants in GEMMs of RMS, and subsequent analysis of additional sequencing data, will further elucidate what role this gene may play in increasing risk or progression of pediatric cancers. Citation Format: P. Adam Weinstein, Ryan Zabriskie, Cem Ozdemir, Kyle M. Miller, Olga Medina-Martinez, Jason D. Heaney, Aniko Sabo, Philip J. Lupo, Sharon E. Plon. Investigating the role of BRCA2 in the development of pediatric rhabdomyosarcoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 5610.

Abstract 6924: Structural variant dynamics in melanoma: Unraveling tumor heterogeneity and evolution

Abstract Melanoma is characterized by significant intratumoral heterogeneity and complex evolutionary dynamics. This diversity in genomic alterations leads to the emergence of various subclonal populations within a single tumor. In our research, we established a new model system composed of 24 single-cell-derived clonal sublines (C1-C24), originating from the M4 melanoma model. This model was developed using a genetically engineered hepatocyte growth factor (HGF)-transgenic mouse. We employed Trisicell (Triple-toolkit for single-cell intratumor heterogeneity inference), a cutting-edge computational tool for scalable analysis of intratumor heterogeneity and evaluation based on single-cell RNA mutations. This enabled us to construct a phylogeny tree, revealing melanoma’s intricate branching evolutionary patterns. These patterns show ancestral clones evolving into genetically distinct subclones, which demonstrate varied phenotypic traits such as drug sensitivity or resistance, cellular plasticity, and immunogenicity. In our study, we conducted long-read sequencing on these clonal sublines in the phylogeny and identified structural variants (SVs) using Severus, a tool optimized for phasing in long-read sequencing. The types of SVs we discovered include deletions, insertions, amplifications, translocations, and inversions. We explored their roles in subclonal evolution, particularly focusing on how they disrupt genes and accumulate during melanoma progression. Our initial data from eleven sublines indicated a higher prevalence of ancestral SVs, shared by all sublines, compared to subline-specific SVs, representing later events. Notably, the individual sublines showed a higher rate of gene disruption by SVs, hinting at potential functional selection. Our analysis further revealed that SVs common to all sublines are linked with genes in key cell growth pathways, such as Rap1, Hippo, and calcium signaling pathways. In contrast, subline-specific SVs primarily affect genes involved in neurophysiological pathways, such as glutamatergic synapse and morphine addiction pathways. These findings suggest that different genes, associated with various pathways, are disrupted at distinct stages of melanoma progression, providing insights into the genetic factors that may predispose individual melanocytes to melanomagenesis. This methodology presents a comprehensive tool for characterizing tumor genomes and understanding their relationship with disease progression and therapy resistance. We are continuing our analysis to map SVs across the entire mutation-based phylogeny of all sublines, which will further our understanding of melanoma’s genomic landscape. Citation Format: Xiwen Cui, Ayse G. Keskus, Farid R. Mehrabadi, Salem Malikic, Mikhail Kolmogorov, Chi-Ping Day, Glenn Merlino, S. Cenk Sahinalp. Structural variant dynamics in melanoma: Unraveling tumor heterogeneity and evolution [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6924.

A novel tumor mutation-related long non-coding RNA signature for predicting overall survival and immunotherapy response in lung adenocarcinoma

BackgroundImmunotherapy has changed the treatment landscape for lung cancer. This study aims to construct a tumor mutation-related model that combines long non-coding RNA (lncRNA) expression levels and tumor mutation levels in tumor genomes to detect the possibilities of the lncRNA signature as an indicator for predicting the prognosis and response to immunotherapy in lung adenocarcinoma (LUAD). MethodsWe downloaded the tumor mutation profiles and RNA-seq expression database of LUAD from The Cancer Genome Atlas (TCGA). Differentially expressed lncRNAs were extracted based on the cumulative number of mutations. Cox regression analyses were used to identify the prognostic lncRNA signature, and the prognostic value of the five selected lncRNAs was validated by using survival analysis and the receiver operating characteristic (ROC) curve. We used qPCR to validate the expression of five selected lncRNAs between human lung epithelial and human lung adenocarcinoma cell lines. The ImmuCellAI, immunophenoscore (IPS) scores and Tumor Immune Dysfunction and Exclusion (TIDE) analyses were used to predict the response to immunotherapy for this mutation related lncRNA signature. ResultsA total of 162 lncRNAs were detected among the differentially expressed lncRNAs between the Tumor mutational burden (TMB)-high group and the TMB-low group. Then, five lncRNAs (PLAC4, LINC01116, LINC02163, MIR223HG, FAM83A-AS1) were identified as tumor mutation-related candidates for constructing the prognostic prediction model. Kaplan‒Meier curves showed that the overall survival of the low-risk group was significantly better than that of the high-risk group, and the results of the GSE50081 set were consistent. The expression levels of PD1, PD-L1 and CTLA4 in the low-risk group were higher than those in the high-risk group. The IPS scores and TIDE scores of patients in the low-risk group were significantly higher than those in the high-risk group. ConclusionOur findings demonstrated that the five lncRNAs (PLAC4, LINC01116, LINC02163, MIR223HG, FAM83A-AS1) were identified as candidates for constructing the tumor mutation-related model which may serve as an indicator of tumor mutation levels and have important implications for predicting the response to immunotherapy in LUAD.

Abstract 136: Determining the genomic basis of PAX3-FOXO1 mediated transformation in rhabdomyosarcoma

Abstract Rhabdomyosarcoma (RMS) is the most common pediatric soft tissue sarcoma. Despite rigorous scientific advances and clinical trials, the survival rate for high-risk RMS has not increased above 20% in the last three decades. A deeper understanding of the basic biology driving RMS tumorigenesis is needed to discover novel therapeutic targets. RMS can be divided into two major histologic subtypes: embryonal RMS (ERMS) and alveolar RMS (ARMS). ARMS accounts for 20% of pediatric cases and has poorer survival. The majority of ARMS tumors harbor t(2;13)(q35;q14) or t(1;13)(p36;q14) chromosomal translocations resulting in the PAX3-FOXO1 (P3F) and PAX7-FOXO1 (P7F) fusion oncoproteins, respectively. ARMS tumors that lack oncofusions more closely resemble ERMS. Classifying on fusion status more accurately describes the disease and highlights the significance of P3/7F. ARMS tumor genomes only contain an average of 6.4 somatic mutations suggesting the oncofusion protein is the key oncogenic driver and a potential candidate for therapeutic intervention. RMS is commonly thought to arise from skeletal muscle progenitor cells that fail to terminally differentiate. However, RMS occurs throughout the body in sites devoid of skeletal muscle, underscoring the potential for alternative cells of origin. Our lab established that endothelial progenitors are indeed a cell of origin for ARMS by generating a genetically engineered mouse model with endothelial cell specific P3F expression coupled with Cdkn2a loss. We further generated a human model system derived from TP53-null human induced pluripotent stem cells (iPSCs), in which we direct differentiation to endothelial cell fate and then force P3F expression. P3F expression during endothelial cell directed differentiation blocked endothelial maturation and instead reprogrammed cells shifting cell fate to skeletal muscle-like cells that form ARMS tumors in mice. Generation of iPSC-derived ARMS (iARMS) occurs over 15 days and provides a unique model to determine the kinetics and specific genomic contributions facilitated by P3F during RMS transformation. To investigate the genomic changes occurring during iARMS transformation I utilized CUT&RUN for P3F and H3K27ac to assess P3F occupancy and enhancer landscape, ATAC-seq to assess chromatin accessibility, and RNA-seq to assess gene expression changes at timepoints throughout the transformation event. These results define the fundamental genomic changes needed for transformation into FP-RMS by P3F and uncovers the key determinants of ARMS cell identity. The iARMS system provides a unique platform to uncover how P3F establishes and maintains RMS cell state and to dissect the specific dependencies required to maintain the ARMS cell state. Understanding the fundamental mechanism of P3F-driven tumorigenesis could provide greater resolution into the key disease determinants to focus pre-clinical efforts on pivotal targets. Citation Format: Bradley T. Stevens, Yang Zhang, Randolph K. Larsen, Grace E. Adkins, Jack D. Hopkins, Darden W. Kimbrough, Matthew R. Garcia, Brian J. Abraham, Mark E. Hatley. Determining the genomic basis of PAX3-FOXO1 mediated transformation in rhabdomyosarcoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 136.

Abstract 3664: Evaluation of on-treatment tumor molecular dynamics using plasma whole genome sequencing (pWGS) and whole exome sequencing (pWES) of circulating tumor DNA (ctDNA) in patients (pts) with metastatic non-small cell lung cancer (NSCLC): An exploratory analysis of KEYNOTE-598

Abstract Background: ctDNA-based pWES is a promising tool to examine tumor genomics using a non-invasive liquid biopsy to serially monitor changes during tumor evolution and treatment. Here, we evaluate on-treatment tumor molecular dynamics using ctDNA-based pWGS and pWES in pts with metastatic NSCLC that responded to pembrolizumab (pembro) monotherapy and later experienced PD in KEYNOTE-598 (NCT03302234). Methods: In KEYNOTE-598, pts had metastatic NSCLC (PD-L1 tumor proportion score ≥50%) and received first-line pembro ± ipilimumab for ≤35 cycles. Cell-free DNA (cfDNA) was isolated and sequenced from plasma samples at baseline (BL), during on-treatment response (cycle 5 [C5]), and at PD (discontinuation [D]) from pts with objective response to pembro monotherapy and subsequent PD. ctDNA burden (tumor fraction [TF]) was estimated using low-pass pWGS by quantifying tumor content in cfDNA as a function of aneuploidy. All samples with TF >1.5% underwent high-depth pWES. ctDNA burden was estimated from pWES data using maximum somatic allele frequency (MSAF). Correlation between TF and MSAF was evaluated using the Spearman coefficient (r). Variant allele frequency (VAF) and normalized VAF (nVAF) were used to determine variant clonality. Results: In total, 23 pts with initial response to pembro and subsequent PD had plasma samples available from BL, C5, and D. Samples from 22 pts had TF >1.5% by pWGS and underwent pWES. MSAF by pWES was correlated with TF by pWGS at BL (r = 0.68), C5 (0.80), and D (0.59). ctDNA TFs at BL and D (mean = 8.1% and 4.9%, respectively) were higher than C5 (2.0%). In a pooled variant-level analysis (nVAF >0.25) of samples from pts with high TF (≥2.5%) at both BL and D (n = 10), 786 (46%) of variants were present at BL and D, 391 (23%) were present at BL only, and 540 (32%) were present at D only. Variants shared between BL and D had higher clonality (mean nVAF = 0.46) than BL-only (0.38) or D-only variants (0.4). In general, shared variants had similar nVAFs at BL and D but in some pts nVAFs increased at D, suggesting evolution of tumor subclones during treatment. TP53 was mutated in 9 of 10 pts with high TF and was shared at BL and D in 6 pts (mean nVAF = 0.49), present at BL only in 1 pt (0.11) and at D only in 2 pts (0.43). KRAS G12V and Q61H hotspot mutations were observed in 2 pts at D-only and may be related to immune editing. Conclusions: Comprehensive pWES is feasible with sufficient ctDNA burden (TF ≥2.5%). ctDNA dynamics demonstrated molecular profiles characteristic of recurrence with high TF at BL and D and were consistent with radiographic determination of response/progression. Tumor variants at BL and D generally had a common clonal trunk with clones disappearing or appearing de novo. Citation Format: Minita Shah, EJ Dettman, Andrew Albright, Cai Chen, Julie Kobie, Vinay Varadan, Ayman Samkari, Carol Peña, Razvan Cristescu, Z Alexander Cao. Evaluation of on-treatment tumor molecular dynamics using plasma whole genome sequencing (pWGS) and whole exome sequencing (pWES) of circulating tumor DNA (ctDNA) in patients (pts) with metastatic non-small cell lung cancer (NSCLC): An exploratory analysis of KEYNOTE-598 [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 3664.

Abstract 3951: Timing the development of chemoresistance in relapsed pediatric cancer

Abstract Survivors of pediatric cancer face lifelong battles with severe morbidities, which includes a significant risk of recurrence. Pre-existing genetic variation within primary tumors may offer certain cell populations an evolutionary advantage, increasing the likelihood that some will be resistant to treatment. But the inherent complexity of the tumor means finding these markers of resistance has yet to be fully explored across childhood cancer. To address this, we developed a pipeline that combines clonal evolution reconstruction with mutational signature extraction at the subclonal level to elucidate changes in mutational processes. Using this pipeline to analyze 1,743 pediatric tumor genomes we found a common pattern of parallel evolution in most pediatric cancers, irrespective of their tissue of origin. Further, we detected therapy-associated mutational signatures linked to specific treatments such as temozolomide, platinum, fluorouracil, and thiopurine in post-treatment tumor subpopulations. Of 235 samples with known treatment exposure, 37.9% displayed one or more treatment-associated signatures. In the context of the phylogenetic tree, we identified specific subpopulations with therapy signatures. In multi-sample cases, resistant subpopulations had ancestors present in primary tumors at the time of diagnosis, suggesting certain subpopulations possessed the ability to withstand chemotherapy-induced pressures from an early stage and then expanded following treatment. Finally, we observed differences in the timing of mutational signatures between antimetabolite and alkylating agent treatments, reflecting the mechanisms of these drugs. Antimetabolite agents showed subclonal mutational signatures occurring in later branches, likely due to dependence on cell cycling. In contrast, alkylating agents displayed signatures in both early and late subpopulations, as they are unaffected by cell cycle state. Thus, investigating mutational signatures at the subclonal level unveils new insights into the clonal dynamics of pediatric cancers and the development of chemoresistance. This research offers valuable insights for identifying pediatric cancer patients at greater risk of recurrence, enabling the development of more effective treatment plans to enhance their quality of life and overall survival rates. Citation Format: Sasha Blay, Mehdi Layeghifard, Scott Davidson, David Chen, Astra Schwertschkow, Vijay Ramaswamy, Michael Taylor, Elli Papaemmanuil, Anita Villani, David Malkin, Ludmil Alexandrov, Mark Cowley, Adam Shlien. Timing the development of chemoresistance in relapsed pediatric cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 3951.

Abstract 4911: PAIRWISE: Deep learning-powered personalized medicine approach identifies actionable drug combinations in B cell lymphoma

Abstract Combination drug therapies offer promise for improving cancer treatment. However, identifying actionable drug combinations for specific cancer subtypes and individual patients remains challenging due to the vast number of possibilities. Existing approaches like high-throughput screening (HTS) and rule-based drug additions have limitations in their ability to predict clinically translatable synergistic combinations. Therefore, advanced machine learning focused on patient-centric modelling have the potential to optimize combination therapies. PAIRWISE is a patient-centric deep learning-powered drug combination recommendation platform. Given gene expression along with chemical structures and targets from a set of two drugs being applied, the output is the binary combination outcome. PAIRWISE introduces a new approach of representing tumor genomics using the unsupervised pretraining design on 11K TCGA tumors. The power of PAIRWISE is derived from knowledge-guided transfer learning by learning from in vitro experiments with drug-drug-cell interactions and cellular heterogeneity. PAIRWISE can accurately predict combination synergy with area under the curve (AUC) > 0.85 across 10 tissue types in a large dataset of 200K in vitro drug combinations and significantly outperform all state-of-the-art methods. Diffuse large B cell lymphoma (DLBCL) is the most common B cell non-Hodgkin lymphoma and remains incurable in around 40% of patients. Finding optimal combination therapy for each DLBCL patient is a critical unmet need. As independent validation, we applied PAIRWISE to an external HTS study of Bruton Tyrosine Kinase inhibitor (BTKi) combinations (Griner et al., 2014), confirming the reproducibility and accuracy of high-confidence predictions (AUC = 0.72). PAIRWISE was also applied to transcriptomic data of 29 DLBCL cell lines (Bal et al., 2022) and predicted BTKi-containing combinations across ABC subtypes with higher synergy scores than GCB (P = 0.02). This is consistent with expected better overall efficacy of BTK inhibitors in ABC subtype (Younes et al., 2019; Davis et al., 2010; Wilson et al., 2015). To predict opportunities in DLBCL, PAIRWISE was used to identify BTKi partners for an additional 211 anti-cancer drugs using 562 published ex vivo DLBCL samples (Schmitz et al., 2018). PAIRWISE effectively identified novel DLBCL sub-population with significantly higher response scores for predicted BTKi combinations and favorable survival. Notably, patients predicted as responsive had significantly better outcomes upon first-line chemotherapy treatments than non-responsive ones (P < 0.0001). Collectively, we demonstrate that PAIRWISE is a highly efficient and generalizable framework that enables confident predictions of pairwise drug combinations at disease subtype level, with the potential of informing treatment options and clinical trial design. Citation Format: Chengqi Xu, Jake Cohen-Setton, Zhouyang Shen, Marta Milo, Ilkay US, Ben Sidders, Jude Fitzgibbon, Heng Pan, Olivier Elemento, Krishna Bulusu. PAIRWISE: Deep learning-powered personalized medicine approach identifies actionable drug combinations in B cell lymphoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 4911.

Abstract 3401: Uncovering virus induced dysregulated processes in cancer and their impact on outcomes

Abstract The National Cancer Institute has designated several viruses such as EBV, HBV, HCV, HIV, HPV, and HTLV as agents that either cause cancer or increase the risk of developing cancer. However, extensive analyses of tumor genome and transcriptome datasets have identified numerous additional virus types, the functional implications of which remain largely unknown. Viral infection often dysregulates unique molecular processes within tumor cells and alters the composition of immune cells in the tumor microenvironment. For instance, HPV-positive tumors have been shown to exhibit a significantly elevated expression of the DNA editing enzyme APOBEC3B compared to HPV-negative tumors. Furthermore, HPV positive tumors display and increased number of C-to-T mutations induced by this protein. The genome and transcriptome of tumor tissue biopsies contain a wealth of information about these virus-associated signatures. However, advanced signal processing approaches are needed to extract these viral signatures from complex tumor datasets, and elucidate their specific roles in cancer. Here, we developed a quantitative approach to investigate the link between viral infection and dysregulation of a seven-membered A3 enzyme family (A3A/B/C/D/F/G/H). The primary role of these enzymes is to restrict viral infections by inducing C-to-U mutations in viral genomes. However, at least three members of this enzyme family (A3A/B/H) can also induce mutations in cellular DNA, and drive tumor formation and evolution. We used NMF (Nonnegative Matrix Factorization) to deconvolute the expression profiles of A3 enzymes in >10,000 bulk RNAseq datasets from diverse TCGA tumors. This analysis revealed the expression profiles of A3 enzymes within tumor and tumor microenvironment cells including various immune cells. Next, we conducted an association analysis to identify the link between the presence of viral infection and the proportion of each of these tumor and tumor microenvironment expression profiles. As expected, our analyses revealed strong associations between tumor-specific A3 expression profiles and infections by oncolytic viruses such as HPV. Notably, and somewhat unexpectedly, our analyses also revealed associations with viruses about which little is known in the field of cancer. For example, in certain tumors, we identified a robust association between bacteriophages and a tumor-specific A3 expression profile characterized by A3B upregulation. Taken together, our analyses provide evidence for the link between A3 dysregulation and infection by viruses including bacteriophages. Citation Format: Mohadeseh Soleimanpour moghadam, Azad Khosh, Diako Ebrahimi. Uncovering virus induced dysregulated processes in cancer and their impact on outcomes [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 3401.

CELSR3 is a prognostic marker in HNSCC and correlates with immune cell infiltration in the tumor microenvironment

PurposeTo look at the diagnostic value of the CELSR receptor 3 (CELSR3) gene in head and neck squamous cell carcinoma (HNSCC) and its effect on tumor immune invasion, which is important for enhancing HNSCC treatment.MethodsSeveral bioinformatics tools were employed to investigate CELSR3’s putative oncogenic pathway in HNSCC, and datasets from The Tumor Genome Atlas (TCGA), Tumor Immune Estimation Resource (TIMER), Gene Expression Profile Interaction Analysis (GEPIA) and LinkedOmics were extracted and evaluated. CELSR3 has been linked to tumor immune cell infiltration, immunological checkpoints, and immune-related genes. CELSR3’s putative roles were investigated using Gene ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), and pathway enrichment analysis. The expression level of CELSR3 in HNSCC tissues and cells was detected by RT-qPCR. The effects of CELSR3 on proliferation of HNSCC cells were detected by CCK-8 assay.ResultsCELSR3 was shown to be expressed differently in different types of cancer and normal tissues. CELSR3 gene expression was linked to pN-stage and pM-stage. Patients with high CELSR3 expression also have a well prognosis. CELSR3 expression was found to be an independent predictive factor for HNSCC in both univariate and multivariate Cox regression analyses. We discovered the functional network of CELSR3 in HNSCC using GO and KEGG analysis. CELSR3 expression levels were found to be favorably associated with immune cell infiltration levels. Furthermore, CELSR3 expression levels were significantly correlated with the expression levels of many immune molecules, such as MHC genes, immune activation genes, chemokine receptors, and chemokines. CELSR3 is highly expressed in HNSCC tissues and cells. CELSR3 overexpression significantly inhibited the proliferation of HNSCC cells. CELSR3 expression may affect the immune microenvironment and, as a result, the prognosis of HNSCC.ConclusionCELSR3 expression is elevated in HNSCC tumor tissues, and high CELSR3 expression is associated with well prognosis, which inhibited the proliferation of NHSCC cells. CELSR3 has the potential to influence tumor formation by controlling tumor-infiltrating cells in the tumor microenvironment (TME). As a result, CELSR3 may have diagnostic significance in HNSCC.

Mesoscale DNA features impact APOBEC3A and APOBEC3B deaminase activity and shape tumor mutational landscapes

Antiviral DNA cytosine deaminases APOBEC3A and APOBEC3B are major sources of mutations in cancer by catalyzing cytosine-to-uracil deamination. APOBEC3A preferentially targets single-stranded DNAs, with a noted affinity for DNA regions that adopt stem-loop secondary structures. However, the detailed substrate preferences of APOBEC3A and APOBEC3B have not been fully established, and the specific influence of the DNA sequence on APOBEC3A and APOBEC3B deaminase activity remains to be investigated. Here, we find that APOBEC3B also selectively targets DNA stem-loop structures, and they are distinct from those subjected to deamination by APOBEC3A. We develop Oligo-seq, an in vitro sequencing-based method to identify specific sequence contexts promoting APOBEC3A and APOBEC3B activity. Through this approach, we demonstrate that APOBEC3A and APOBEC3B deaminase activity is strongly regulated by specific sequences surrounding the targeted cytosine. Moreover, we identify the structural features of APOBEC3B and APOBEC3A responsible for their substrate preferences. Importantly, we determine that APOBEC3B-induced mutations in hairpin-forming sequences within tumor genomes differ from the DNA stem-loop sequences mutated by APOBEC3A. Together, our study provides evidence that APOBEC3A and APOBEC3B can generate distinct mutation landscapes in cancer genomes, driven by their unique substrate selectivity.

Tumor Genomics and the Association with Survival in Recurrent/metastatic Head and Neck Cancer Patients

Tumor Genomics and the Association with Survival in Recurrent/metastatic Head and Neck Cancer Patients

Primary Resistance to RET Inhibition in a RET Fusion-Positive Pancreatic Neuroendocrine Carcinoma.

We present a 54-year-old White male with a diagnosis of stage IV pancreatic neuroendocrine carcinoma. Next-generation sequencing of the tumor/blood identified a complex tumor genome, which included a rearranged during transfection (RET) gene fusion. The patient initially received cytotoxic chemotherapy with a significant radiographic response. After 4 cycles of chemotherapy, the patient was transitioned to a clinical trial using selpercatinib, a RET inhibitor, as maintenance therapy. Unfortunately, our patient developed progression of disease at the first treatment monitoring scan. Our patient suffered primary resistance to RET-targeted therapy. Proposed mechanisms of resistance include intrinsic resistance of the nuclear receptor co-activator 4-RET fusion to RET inhibition, the RET fusion representing a passenger alteration to another tumorigenic driver pathway and/or decreased efficacy of RET inhibition after platinum-based chemotherapy. Our patient's clinical course highlights the fact that "actionable" genomic alterations do not always equate to patient benefit.

High-throughput evaluation of genetic variants with prime editing sensor libraries.

Tumor genomes often harbor a complex spectrum of single nucleotide alterations and chromosomal rearrangements that can perturb protein function. Prime editing has been applied to install and evaluate genetic variants, but previous approaches have been limited by the variable efficiency of prime editing guide RNAs. Here we present a high-throughput prime editing sensor strategy that couples prime editing guide RNAs with synthetic versions of their cognate target sites to quantitatively assess the functional impact of endogenous genetic variants. We screen over 1,000 endogenous cancer-associated variants of TP53-the most frequently mutated gene in cancer-to identify alleles that impact p53 function in mechanistically diverse ways. We find that certain endogenous TP53 variants, particularly those in the p53 oligomerization domain, display opposite phenotypes in exogenous overexpression systems. Our results emphasize the physiological importance of gene dosage in shaping native protein stoichiometry and protein-protein interactions, and establish a framework for studying genetic variants in their endogenous sequence context at scale.

Syngeneic murine models with distinct immune microenvironments represent subsets of human intrahepatic cholangiocarcinoma

Cholangiocarcinoma (CCA) is a poorly immunogenic malignancy associated with limited survival. Syngeneic immunocompetent mouse models of CCA are an essential tool to elucidate the tumor immune microenvironment (TIME), understand mechanisms of tumor immune evasion, and test novel immunotherapeutic strategies. The scope of this study was to develop and characterize immunocompetent CCA models with distinct genetic drivers, and correlate tumor genomics, immunobiology, and therapeutic response. A multifaceted approach including scRNA-seq, CITE-seq, whole exome and bulk RNA sequencing was employed. FDA-approved PD-1/PD-L1 antibodies were tested in humanized PD-1/PD-L1 mice (HuPD-H1). A genetic mouse model of intrahepatic CCA (iCCA) driven by intrabiliary transduction of Fbxw7ΔF/Akt that mimics human iCCA was generated. From the Fbxw7ΔF/Akt tumors, a murine cell line (FAC) and syngeneic model with genetic and phenotypic characteristics of human iCCA were developed. Established SB1 (YAPS127A/Akt) and KPPC (KrasG12Dp53L/L) models were compared to the FAC model. Although the models had transcriptomic similarities, they had substantial differences as well. Mutation patterns of FAC, SB1, and KPPC cells matched different mutational signatures in Western and Japanese CCA patient cohorts. KPPC tumors had a high tumor mutation burden. FAC tumors had a T cell-infiltrated TIME, while SB1 tumors had apreponderance of suppressive myeloid cells. FAC, SB1, and KPPC tumors matched different immune signatures in human iCCA cohorts. Moreover, FAC, SB1, and KPPC tumor-bearing HuPD-H1 mice displayed differential responses to nivolumab or durvalumab. Syngeneic iCCA models display a correlation between tumor genotype and TIME phenotype, with differential responses to FDA-approved immunotherapies. This study underscores the importance of leveraging multiple preclinical models to understand responses to immunotherapy in different genetic subsets of human CCA. Understanding the relationship between tumor genotype and the phenotype of the immune microenvironment is an unmet need in cholangiocarcinoma (CCA). Herein, we use syngeneic murine models of intrahepatic CCA with different genetic drivers to demonstrate a correlation between tumor genotype and immune microenvironment phenotype in murine models, which is associated with differential responses to FDA-approved immunotherapies. This information will help guide other preclinical studies. Additionally, it emphasizes that immune checkpoint inhibition in patients with CCA is not a "one-size-fits-all" approach. Our observations suggest that, as for targeted therapies, patients should be stratified and selected for treatment according to their tumor genetics.

Identifying Oncology Clinical Trial Candidates Using Artificial Intelligence Predictions of Treatment Change: A Pilot Implementation Study.

Precision oncology clinical trials often struggle to accrue, partly because it is difficult to find potentially eligible patients at moments when they need new treatment. We piloted deployment of artificial intelligence tools to identify such patients at a large academic cancer center. Neural networks that process radiology reports to identify patients likely to start new systemic therapy were applied prospectively for patients with solid tumors that had undergone next-generation sequencing at our center. Model output was linked to the MatchMiner tool, which matches patients to trials using tumor genomics. Reports listing genomically matched patients, sorted by probability of treatment change, were provided weekly to an oncology nurse navigator (ONN) coordinating recruitment to nine early-phase trials. The ONN contacted treating oncologists when patients likely to change treatment appeared potentially trial-eligible. Within weekly reports to the ONN, 60,199 patient-trial matches were generated for 2,150 patients on the basis of genomics alone. Of these, 3,168 patient-trial matches (5%) corresponding to 525 patients were flagged for ONN review by our model, representing a 95% reduction in review compared with manual review of all patient-trial matches weekly. After ONN review for potential eligibility, treating oncologists for 74 patients were contacted. Common reasons for not contacting treating oncologists included cases where patients had already decided to continue current treatment (21%); the trial had no slots (14%); or the patient was ineligible on ONN review (12%). Of 74 patients whose oncologists were contacted, 10 (14%) had a consult regarding a trial and five (7%) enrolled. This approach facilitated identification of potential patients for clinical trials in real time, but further work to improve accrual must address the many other barriers to trial enrollment in precision oncology research.

Advancing Precision Oncology in Hereditary Paraganglioma-Pheochromocytoma Syndromes: Integrated Interpretation and Data Sharing of the Germline and Tumor Genomes.

Standard methods of variant assessment in hereditary cancer susceptibility genes are limited by the lack of availability of key supporting evidence. In cancer, information derived from tumors can serve as a useful source in delineating the tumor behavior and the role of germline variants in tumor progression. We have previously demonstrated the value of integrating tumor and germline findings to comprehensively assess germline variants in hereditary cancer syndromes. Building on this work, herein, we present the development and application of the INT2GRATE|HPPGL platform. INT2GRATE (INTegrated INTerpretation of GeRmline And Tumor gEnomes) is a multi-institution oncology consortium that aims to advance the integrated application of constitutional and tumor data and share the integrated variant information in publicly accessible repositories. The INT2GRATE|HPPGL platform enables automated parsing and integrated assessment of germline, tumor, and genetic findings in hereditary paraganglioma-pheochromocytoma syndromes (HPPGLs). Using INT2GRATE|HPPGL, we analyzed 8600 variants in succinate dehydrogenase (SDHx) genes and their associated clinical evidence. The integrated evidence includes germline variants in SDHx genes; clinical genetics evidence: personal and family history of HPPGL-related tumors; tumor-derived evidence: somatic inactivation of SDHx alleles, KIT and PDGFRA status in gastrointestinal stromal tumors (GISTs), multifocal or extra-adrenal tumors, and metastasis status; and immunohistochemistry staining status for SDHA and SDHB genes. After processing, 8600 variants were submitted programmatically from the INT2GRATE|HPPGL platform to ClinVar via a custom-made INT2GRATE|HPPGL variant submission schema and an application programming interface (API). This novel integrated variant assessment and data sharing in hereditary cancers aims to improve the clinical assessment of genomic variants and advance precision oncology.

Extrachromosomal DNA in cancer.

Extrachromosomal DNA (ecDNA) has recently been recognized as a major contributor to cancer pathogenesisthat is identified in most cancer types and isassociated with poor outcomes. When it wasdiscovered over 60 years ago, ecDNA was considered to be rare, and its impact on tumour biology was not well understood. The application of modern imaging and computational techniques has yielded powerful new insights into the importance of ecDNA in cancer. The non-chromosomal inheritance of ecDNA during cell division results in high oncogene copy number, intra-tumoural genetic heterogeneity and rapid tumour evolution thatcontributes to treatment resistance and shorter patient survival. In addition, the circular architecture of ecDNA results in altered patterns of gene regulation that drive elevated oncogene expression, potentially enabling the remodelling of tumour genomes. The generation of clusters of ecDNAs, termed ecDNA hubs, results in interactions between enhancers and promoters in trans, yielding a new paradigm in oncogenic transcription. In this Review, we highlight the rapid advancements in ecDNA research, providing new insights into ecDNA biogenesis, maintenance andtranscription and itsrole in promoting tumour heterogeneity. To conclude, we delve into a set of unanswered questions whose answers will pave the way for the development of ecDNA targeted therapeutic approaches.

Identification of breast cancer subtypes based on graph convolutional network

Identification of molecular subtypes of malignant tumors plays a vital role in individualized diagnosis, personalized treatment, and prognosis prediction of cancer patients. The continuous improvement of comprehensive tumor genomics database and the ongoing breakthroughs in deep learning technology have driven further advancements in computer-aided tumor classification. Although the existing classification methods based on gene expression omnibus database take the complexity of cancer molecular classification into account, they ignore the internal correlation and synergism of genes. To solve this problem, we propose a multi-layer graph convolutional network model for breast cancer subtype classification combined with hierarchical attention network. This model constructs the graph embedding datasets of patients' genes, and develops a new end-to-end multi-classification model, which can effectively recognize molecular subtypes of breast cancer. A large number of test data prove the good performance of this new model in the classification of breast cancer subtypes. Compared to the original graph convolutional neural networks and two mainstream graph neural network classification algorithms, the new model has remarkable advantages. The accuracy, weight-F1-score, weight-recall, and weight-precision of our model in seven-category classification has reached 0.851 7, 0.823 5, 0.851 7 and 0.793 6 respectively. In the four-category classification, the results are 0.928 5, 0.894 9, 0.928 5 and 0.865 0 respectively. In addition, compared with the latest breast cancer subtype classification algorithms, the method proposed in this paper also achieved the highest classification accuracy. In summary, the model proposed in this paper may serve as an auxiliary diagnostic technology, providing a reliable option for precise classification of breast cancer subtypes in the future and laying the theoretical foundation for computer-aided tumor classification.

ARID1A loss is associated with increased NRF2 signaling in human head and neck squamous cell carcinomas.

Prior to the next generation sequencing and characterization of the tumor genome landscape, mutations in the SWI/SNF chromatin remodeling complex and the KEAP1-NRF2 signaling pathway were underappreciated. While these two classes of mutations appeared to independently contribute to tumor development, recent reports have demonstrated a mechanistic link between these two regulatory mechanisms in specific cancer types and cell models. In this work, we expand upon these data by exploring the relationship between mutations in BAF and PBAF subunits of the SWI/SNF complex and activation of NRF2 signal transduction across many cancer types. ARID1A/B mutations were strongly associated with NRF2 transcriptional activity in head and neck squamous carcinomas (HNSC). Many additional tumor types showed significant association between NRF2 signaling and mutation of specific components of the SWI/SNF complex. Different effects of BAF and PBAF mutations on the polarity of NRF2 signaling were observed. Overall, our results support a context-dependent functional link between SWI/SNF and NRF2 mutations across human cancers and implicate ARID1A inactivation in HPV-negative HNSC in promoting tumor progression and survival through activation of the KEAP1-NRF2 signaling pathway. The tumor-specific effects of these mutations open a new area of study for how mutations in the KEAP1-NRF2 pathway and the SWI/SNF complex contribute to cancer.