Tumor Evolution Research Articles

Abstract 1617: Tumor evolution during therapy in ovarian cancer

Abstract The evolution of tumors from diagnosis to relapse in ovarian high-grade serous carcinoma (HGSC) patients is poorly understood. Consequently, it is unknown whether diagnostic samples are representative enough to guide treatment directions at relapse, or whether acquired changes during therapy provide targets in patients with relapses. To answer these questions and characterize tumor evolution during treatment, we used data from >700 tissue and plasma samples from 200 HGSC patients belonging to the observational, longitudinal DECIDER trial (NCT04846933). Mutations from whole-genome sequencing data from tissue samples (2-12 per patient) were used to construct phylogenetic trees. The trees were corrected by the copy-number alterations and curated to integrate circulating tumor DNA (ctDNA) exome- and targeted-sequencing data from plasma samples. Evolutionary states presenting within and between sample heterogeneity were evaluated through clonal complexity and divergence. Additionally, genetic differences between persistent and non-persistent subclones were studied to identify features affecting treatment resistance. Our results indicate that the majority of subclones survive chemotherapy treatment and major subclonal replacement takes place in only a few patients. Within-sample heterogeneity was higher in plasma samples than in ascites, highlighting the importance of ctDNA samples to estimate persistent subclones. Additionally, our recently reported evolutionary states (Lahtinen et al. Cancer Cell 2023) were often altered by the treatment induced selection pressure. At late relapses, subclone compositions typical of maintaining evolutionary state tumors were less common, and a higher number of acquired subclones were detected, suggesting further evolution of the persisting clones. In summary, chemotherapy does not suppress heterogeneity in HGSC tumors. Interestingly, patients with the worst prognosis have a stable composition of multiple clones in their tumors, and tumors with low heterogeneity responded better to primary treatment. Relapse samples show lower within-sample heterogeneity and further evolution. Furthermore, we detected several evolutionary trajectories from stable, persistent cancer subclone populations to highly dynamic subclone structures. This renders diagnostic samples suitable to guide relapse treatment decisions in a subset of patients with HGSC but not for all. Our characterization of the dynamics of tumor clonal heterogeneity during treatment allows to define better therapeutic interventions for HGSC patients with relapsed disease. Citation Format: Sampsa Kalervo Hautaniemi, Jaana Oikkonen, Alexandra Lahtinen, Mai T.N. Nguyen, Giovanni Marchi, Kari Lavikka, Anna Rajavuori, Sakari Hietanen, Anni Virtanen, Johanna Hynninen. Tumor evolution during therapy in ovarian 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 1617.

Abstract 419: Evolutionary trajectories of complex genome rearrangements in osteosarcoma

Abstract Whole-genome sequencing (WGS) studies have revealed that osteosarcoma genomes are riddled by remarkably intricate forms of structural variants (SV), collectively known as complex genomic rearrangements (CGR). Yet, most CGR detected in osteosarcoma remain unexplained, hinting at the possibility of undiscovered mechanisms that might lead to novel therapeutic strategies. To elucidate the mechanisms underpinning cancer genome complexity in osteosarcomas and the downstream consequences of CGR during tumor evolution, we uniformly processed exiting osteosarcoma WGS data sets and performed high-depth (>120x) multi-regional short- and long-read WGS for hundreds of osteosarcomas. We found that whole genome doubling and diverse types of CGR are frequent clonal events in most osteosarcomas. Through the integration of multi-regional WGS data, we discovered that in tumors with whole-genome doubling, cancer genome evolution and oncogene amplification are the consequence of multi-generational CGR triggered by whole-genome doubling and involving multiple mechanisms, especially chromothripsis and breakage-fusion-bridge cycles. By contrast, in tumors without whole-genome doubling, biallelic inactivation of TP53 and oncogene amplification frequently occur as part of a cascade of genomic instability initiated by a single double-strand break upstream of the TP53 locus. This process results in clonal diversification and rapid tumor growth suggestive of punctuated evolution. In addition, we found that subclonal CGRs are also frequent events. Indeed, clonal expansions triggered by the acquisition of subclonal CGR often colonize distant tumor regions, suggesting that CGR act as subclonal driver events. Finally, we show that some recurrent genomic alterations are prognostic of poor survival, which we validate in independent cohorts. In sum, using WGS data, we have elucidated the mechanisms underpinning the complexity and evolution of osteosarcoma genomes. Our results indicate that diverse types of CGR, such as chromothripsis, are triggered by whole genome doubling and occur throughout tumor evolution, which primes the cancer genome for ongoing genomic instability, rapid karyotype evolution and clonal diversification. These results have implications for our understanding of the molecular basis of intra-tumor heterogeneity and drug resistance in osteosarcomas and other cancers driven by genomic instability. Citation Format: Isidro Cortes Ciriano, Jose Espejo Valle-Inclán, Solange de Noon, Katherine Trevers, Hillary Elrick, Adrienne M. Flanagan. Evolutionary trajectories of complex genome rearrangements in osteosarcoma [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 419.

Abstract 1201: Pervasive HLA disruption fuels cancer evolution

Abstract Disruption of the human leukocyte antigen (HLA) molecules has important implications for immune evasion and tumor evolution. However, although genomic loss of HLA is frequent in non-small cell lung cancer (NSCLC) and in some types of breast cancer, the extent and importance of transcriptomic disruption to HLA presentation, including transcript repression and alternative splicing (AS), remains unclear. To address this, we developed MHC Hammer, a tool to call HLA allele-specific mutations, loss of heterozygosity (LOH), transcriptional repression, and alternative splicing. We first assessed HLA allelic expression and the prevalence of AS in normal tissue. Applying MHC Hammer to 510 normal lung samples and 407 normal breast samples from the Genotype-Tissue Expression cohort, we identified extensive heterogeneity in HLA allelic expression depending on both the allele and tissue type. In addition, HLA allelic AS, occurring in at least 20% of the allele transcripts, was observed in 19% of normal lung and 13% of normal breast samples. Next, we applied MHC Hammer to 413 TRACERx NSCLCs and 743 TCGA breast tumours. While mutations were uncommon, HLA LOH was a frequent event, occurring in 32% of lung adenocarcinomas (LUAD) and 57% of lung squamous cell carcinomas (LUSC). We observed a similar rate of HLA LOH in estrogen receptor positive (ER+) and negative (ER-) tumors (ER+: 14%, ER-: 15%), while in triple negative breast tumours we observed a higher rate of LOH (31%). Due to the heterogeneity in HLA expression observed in the normal samples, to measure tumor specific HLA repression and alternative splicing, we restricted our analysis to tumours with WES, tumor and tumor-adjacent normal RNAseq data. This resulted in a cohort of 88 NSCLCs from the TRACERx study as well as 44 ER+ breast tumours from the TCGA cohort. 61% of LUAD, 75% of LUSC and 35% of ER+ tumours harbored class I HLA transcriptional repression that could not be explained by underlying genomic loss. 27% of LUAD, 7% of LUSC and 55% of ER+ breast tumor patients exhibited no HLA disruption. Tumor specific AS affecting HLA exons 2, 3 or 4, potentially disrupting HLA antigen presentation, was identified in 8% of LUADs, 11% of LUSC and 7% of ER+ breast tumours. Tumor specific AS of exon 5, yielding soluble HLA molecules without a transmembrane domain was identified in 20% of LUAD, 14% of LUSCs and 18% of ER+ breast tumours. LUAD and LUSC tumor regions without HLA LOH or repression were enriched for somatic HLA AS events and alleles with somatic HLA AS had a higher neoantigen burden than those without. We found no relationship between survival and HLA expression or AS in the normal tissue. However, consistent with the importance of HLA dysfunction in tumor evolution, LUADs with low HLA expression in at least one tumor region were associated with shorter disease-free survival, HLA LOH was more common in LUAD’s that metastasized and HLA repression was enriched in LUSC primary regions seeding metastases. Citation Format: Clare Puttick, Thomas P. Jones, Michelle Leung, Oriol Pich, Felipe Galvez Cancino, Jiali Liu, Manuel Varas-Godoy, Andrew Rowan, Carlos Martinez-Ruiz, Robert Bentham, Krijn K. Dijkstra, Rachel Rosenthal, Nnennaya Kanu, Kevin Litchfield, Roberto Salgado, David A. Moore, Peter Van Loo, Mariam Jamal-Hanjani, Sergio A. Quezada, Nicholas McGranahan, Charles Swanton. Pervasive HLA disruption fuels cancer 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 1201.

Abstract 6225: Tumor evolution through selection by ECM stiffness

Abstract In this study, we used experimental evolution to investigate the extent to which the mechanical stiffness of the tumor extracellular matrix (ECM) exerts selection pressure on a genetically diverse population of breast cancer cells. Polyacrylamide hydrogels were fabricated and conjugated with collagen type I to create the ECM with different stiffness. MDA-MB-231 cells were continuously cultured on the hydrogels of the same corresponding stiffness. Growth rate was determined at certain time points by harvesting and counting the cells. Abundance of clonal MDA-MB-231 cells labeled with heritable DNA barcodes underwent the same selection was identified with sequencing of DNA barcode tags. Cellular phenotypes were characterized through immunofluorescence staining and confocal microscopy. 3D migration of the cancer cells was examined using Boyden chamber assay. Cellular contractility was measured through traction force microscopy. Whole genome transcriptome sequencing (RNA-seq), assay for transposase-accessible chromatin with sequencing, and reduced representation bisulfite sequencing were performed for transcriptomic and epigenetic analysis. Sustained culture of MDA-MB-231 breast cancer cells overall three months on soft ECM resulted in an increased proliferation rate and beta1 integrin-dependent spreading area of the cells. Such changes did not occur in clonal cells with low genetic variation and did not occur in ancestral cells on stiff substrates. Barcode analysis revealed two distinct surviving populations post-genetic selection, with one demonstrating robust proliferation on both soft and stiff ECM, and the other exhibiting preferential adaptation to the soft matrix. The soft-selected populations exhibited nuclear localization of the transcriptional regulator yes-associated protein and less wrinkled nuclei compared with ancestral populations when both populations were grown on soft ECM. Aggressive migration was observed in the soft-selected populations, indicating a more malignant selected phenotype. The soft-selected cells showed increased RhoA activity and contractile tension on a soft substrate compared to the ancestral cells. RNA-seq analysis revealed systematic differences between ancestral, stiff-selected, and soft-selected populations corresponding to enriched biological processes related to cell adhesion and ECM organization. Significant differences in DNA methylation and chromatin accessibility were observed in soft selected populations which were correlated with changes in RNA expression. ECM stiffness exerts selection pressure on genetically variable cancer cell populations, selecting specific genetic variants in a cell population optimally adapted for that stiffness. Selected variants on soft ECM tend to be highly migratory, proliferative and generate high traction even on soft ECM. Our results highlight the importance of ECM stiffness-driven evolution in cancer development. Citation Format: Ting-Ching Wang, Suchitaa Sawhney, Daylin Morgan, Richa Rashmi, Marcos R. Estecio, Amy Brock, Irtisha Singh, Richard L. Bennett, Charles F. Baer, Jonathan D. Licht, Tanmay P. Lele. Tumor evolution through selection by ECM stiffness [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 6225.

Abstract 3289: APOBEC3A drives tumor evolution through activation of ERVs in non small cell lung cancer

Abstract Acquired drug resistance to even the most effective anti-cancer targeted therapies remains an unsolved clinical problem. Although many drivers of acquired drug resistance have been identified, the underlying molecular mechanisms shaping tumor evolution during treatment are incompletely understood. We recently demonstrated that lung cancer targeted therapies commonly used in the clinic induce the expression of cytidine deaminase APOBEC3A (A3A), leading to sustained mutagenesis in drug-tolerant cancer cells persisting during therapy. Preventing therapy-induced A3A mutagenesis by gene deletion decreased the chromosomal aberration events and delayed the emergence of drug resistance. Thus, inhibition of A3A may represent a potential therapeutic strategy to prevent acquired resistance. Understanding the regulatory mechanisms of A3A may provide alternative approaches to inhibit A3A mutagenesis. Using a panel drug screen, we found that DNA methyltransferase (DNMT) inhibitor remarkably induces A3A in non-small cell lung cancer cells. RNA-seq profiling revealed that targeted therapy as well as DNMT inhibitor treatment activate expression of distinct classes non-coding repetitive RNA elements, including endogenous retrovirus (ERVs), in drug-tolerant persister cells. Activation of intracellular viral sensing pathways by introduction of exogenous nucleic acids mimicking viral elements induced A3A expression. While the transcription factor NFkB was acutely activated by targeted therapy, TBK1-IRF3 signaling was activated at later time points. These findings suggest that reactivation of ERVs may underlie A3A mutagenic activity and tumor evolution during targeted therapy. Citation Format: Hideko Isozaki, Ramin Sakhtemani, Naveed Nikpour, Susanna Monroe, Michael Lawrence, Aaron Hata. APOBEC3A drives tumor evolution through activation of ERVs in non small cell lung 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 3289.

Abstract 869: Single-cell DNA replication dynamics in genomically unstable cancers

Abstract Background: DNA replication and cell cycle regulation are frequently disrupted as part of a cancer’s progression toward uncontrolled proliferation, generating somatic copy number alterations (CNAs) and producing intratumoral heterogeneity that drives subsequent evolution. Structural variation and CNAs have been shown to impact epigenetic and chromatin states, but our ability to assess their impact on DNA replication timing (RT) and cell proliferation rates remains limited. Single-cell whole genome sequencing (scWGS) is a powerful method for studying clonal heterogeneity and CNAs, and has the potential to provide greater insight into DNA replication dynamics in aneuploid populations. However, computational identification of S-phase cells and distinguishing inherited somatic CNAs from transient DNA replication changes remain challenging. Methods: We present a new method, PERT, which uses a Bayesian framework to model read depth as a combination of somatic copy number, replication, and sequencing bias, enabling estimation of DNA replication profiles and cell cycle phase from scWGS data. Unlike previous approaches, PERT provides unbiased estimates of RT and cell cycle phase which allows for analysis of previously uncharacterized cell types using any scWGS platform. These unique properties enable PERT to perform novel analysis such as estimating clone-specific proliferation rates and studying the interplay between RT and somatic CNAs during tumor evolution. We applied PERT to a cohort of >50,000 scWGS cells obtained from a collection of genomically unstable breast and ovarian cell lines, xenografts and primary cancer tissues. Results: Clone RT profiles correlated with future copy number changes in serially passaged cell lines. Cell type was the strongest determinant of RT heterogeneity, while whole genome doubling (WGD) and mutational signature had weaker RT associations but were associated with accumulation of late S-phase cells. Recurrent CNAs affecting chromosome X had striking impact on RT, with loss of the inactive X allele shifting replication earlier, and loss of inactive Xq resulting in reactivation of Xp. Analysis of time series xenografts illustrated that cell cycle distributions approximate clone proliferation. This relationship enabled us to observe that highly proliferative clones were the most chemosensitive in cisplatin-treated xenografts and, separately, present novel evidence that WGD leads to slower proliferation. Conclusions: Our analysis implicates cell type as the strongest determinant of RT with chrX being the locus of highest RT variation due to X-inactivation. Separately, quantification of S-phase cells enables interrogation of the on- and off-treatment fitness of genetically distinct subclones. This work leads to better understanding of how DNA replication dynamics drive and are further modulated by genomic instability. Citation Format: Adam C. Weiner, Marc Williams, Hongyu Shi, Ignacio Vazquez-Garcia, Sohrab Salehi, Nicole Rusk, Sohrab P. Shah, Andrew McPherson. Single-cell DNA replication dynamics in genomically unstable cancers [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 869.

Abstract 1205: Inferring allele-specific copy number aberrations and tumor phylogeography from spatially resolved transcriptomics

Abstract Tumors evolve through the acquisition of somatic mutations over time and across space. While the temporal aspects of tumor evolution can be inferred by applying phylogenetic techniques to bulk or single-cell DNA sequencing data, the spatial aspects of tumor evolution remain understudied. Spatially resolved transcriptomics (SRT) measures gene expression at thousands of spatial locations in a tumor, but does not directly measure genetic aberrations. Thus, SRT data has not yet been fully utilized for tumor evolution studies. We introduce CalicoST, an algorithm that simultaneously infers allele-specific copy number aberrations (CNAs) and derives a phylogeographic model of tumor evolution from SRT data. By modeling CNA-induced perturbations in both gene expression and allele frequencies, CalicoST identifies important types of CNAs - including copy-neutral loss of heterozygosity (CNLOH) and mirrored subclonal CNAs - that are invisible to total copy number analysis. On SRT data from nine patients from the Human Tumor Atlas Network (HTAN), CalicoST achieves an average accuracy of 86%, approximately 21% higher than existing methods. On two of these patients with multiple adjacent slices, CalicoST reconstructs a tumor phylogeography that describes the spread of cancerous clones in three-dimensional space. CalicoST analysis of multiple SRT slices from a cancerous prostate organ reveals five spatially coherent clones, with mirrored subclonal CNAs distinguishing clones on the two sides of the prostate, and a bifurcating phylogeography in both genetic and physical space that suggests a polyclonal cancer. Citation Format: Cong Ma, Metin Balaban, Jingxian Liu, Siqi Chen, Li Ding, Benjamin J. Raphael. Inferring allele-specific copy number aberrations and tumor phylogeography from spatially resolved transcriptomics [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 1205.

Abstract 5219: Loss of chromosome Y is unrelated to the composition of the tumor microenvironment and patient prognosis in muscle-invasive urothelial bladder cancers

Abstract Background: Loss of chromosome Y (LOY) occurs commonly in cancers of male patients. It was recently proposed to be associated with cancer aggressiveness and altered T-cell function in bladder cancer and it was found to be linked to poor prognosis in a TCGA bladder cancer patient cohort of 300 patients based on a combined analysis of RNAs from Y-chromosome genes. In this study, we aimed to clarify the clinical relevance of LOY in a large cohort of urothelial bladder cancer patients by using fluorescence in-situ hybridization (FISH). Design: X and Y chromosomes were analyzed by FISH on a tissue microarray (TMA) containing 1,819 urothelial carcinomas of the urinary bladder from male patients including 487 patients who had undergone cystectomy for muscle-invasive disease and for which follow-up data were available. Data of tumor microenvironment parameters were obtained in a previous study. Results: LOY was found in 26.0% of 1,479 analyzable cancers. In the subset of non-invasive cancers, the frequency of LOY increased from pTaG2 (14.5%) to pTG3 (28.6%) but these differences were not significant (p=0.1015). In muscle-invasive cancers, the frequency of LOY increased slightly from pT2 (25.4%) to pT4 cancers (32.8%) but this association was again not significant (p=0.1868). Within pT2-4 cancers, LOY was associated with venous invasion (p=0.0010) but unrelated to pT, pN, and L-status, as well as to overall, recurrence-free, and cancer-specific survival. Muscle-invasive urothelial carcinomas with and without LOY did not show significant differences in the number of CD8 positive lymphocytes, the fraction of CD8 positive intraepithelial lymphocytes, the number of macrophages and dendritic cells, as well as the fraction of t helper and t regulatory cells. Conclusion: FISH is the gold standard for determining LOY. That LOY was neither associated with histopathological parameters of cancer aggressiveness and patient prognosis nor with any parameters describing the tumor microenvironment strongly argues against a driving role of LOY for bladder cancer progression and cancer associated immune reaction. The slight increase of LOY from pTaG2 to pT4 cancers is consistent with a continuous (but clinically irrelevant) loss of the dispensable Y chromosomes during tumor evolution over time. Citation Format: Martina Kluth, Henning Plage, Sebastian Hofbauer, Kira Furlano, Sarah Weinberg, Antonia Franz, Annika Fendler, Florian Roßner, Simon Schallenberg, Sefer Elezkurtaj, Niclas C. Blessin, Maximilian Lennartz, Andreas H. Marx, Henrik Samtleben, Margit Fisch, Michael Rink, Marcin Slojewski, Krystian Kaczmarek, Thorsten Ecke, Stefan Koch, Nico Adamini, Ronald Simon, Guido Sauter, Joachim Weischenfeldt, Tobias Klatte, Sarah Minner, David Horst, Thorsten Schlomm, Dr. med Henrik Zecha Zecha. Loss of chromosome Y is unrelated to the composition of the tumor microenvironment and patient prognosis in muscle-invasive urothelial bladder cancers [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 5219.

Abstract 2541: The landscape of circular extrachromosomal amplification in pediatric cancers

Abstract Circular extrachromosomal DNA (ecDNA, also known as double minutes, dm) has emerged as a driver of oncogenic copy number amplification, intratumoral heterogeneity, tumor evolution and therapy resistance. We recently screened the whole genomes of 481 medulloblatomas for ecDNA and concluded that medulloblastoma tumors harboring ecDNA are more than twice as likely to progress within 5 years of diagnosis. To evaluate the frequency and clincical impact of ecDNA across the spectrum of childhood cancers, we performed a retrospective analysis of 1671 tumors from 1482 patients across 47 pediatric tumor types from two large cloud pediatric cancer genomics resources, the Children’s Brain Tumor Network and St. Jude Cloud. ecDNA was detected in 15 pediatric tumor types, including ETMR (100%, n=4), osteosarcoma (47%, n=57), rhabdomyosarcoma (40%, n=35), neuroblastoma (30%, n=106), high-grade glioma (20%, n=157), and retinoblastoma (19%, n=32). Patients with tumors containing ecDNA had significantly poorer 5-year overall survival. In sequential samples from the same tumor at diagnosis and relapse, we observe evidence of ecDNA evolution including sequence instability and generation of new high-copy ecDNA amplification not detected at initial diagnosis. These results comprise the most detailed survey to date of the landscape of extrachromosomal amplifications across childhood cancers. Citation Format: Sunita Sridhar, Owen S. Chapman, Shanqing Wang, Aditi Dutta, Jill P. Mesirov, Lukas Chavez. The landscape of circular extrachromosomal amplification in pediatric cancers [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 2541.

Abstract 6243: Rapid autopsy reveals the interplay of tumor heterogeneity at genomic, transcriptomic, and spatial dimensions

Abstract Incomplete understanding of tumor heterogeneity across primary and metastatic sites often lead to ineffective cancer treatments. To comprehensively interrogate tumor heterogeneity at genomic, transcriptomic, and spatial levels, we established a novel rapid autopsy protocol to obtain multiregional tri-omic data consisted of whole exome sequencing, mRNA-sequencing, and multiplexed immunohistochemistry on primary and metastatic tumors and associated microenvironments (TMEs). We applied this protocol to six patients with various primary tumor types: non-small cell lung cancer (NSCLC), small cell lung cancer (SCLC), urothelial cancer, and one mixed case with both NSCLC and SCLC. A total of 50 tumor samples with paired tri-omic data were obtained across 16 different primary and metastatic organs to interrogate heterogeneities underlying tumor evolution and treatment resistance mechanisms. We constructed mutation-based phylogeny trees to reveal structural differences in tumor evolution trajectories across patients and infer heterogeneous tumor antigen landscape. We uncovered highly entropic phylogenetic structures having unique variants gained at each metastasis as well as low entropic structures with high sharing across sites. The mixed case had a clonal sweep structure with two metastases developed prior to the rest of the metastasis. In this case, older clones were shown to have strong association with APOBEC mutation signatures, enhanced subclonal mutation burden and loss of human leukocyte antigen heterozygosity as mechanisms of immune evasion. Integrating transcriptome and spatial profiles, we revealed that tumor intrinsic expressions and extrinsic immune infiltration levels covary with tissue types rather than with mutation-based phylogenetic clusters. Across patients, samples that had high similarities in mutational landscape varied in their pro-tumor expression programs and the cellular composition of TMEs. In particular, lung samples across patients had the highest effector T cell population and major histocompatibility complex (MHC) class I, II presentations compared to other tissue types. Contrasting small cell carcinoma (SCC) and adenocarcinoma (Adeno) samples from lung tissues further revealed that SCC had more immunosuppressive TMEs with downregulation of MHC class II. Liver metastasis samples were further characterized to have the most aggressive phenotypes enriched in genomic instability measured by aneuploidy compared to other metastatic sites. In sum, we present a novel technology and computational framework that enables a three-dimensional profiling of tumor heterogeneity. This work suggests that single-site biopsies are insufficient to capture tumor heterogeneity and calls for multi-modal integrations to design treatments that can overcome tumor-specific resistance mechanisms. Citation Format: Ning Yao, Rami Vanguri, Juber Patel, Jayon Lihm, Andrew Fisher, Irene Jarchum, Shigeaki Umeda, Hao Li, Scott Ely, Travis Hollmann, Christine Iacobuzio-Donahue, Benjamin Greenbaum. Rapid autopsy reveals the interplay of tumor heterogeneity at genomic, transcriptomic, and spatial dimensions [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 6243.

Abstract 5324: Key antiviral response genes mediate the impact of chromosomal instability on breast cancer stemness

Abstract Chromosomal instability (CIN) is a recognized hallmark of cancer, pivotal in tumor evolution and metastasis. Our research unveils CIN as a stress during tumor initiation, uniquely tolerated by activated breast cancer stem cells (aCSCs). Elevated CIN, specific to tumors from aCSCs, contrasts with baseline levels akin to non-stem cell types. This transient increase in CIN, observed in low burden aCSC tumors, diminishes in established disease, suggesting a dynamic induction during initiation. Consist wit this we have observed increase in CIN tolerance in stem-like vs basal-like cells suggesting heterogeneity in response. CIN triggers a specific response, activating proinflammatory cancer autonomous immune signaling. We identified several CIN-induced proinflammatory stem cell-specific genes from our overlap studies. Knockdown assays targeting RSAD2 and PLAU showed a significant decrease in tumorsphere properties, emphasizing their functional relevance for cancer progression. AXL did not show any effect on tumorsphere but was important to tolerate CIN induced due to reversine. Mechanistically, AXL expression was induced by c-Jun stress signaling activation in aCSCs. These antiviral CIN-induced proinflammatory markers, emerge as a vital player in breast cancer progression. This study establishes a foundation for investigating antiviral genes as crucial contributors to cancer. Citation Format: Shahnawaz A. Baba, Samson Mugisha, Shreyas labhsetwar, Richard klemke, Jay Desgrosellier. Key antiviral response genes mediate the impact of chromosomal instability on breast cancer stemness [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 5324.

Abstract 2268: CREDAC: Copy number-based reference-free expression deconvolution analysis of cancers

Abstract Introduction: Analyses of bulk RNA sequencing data are central to most large-scale tumor sequencing studies. Bulk expression data represent population averages and its interpretation is confounded by both normal cell contamination and somatic copy number alterations. Several computational methods to deconvolve have been developed recently. However, these methods often rely on a set of cell-type-specific reference signatures and ignore the effect of copy number changes. Methods: To address these issues, we have developed a method that formalizes the relationship between allele-specific copy number, expression and sample purity to deconvolve the expression profiles of tumor and normal cells from bulk RNA-seq in an unbiased manner. Our method was applied to sequencing data produced by the TRACERx consortium, a longitudinal study with multi-region whole-exome and RNA-seq of non-small-cell lung cancers. A total of 414 primary tumor regions and 140 adjacent normal tissue samples from 140 patients with matched DNA and RNA-seq data were processed. Results: We were able to directly deconvolve a median of ~2,000 genes per sample and indirectly infer tumor and normal expression profiles of ~10,000 genes using an scaled and weighted mean approach. The accuracy of the deconvolution was validated using (1) in-silico mixtures of patient-derived tumor and normal cells, (2) pseudo-bulk scRNAseq and (3) in regions with loss of heterogeneity (LOH) directly on the bulk sequencing data, where the total fraction of expression attributed to tumor cells can be computed directly using somatic mutations. CREDAC allows for single sample bulk-level tumor-normal differential expression analysis and revealed a strong and constitutive genome-wide overexpression in cancer compared to admixed normal cells, with a more pronounced overexpression in lung squamous cell carcinoma than lung adenocarcinoma (p<0.001). CREDAC adjusts for variations in copy number, thus facilitating the investigation of cancer-specific dosage compensatory responses triggered by aneuploidy. We revealed that a majority of genes (~65%) exhibit proportionality to copy number alterations, while a substantial proportion (~35%) shows anti-scaling or dosage compensation strategies. We show an enrichment of genes involved in cell cycle and expression processes within the dosage compensated group and revealed methylation as a key mechanism driving this compensatory process. Conclusion: Overall, our results suggest that CREDAC is able to accurately disentangle the expression of tumor and normal cells from bulk RNA-seq without any previous knowledge. It has potential applications in many studies that include matched RNA-seq and copy number data and can provide new insights functional characterization, the taxonomy of cancer, and tumor evolution. Citation Format: Carla Castignani, Jonas Demeulemeester, Oriol Pich, Tom Lesluyes, Robert E. Hynds, David R. Pearce, Elizabeth Larose Cadieux, Stefan C. Dentro, TRACERx Consortium, Nnenna Kanu, Charles Swanton, Peter Van Loo. CREDAC: Copy number-based reference-free expression deconvolution analysis of cancers [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 2268.

Abstract 1615: Balanced molecular feature selection unveils universally tumor-lineage informative methylation sites in colorectal cancer

Abstract In the era of precision medicine, performing comparative analysis over diverse patient populations is a fundamental step toward tailoring healthcare interventions. However, the critical aspect of equitably selecting molecular features across multiple patients is often overlooked. To address this challenge, we introduce FALAFL (FAir muLti-sAmple Feature seLection), a novel algorithmic approach based on integer linear programming. FALAFL is designed to bridge the gap between molecular feature selection and algorithmic fairness, ensuring a balanced selection of molecular features across different patient samples. We have applied FALAFL to the problem of selecting lineage-informative CpG sites within a cohort of 9 colorectal cancer patients subjected to low-coverage single-cell methylation sequencing. Our results demonstrate that FALAFL excels at rapidly determining the optimal set of CpG sites, which are well covered by cells across the vast majority of the patients, while ensuring that each patient contributes a similar number of sites to the final selection. An analysis of the FALAFL-selected CpG sites reveals that their lineage informativeness exhibits a strong association across diverse patient profiles, suggesting the non-stochasticity and universality of methylation changes in tumor evolution. Furthermore, these universally lineage-informative sites exhibit subclonal level alterations in the colorectal cancer primary tumors and are strongly associated with lineage-specific gene expression patterns. By integrating equity considerations into the molecular feature selection process, FALAFL is poised to propel equitable healthcare data science practices and advance our understanding of cancer and other complex diseases. Citation Format: Xuan Cindy Li, Yuelin Liu, Alejandro A. Schäffer, Stephen M. Mount, Süleyman C. Sahinalp. Balanced molecular feature selection unveils universally tumor-lineage informative methylation sites in colorectal 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 1615.

Abstract 1765: The genetic landscape of head and neck cancer using brush biopsy

Abstract Oral premalignant lesions (OPLs) with genomic alterations have a heightened risk of evolving into oral squamous cell carcinoma (OSCC). Currently, genomic data are obtained through invasive tissue biopsy. Brush biopsy has been utilized for diagnosing dysplasia but its effectiveness in reflecting the complete genomic landscape of OPLs remains uncertain. This study investigates the potential of brush biopsy samples in accurately reconstructing the genomic profile of OPLs. The evolution and heterogeneity were assessed by assessing SNVs, copy number analysis, and subclonal architecture reconstruction of paired tissue-brush biopsy samples of oral epithelium, dysplastic lesion, and OSCC lesion. We found that brush biopsy accurately reflects the genomic landscape of oral lesions, mirroring about 90% of SNVs and comparable CNA profiles found in tissue biopsies. Genomic profiling with brush biopsy was tissue-specific, as SNVs identified in OPL or OSCC lesions were not found in adjacent normal mucosa. Shared SNVs and CNAs were observed between OPL and OSCC samples. This suggested a common ancestor giving rise to these lesions. Subclonal architecture reconstruction confirmed that both lesion types shared a common ancestor clone and then diverged evolutionarily. These findings underscore the potential of brush biopsies in accurately reconstructing the genomic profile of OPL and OSCC, highlighting their usefulness in understanding the biological processes involved in tumor evolution. Citation Format: Evit John, Tom Lesluyes, Toby Baker, Maxime Tarabichi, Peter Van Loo, Xiao Zhao. The genetic landscape of head and neck cancer using brush biopsy [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 1765.

Abstract 1395: Unraveling the interplay between tumor heterogeneity and microenvironment in pancreatic ductal adenocarcinoma evolution and its clinical implications

Abstract Intratumoral heterogeneity (ITH) and tumor microenvironment (TME) of pancreatic ductal adenocarcinoma (PDAC) play important roles in tumor evolution and patient outcomes. However, the precise characterization of diverse cell populations and their crosstalk associated with PDAC progression and metastasis is still challenging. We performed single-cell RNA sequencing of primary PDAC samples with and without paired liver metastasis samples to understand the interplay between ITH and TME in the PDAC evolution and its clinical associations. We discovered that even a small proportion (~30%) of basal-like ductal cells in primary PDAC was associated with poor chemotherapy response and patient survival and that epithelial-mesenchymal transition programs were largely subtype specific. The clonal homogeneity significantly increased with more prevalent and pronounced copy number gains of oncogenes, such as KRAS and ETV1, and losses of tumor suppressor genes, such as SMAD2 and MAP2K4, along PDAC progression. Diverse immune cell populations, including naive SELLhi regulatory T cells (Tregs) and activated TIGIThi Tregs, contributed to shaping immunosuppressive TMEs of PDAC through cellular interactions with malignant ductal cells. The proportion of basal-like ductal cells negatively correlated with that of immunoreactive cell populations, such as cytotoxic T cells, but positively correlated with that of immunosuppressive cell populations, such as Tregs. Overall, our study provides a valuable resource for better understanding the roles of ITH and TME in PDAC evolution and their clinical relevance in the treatment of PDAC. Citation Format: Hyemin Kim, Hyong-oh Jeong, Se-Hoon Lee, Kwang Hyuck Lee, Kyu Taek Lee, Jong Kyun Lee, Min Woo Lee, Semin Lee, Joo Kyung Park. Unraveling the interplay between tumor heterogeneity and microenvironment in pancreatic ductal adenocarcinoma evolution and its clinical implications [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 1395.

Abstract 7602: Discovery of GH1581, a potent and selective WRN inhibitor as a cancer therapy

Abstract Defects in DNA mismatch repair (MMR) proteins could induce damage of genome, resulting microsatellite instability (MSI). MSI is highly correlated with the occurrence of several types of cancers, mainly in colon, gastric, endometrial, and ovarian cancers. Despite advances in the treatment of MSI-high cancers, tumor evolution and drug resistance remain the leading causes of treatment failure and death in cancer patients. About 50% MSI cancers have no response to immune checkpoint inhibitors. There is still unmet needs to be fulfilled for MSI-high patients. WRN is a RecQ family helicase involves in genome stability maintenance, DNA repair, replication, transcription, and telomere maintenance. Recent studies have identified WRN as a synthetic lethal target of MSI-high cancers. Employing structure-based drug design strategy, we have discovered a potent and selective WRN inhibitor, GH1581. In vitro, GH1581 displayed specific < 10 nM potency in ADP-Glo assay and < 10 nM potency in cell proliferation assay using MSI cancer cell lines. In vivo, GH1581 demonstrated favorable pharmacokinetic properties in different species. In SW48 xenograft model, it inhibited the tumor growth with 100% TGI at QD dosing of 10 mg/kg without severe adverse effect on body weight. In summary, we discovered GH1581 as a novel inhibitor of WRN with high in vitro and in vivo efficacy. Citation Format: Jiapeng Li, Guiping Zhang, Jie Jack Li. Discovery of GH1581, a potent and selective WRN inhibitor as a cancer therapy [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 7602.

Abstract 1759: Integrative longitudinal genomic analysis of therapy-resistant and metastatic pediatric cancers

Abstract Pediatric cancer patients are commonly profiled with gene-panel sequencing tests that yield few actionable results, in part due to the complex genomic alterations that define these malignancies. We hypothesized that integration of whole-genome (WGS) and RNA sequencing (RNAseq), would lead to a more comprehensive understanding of these diseases. Our study is uniquely focused on metastatic and relapsed disease, whereas previous studies focused on primary cases. We also prioritized longitudinal profiling, including with deep sequencing, to capture tumor evolution across primary and metastatic sites and to quantify the utility of resampling. We assembled a cohort of 219 high-risk pediatric oncology patients, including solid tumors, CNS tumors, leukemias/lymphomas, patients with relapsed/refractory disease (75), metastatic disease at diagnosis (7), rare diagnoses (17), prior cancer history, and estimated overall survival <50%. We characterized 286 samples with WGS (tumor ~60X; germline ~30X) and/or RNAseq (polyA selected, ≥20 million reads), including 95 samples taken from 44 patients at different time points (diagnosis, relapse, etc.). Variants, structural rearrangements, mutational signatures, and copy-number alterations were identified using WGS. RNAseq was used to profile gene expression outliers, gene fusions, and expression of variants. Integrated results were used to prioritize potentially actionable variants. For 20 patients (44 samples), we performed targeted deep sequencing of the DNA (~500X). RNAseq identified potentially druggable outlier gene expression and fusions, including 102 novel fusions where the 3’ gene is overexpressed. WGS identified aneuploidy, loss of heterozygosity and whole genome duplication across histotypes. Mutational burden and mutational signatures analyses identified profound effects from platinum drugs on tumor evolution, but treatment effects were not universal. Multiple sampling per patient uncovered drastic spatial and temporal differences in the genomes and transcriptomes of these tumors. Custom deep sequencing confirmed these findings, captured evolution missed by WGS, and furthered our understanding of the complex impacts of treatment on clonal evolution. Histotypes differed by whether actionability was higher for WGS or RNAseq alone, but in 66% of samples modality integration increased actionability. Integration also identified a subset of tumors that may be amenable to immunotherapy, but which lack canonical markers of response. Longitudinal analysis highlighted both the opportunities and risks of targeted therapy, with targetable variants gained and lost between timepoints. Our study shows that integrated multi-modality sequencing can elucidate novel insights into the biology of pediatric cancers and identify potential therapeutic targets not detected using gene-panel testing alone. Citation Format: Henry J. Martell, Avanthi T. Shah, Alex G. Lee, Stanley G. Leung, Soo-Jin Cho, María Pons Ventura, Ana Golla, Amanda E. Marinoff, Elizabeth P. Young, Bogdan Tanasa, Inge Behroozfard, Heng-Yi Liu, Aviv Spillinger, Michelle L. Turski, Nicole Elzie-Tuttle, Carlos Espinosa-Mendez, Arun Rangaswami, Tabitha M. Cooney, Cassie Kline, Anurag Agrawal, Jennifer Michlisch, Elliot Stieglitz, Mignon Loh, Amit J. Sabnis, Kieuhoa T. Vo, Sheri Spunt, Norman Lacayo, Holly C. Beal, Florette K. Hazard, Sophie Salama, David Haussler, Olena M. Vaske, Marcus R. Breese, E. Alejandro Sweet-Cordero. Integrative longitudinal genomic analysis of therapy-resistant and metastatic pediatric cancers [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 1759.

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 5665: Native nanopore sequencing of multiple tumor sites reveals genetic and epigenetic intra-tumor heterogeneity in canine osteosarcoma

Abstract Osteosarcoma (OS) is an aggressive bone malignancy affecting children and young adults. Implementation and optimization of neoadjuvant chemotherapy led to significant improvements in outcome, but patients with recurrence or metastases at diagnosis continue to have poor prognosis. Whole genome sequencing (WGS) revealed that a hallmark of OS is a highly rearranged genome with complex copy number aberrations rather than point mutations. A better understanding of the key genomic and epigenetic factors driving clonal evolution may provide insights into the tumor progression and the therapy response. Canine OS recapitulates the majority of the hallmark genetic and molecular factors of human OS and is recognized as an informative model. A significant difference is that initial surgery in canine OS allows us to explore tumor evolution unperturbed by treatment, providing invaluable resources to study the genetic and molecular landscape of tumor evolution. Here, we present the first study to date in which we performed bulk WGS of multiple tumor sites and distal normal tissue in two dogs using nanopore long-read sequencing to analyze the genetic and epigenetic evolution of OS. In both subjects, somatic structural variation detection using Severus revealed chromothripsis-like rearrangements in chr11, leading to the deletion of Cdkn2a, a known cell cycle regulator and tumor suppressor. Interestingly, in one of the subjects, the majority of the SVs (n = 131) shared across all the samples were localized on chr5, coinciding with perturbed copy number profiles and bi-allelic frequency of the SNVs, with all mutations identified using complementary bioinformatics analysis of the same native nanopore reads. The multi-site sample analysis identified subclonal structural alterations in TP53 and ATRX genes. With phased nanopore reads, we observed subject-specific copy-neutral loss-of-heterozygosity events and long identical-by-descent blocks. Haplotype-specific copy-number analysis of tumors showed different aneuploidy patterns in both subjects, highlighted clonal compositions of samples from distinct sites, and identified regions of copy-number heterogeneity. Furthermore, phased nanopore reads also revealed haplotype-specific 5mC and 5hmC methylation differences across tumor samples, where samples from different sites of the same tumor demonstrate distinct methylation patterns that overlap DLG2, RUNX2, KDR, and other genes known as mutational hotspots in canine/human OS. Overall, we showcase how a single sequencing assay allows for a joint analysis of genetic and epigenetic tumor alterations, thus providing a richer understanding of the drivers of somatic evolution and their association with the molecular landscape in the canine OS. These findings can be translated into human disease and shed light on tumor progression and therapy response. Citation Format: Ayse Keskus, Sergey Aganezov, Tanveer Ahmad, Xiaoguang Dai, Robert Walker, John Beaulaurier, Scott Hickey, Sissel Juul, Amy Kathleen LeBlanc, Paul S. Meltzer, Mikhail Kolmogorov. Native nanopore sequencing of multiple tumor sites reveals genetic and epigenetic intra-tumor heterogeneity in canine osteosarcoma [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 5665.

Abstract 4093: Modeling tumor immunoediting under immune selective pressure to inform neoantigen landscape dynamics for effective cancer vaccines

Abstract Tumor specific neoantigens, encoded by somatic mutations, are recognized by T cells, inducing anti-tumor immune responses. This renders neoantigens viable targets for personalized cancer vaccines. However, the identification of immunogenic neoantigens remains suboptimal. Understanding how immune selective pressure-mediated tumor evolution and neoantigen editing contribute to emergence of immunogenic neoantigens may improve prediction accuracy. In this study, we aimed to model tumor evolution and heterogeneity under immune pressure using three preclinical models of lung cancer with common driver mutations, Kras G12D/+ and p53 -/-, namely KPA, KPC, and HKP1, subcutaneously implanted in immunocompetent C57BL/6 (B6) and immune-deficient Rag1-/- mice. Our data suggest that despite shared driver mutations identified by whole exome and RNA-seq, tumor growth in vivo varied between tumor models. KPA tumors were immunogenic, as they were controlled in B6 mice but rapidly progressed in Rag1-/- mice, suggesting immunoediting of KPA-specific neoantigens. In contrast, HKP1 was non-immunogenic, with tumors progressing regardless of immune competency. KPC showed moderate tumor control in B6 mice, which was lost in Rag1-/- mice. We then estimated the cancer cell fraction on every mutation based on phylogenetic reconstruction. Immunogenic KPA tumors had significantly lower mutation burden in B6 than in Rag1-/- mice and the fraction of cell line mutations edited in vivo was significantly higher, suggesting active immunoediting in the former. Non-immunogenic HKP1 tumors did not show significant differences in mutation burden nor fraction of mutations edited between B6 and Rag1-/- mice. The immunogenicity was reflected in the immune infiltrate levels within tumors, with KPA being highly infiltrated by activated CD4 and CD8 T cells compared to KPC. HKP1 tumors showed increased infiltration of suppressive regulatory T cells. T cell receptor (TCR) sequencing on lung tumors showed that clonal expansion is strongly associated with immunogenicity. KPA tumors showed significantly higher TCR clustering with less diversity compared to KPC and more than HKP1. Following a simple approach to study the immunogenicity of shared neoantigens to elicit cross-protective anti-tumor responses, mice were immunized with whole irradiated (IR) tumor cell lines and implanted with live tumor cells (matched and unmatched). While immunization with IR-KPA protected against corresponding tumor implant and moderately against KPC, it did not protect against HKP1, indicating lack of immunogenicity of mutations shared between KPA and HKP1. Further studies testing the anti-tumor responses and characteristics of neoantigens shared between these three related tumor cell lines will inform the conditions under which tumors may escape or regress and help improve neoantigen identifying algorithms. Citation Format: Mariam Mathew George, Jayon Lihm, Hyejin Choi, Yuval Elhanati, Stephen Martis, Marta Luksza, Benjamin Greenbaum, Jedd D. Wolchok, Taha Merghoub. Modeling tumor immunoediting under immune selective pressure to inform neoantigen landscape dynamics for effective cancer vaccines [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 4093.