My laboratory has a long-standing interest in the genetics of human cancer, including genes implicated in Wilms' tumor, a pediatric kidney cancer, and in susceptibility to breast cancer, as well as somatic mutations associated with specific sensitivity to targeted cancer therapies. We have two major research foci: identification of genetic lesions underlying cancer drug susceptibility through the molecular analysis of circulating tumor cells and analysis of aberrant genetic pathways in the differentiation of pluripotent renal precursors leading to Wilms' tumorigenesis.
Molecular Genetics Underlying Targeted Cancer Therapeutics
Somatic activating mutations in the epidermal growth factor receptor (EGFR) were identified in our laboratory in the subset of non-small-cell lung cancer (NSCLC) with dramatic clinical responses to the tyrosine kinase inhibitors targeting EGFR, gefitinib (Iressa) and erlotinib (Tarceva). These mutations lead to altered ligand-dependent activation of the receptor, as well as to enhanced inhibition by gefitinib. Indeed, NSCLC tumor cells harboring activating mutations of the EGFR kinase appear to be extraordinarily dependent on survival signals that are differentially mediated by these mutant receptors, suggesting that the phenomenon of oncogene addiction may underlie their exceptional sensitivity to kinase inhibitors. Like chronic myeloid leukemia harboring the BCR-ABL translocation, which is sensitive to imatinib (Gleevec), and gastrointestinal stromal tumors with mutations of KIT, EGFR-mutant lung cancer constitutes a genetically defined subset of a common epithelial cancer that is uniquely susceptible to targeted cancer therapy with tyrosine kinase inhibitors.
Clinical studies have indicated that EGFR-mutant NSCLC constitutes a specific subset of lung adenocarcinomas (often with bronchoalveolar histological features) that arise in nonsmokers and are more prevalent in women and in individuals of East Asian ethnic background. Although treatment with EGFR inhibitors may lead to a modest survival improvement in chemotherapy refractory patients with EGFR–wild-type NSCLC, treatment of EGFR-mutant cases typically results in rapid and profound tumor shrinkage, consistent with the apparent tumor dependence on EGFR signaling.
The identification of molecularly defined subsets of epithelial cancers with unique genetic lesions underlying specific drug susceptibility poses both an opportunity and a challenge for successful targeted therapy for cancer. Matching drugs and genotypes is particularly important because relatively rare genotypes are identified across different histological subtypes of tumors. For instance, we recently found that high-level amplification of the growth factor receptor gene c-MET identifies a subset of gastric adenocarcinomas likely to respond to novel inhibitors of the MET tyrosine kinase, leading to the initiation of a genotype-directed clinical trial.
To extend these modeling studies of targeted cancer therapies, we are collaborating with Jeff Settleman, director of the Massachusetts General Hospital Center for Molecular Therapeutics, to conduct high-throughput drug screens on a very large panel of genetically characterized cell lines designed to capture the genetic heterogeneity of diverse human cancers. Early results have confirmed drug/genotype correlations in EGFR-mutant and MET-amplified cancers across different tumor types and uncovered additional predictors of therapeutic potential.
The response of EGFR-mutant NSCLC to EGFR kinase inhibitors is often limited by the rapid emergence of drug resistance, often associated with secondary EGFR mutations that affect drug-binding affinity. Our studies of NSCLC cells with acquired resistance to gefitinib led to the characterization of a novel class of irreversible inhibitors of EGFR, which appear to circumvent such resistance and are now entering clinical trials. These approaches highlight the need to serially monitor tumor genotypes as these evolve during therapy and progression of epithelial cancers.
We are collaborating with Mehmet Toner, director of the Massachusetts General Hospital BioMEMS laboratory, in characterizing a novel nanofluidic device capable of isolating circulating tumor cells (CTCs) from the blood of patients with cancer. This CTC chip relies on flow of blood through microposts coated with antibody to the epithelial marker EpCAMto capture CTCs with high efficiency (captured cells, 50–500 cells/ml; purity, 50–80 percent). We have shown that during the course of therapy, the number of captured cells correlates with radiological measurement of tumor response and that the isolated CTCs can readily be used to define molecular markers characteristic of the underlying malignancy. In addition to its potential applications in early detection of cancer and in the identification of prognostic markers, this approach offers an unprecedented opportunity to noninvasively monitor tumor genotypes over the course of treatment, which we believe to be critical to the successful application of targeted cancer therapies.
The ability to isolate CTCs from the blood of patients with diverse epithelial cancers also provides an exceptional opportunity to study the mechanism of blood-borne metastasis, the primary cause of cancer-related deaths. Epithelial-to-mesenchymal transition (EMT) is thought to constitute a key mediator of tumor invasiveness, reflecting the acquisition of migratory properties enhancing vascular invasion. In a genome-wide screen for gene copy-number alterations in a mouse tumorigenesis model, we identified specific amplification of the gene encoding YAP, a critical effector of the recently uncovered Hippo signaling pathway. YAP is a potent inducer of EMT, providing a powerful link between this alteration in epithelial cell fate and invasive tumorigenesis. Current studies are aimed at defining molecular mechanisms underlying EMT and defining its role in human cancer metastasis through the analysis of circulating tumor cells.
Wilms' Tumor
In contrast to EMT, mesenchymal-to epithelial transition (MET) underlies development of the normal kidney and is a process that appears to be disrupted in the pluripotent embryonal renal cancer of childhood, called nephroblastoma or Wilms' tumor. Wilms' tumor has been linked to inactivation of the WT1 gene, encoding a zinc finger transcription factor that is required for development of the kidney, gonads, and mesothelial tissues. Germline mutations in WT1 result in genetic predisposition to Wilms' tumor; somatic mutations and a chimeric fusion with the EWS gene contribute to the development of Wilms' and mesothelial tumors, respectively. Retrovirally directed expression of WT1 in primary hematopoietic cells triggers lineage-specific differentiation.
We recently used comparative genomic hybridization to identify a novel Wilms' tumor–suppressor gene located on the X chromosome, which we named WTX. Like WT1, WTX is expressed in early renal precursors and shows a tight developmental regulation of expression. Mutations in WTX are found in up to 30 percent of sporadic Wilms' tumors, and interestingly, deletions and intragenic mutations targeting WTX display one-hit inactivation, targeting the single X chromosome in males or the single active X chromosome in females. WTX is the first tumor-suppressor gene identified on the X chromosome, and current studies are aimed at defining its functional properties.
We are expanding our search for genetic causes of Wilms' tumor by examining alterations in loci encoding microRNAs. These small noncoding RNAs have been implicated in cellular differentiation in multiple systems, and their contribution to human cancer is under intense investigation. We recently described biological mechanisms underlying the regulation of their biogenesis by Argonaute 2 (AGO-2). Further studies are aimed at defining potential roles for subsets of microRNAs in normal renal development and the genesis of Wilms' tumor.
