Abstract SY15-02: Comprehensive analysis of genetic and epigenetic factors contributing to tumorigenesis in retinoblastoma

Abstract

Abstract Decades of research on tumor cell biology has led to consensus that most cancers can be defined in terms of a small number of shared biological properties or hallmarks. Tumor cells activate growth-signaling pathways, evade cell death and senescence, acquire limitless replicative potential, sustain angiogenesis and may eventually invade surrounding tissues and metastasize. Molecular genetic studies have provided important examples of the direct connection between genetic lesions in tumor suppressor genes or proto-oncogenes and the corresponding pathways that are deregulated during tumorigenesis. For example, TP53 mutations allow cancer cells to evade cell death and senescence while RB1 mutations promote limitless replicative potential. We sought to simplify the integration of cancer genetics with the hallmarks of cancer by focusing on a simple, homogeneous tumor type that progresses rapidly and initiates with a common genetic lesion. Retinoblastoma is a rare cancer of the retina that begins during fetal development and is diagnosed in the first years of life. Virtually all retinoblastomas are believed to initiate with biallelic RB1 gene inactivation, which is rate-limiting for tumorigenesis. Here we present comprehensive genetic and epigenetic analysis of 4 primary retinoblastomas, their matched peripheral blood and an orthotopic xenograft derived from one of the primary tumors. Genetic or epigenetic inactivation of both copies of the RB1 gene was confirmed in all 4 primary tumors. Overall, our data suggest that epigenetic mechanisms may be playing a more important role in retinoblastoma tumorigenesis than previously appreciated. In addition, the orthotopic xenograft retained genomic stability through multiple passages in animals. Therefore, our orthotopic retinoblastoma xenograft provides an ideal experimental system to begin to directly connect the molecular genetics of cancer with sensitivity to novel chemotherapeutic agents. Recent advances in high-throughput sequencing technologies have allowed researchers to produce comprehensive profiles of the mutations and chromosomal lesions in individual tumors for the first time. This is important because in theory, the complete spectrum of genetic lesions in a particular tumor can now be directly mapped to each of the corresponding biological hallmarks of that tumor. Eventually, such analyses could be used to select the most appropriate treatment regimens for individual patients based on the spectrum of genetic lesions in their tumor and our understanding of how those lesions relate to the functional properties of the tumor cells themselves. However, cancer genome analysis has proven to be more complex than previously anticipated. In particular, the magnitude and complexity of the genetic lesions in some human cancers has made it difficult to unambiguously assign individual genetic lesions to the corresponding pathways that contribute to the hallmarks of cancer. For example, in a recent paired genome sequence analysis of a tumor from a lung cancer patient, 530 single nucleotide variants were identified in the tumor from 392 different coding regions. In addition, this tumor contained 43 large-scale structural variations. This genetic complexity is further complicated by difference across tumor types, differences among individual tumors for a particular tumor type and in some cases, lack of a clear understanding of the order of genetic lesions during tumorigenesis. Moreover, in tumors such as the aforementioned lung tumor, with high rates of genome instability that may take decades to be diagnosed, it is difficult to definitively separate those lesions that directly promote tumorigenesis from those that play no role in cancer initiation or progression. In this study, we sought to simplify the integration of cancer genetics with the hallmarks of cancer by focusing on a simple, homogeneous tumor type that progresses rapidly and initiates with a common genetic lesion. We performed whole genome and transcriptome sequence analysis of human retinoblastoma. There are several advantages to using retinoblastoma for this type of study. First, retinoblastomas are relatively homogeneous developmental tumors with little if any difference in incidence across gender, race or region. Second, all retinoblastomas are believed to initiate with the same genetic lesion_RB1 gene inactivation. This simplifies the cancer genetics tremendously because we do not need to parse out subgroups that initiate with different lesions. Third, RB1 gene inactivation is rate-limiting for tumorigenesis suggesting that the secondary and tertiary mutations occur rapidly as a direct or indirect result of RB1 inactivation. This may help us with our analysis because the time for initiating genetic lesion to tumor diagnosis is rapid and may minimize the accumulation of mutations or chromosomal lesions that do not contribute to tumorigenesis. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr SY15-02. doi:10.1158/1538-7445.AM2011-SY15-02