Abstract Gene expression profiling has been widely used as a tool to characterize genome-wide changes in expression level in various cancers. Expression profiling is generally used to identify specific genes with altered expression or to identify patterns that distinguish cancer subtypes. More recently, several approaches to use expression data to identify interactions between genes or build cancer-specific networks have been described. We have previously demonstrated that cross-species gene expression analysis can be used as a tool to identify oncogene-specific alterations in gene expression. Cross-species analysis is useful for this because it allows for filtering of noise inherent in human gene expression analysis. The identification of an oncogenic Kras-specific gene expression signature was accomplished using this approach. A central focus of our laboratory has been to use this and other oncogene-specific signatures to identify new critical downstream pathways necessary for oncogene function. Using several computational approaches, we identified several upstream transcriptional regulators of our previously identified oncogenic Kras signature. We hypothesized that some of the transcriptional regulators represent specific dependencies that are required for Kras to function as an oncogene. Thus, inhibition of the transcription factors could uncover novel synthetic lethal interactions specific for cells that express oncogenic Kras. Using an shRNA pooled library, we have confirmed one such synthetic lethal interaction betweeen oncogenic Kras and the Wilms-tumor 1 (Wt1), a gene known to function as a tumor suppressor in some contexts. Loss of Wt1 induces senescence in primary cells expressing oncogenic Kras. Human non-small cell lung cancer cells expressing oncogenic Kras also become senescent when Wt1 is knocked-down using shRNA. Thus, synthetic senescence is induced on Wt1 loss in oncogenic Kras expressing cells. A role for Wt1 in Kras-driven oncogenesis was further confirmed using a mouse model for lung cancer in which a mouse carrying a conditional Kras allele was crossed a mouse carrying an allelel that allows for conditional deletion of Wt1. The biological signficance of this intereation was further demonstrated by using a “Wt1 “high” or “Wt1 low” signature to classify human NSCLC gene expression profiles. This allowed us to determine that human tumor samples that express oncogenic Kras and express high levels of Wt1-regulated genes have a worse prognosis compared to samples expressing oncogenic Kras but having low low level of Wt1-regulated genes. In human NSCLC cell lines, loss of Wt1 leads to significant upregulation of the cell cycle regulator p21. Wt1 appears to regulate p21 levels posttranscriptionally, as p21 mRNA levels did not change. Further studies to elucidate the mechanisms leading to p21 upregulation upon Wt1 loss are underway and preliminary results will be discussed. To take these studies further, we began by generating a more refined Kras-specific signature using FACs-sorted tumor cells from the LSL-KrasG12D mouse model crossed to a fluorescent reporter. We have used this mouse-generated Kras signature for cross-species analysis using a much larger NSCLC dataset than was possible in our previously published studies. We then applied the ARACNE algorithm to identify transcription factors that regulate this Kras-signature. Using this approach, we have identified other transcription factors that are potential regulators of oncogenic Kras. The fact that Wt1 once again was identified as a regulator of oncogenic Kras using this novel method suggests the validity of this approach. Preliminary results testing the functional significance of these transcription factors for Kras-driven oncogenesis will be discussed. 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 SY01-03. doi:10.1158/1538-7445.AM2011-SY01-03
Read full abstract