Studies of genomic alterations and therapeutic vulnerabilities in cancer

Abstract

Next-generation sequencing (NGS) is a key technology for multiple fields of cancer research. The output of NGS has expanded diagnostic, prognostic, and therapeutic implications for precision oncology. The introduction of NGS has enabled high-throughput and accurate discoveries of genetic, epigenetic, and transcriptional alterations in cancer genomes, rapidly unraveling the landscape of tumors harboring complex genetic architectures. The ability to use NGS to quantify experiments performed with RNA interference or clustered regularly interspaced short palindromic repeats (CRISPR)-Cas9 systems allows pooled genome-wide screens to interrogate gene functions systematically in diverse cancer contexts, leading to the identification of numerous promising therapeutic targets and subsequent target-specific drug discovery. Ileal neuroendocrine tumor (NET) is the most common tumor of the small intestine. About half of ileal NET patients develop multiple synchronous tumors at diagnosis. However, whether these tumors arise due to germline predisposition or as independent clones, or as disseminations of a single primary tumor is largely unknown. I was interested in leveraging the use of NGS to shed light on the clonal relationship between individual tumors from the same patients. In Chapter 2, I describe a high-throughput sequencing analysis of 40 multifocal ileal NET from three patients. I found that chromosome 18 loss of heterozygosity occurs in 55% of tumors profiled. I further phased the germline single nucleotide polymorphisms (SNPs) on chromosome 18 and show that individual tumors from the same patients may lose different alleles of chromosome 18. Our data support the theory that multifocal ileal NETs are likely polyclonal. Besides conducting NGS experiments to decipher the genomic underpinnings of tumors in Chapter 2, we are also interested in uncovering novel tumor vulnerabilities and therapeutic targets in cancer through combined analysis with genome-wide functional screening data, cancer patient data, and functional validation. Dysregulation of the cyclin-dependent kinase (CDK) 4 and 6 pathway is commonly observed in human cancer; inhibitors of CDK4/6 have been clinically approved for the treatment of hormone-receptor-positive breast cancer. However, the clinical benefit of CDK4/6 inhibitors can be hindered by on-target hematological adverse effects as well as inadequate effectiveness resulting from primary and adapted resistance to CDK4/6 inhibition. Therefore, understanding the requirement for CDK4/6, and elucidating diverse resistance mechanisms to CDK4/6 inhibition, would provide important insight into future CDK4/6-based therapy optimization. In Chapter 3, I sought to understand the essentiality of CDK4 and CDK6 across human cancer cell lines. Through analysis of genome-wide loss-of-function CRISPR/RNAi data combined with functional assays, I delineated a CDK4/6 “essentiality map”. I identified that CDK6 gene expression is the best single predictor for CDK4 and CDK6 gene dependencies, and that cancer subtypes with higher CDK6 expression are more dependent on CDK6 for cell proliferation, while CDK4 is more important for cancer types with lower CDK6 expression. We further show that the CDK2-CCNE1 axis is an important alternative dependency for cell lines that require neither CDK4 nor CDK6. In Chapter 4, we show that the use of CDK4/6 inhibition alone is not effective enough long-term in esophageal squamous cell carcinoma (ESCC) models, even though the majority of ESCC patients harboring genomic alterations affecting CDK4/6-RB-CDKN2A pathway. Through data mining and functional genomics, we demonstrated that ERBB kinase family proteins represent a preferential baseline dependency for ESCC cells, and inhibitors of the ERBB family proteins reveal enriched sensitivity to the same cell line group. We then demonstrated that combining ERBB inhibitor and CDK4/6 inhibitor significantly improves growth inhibition in ESCC models in vitro and in vivo. In summary, results from Chapter 2 suggest that there is not a particular germline chr18 allele that predisposes somatic LOH; and these individual tumors likely arise independently. Results from Chapter 3 report variable CDK4/6 gene dependencies across human cancer cell lines, which may shed light on novel CDK4/6-based therapy development. Our work from Chapter 4 provides preclinical evidence that combining pan-ERBB inhibitor and CDK4/6 inhibitor may be a promising therapeutic strategy for ESCC patients. Taken together, these studies highlight the use of NGS-enabled large-scale genomic sequencing and genome-wide functional screening in cancer research.