Research Areas

Prostate cancer subtypes, intratumor heterogeneity and genome evolution

Prostate cancer can be lethal, but is more commonly indolent, with PSA screening having led to over-diagnosis and overtreatment. A clinically important question is how to distinguish lethal from indolent disease. Our lab was among the first to identify clinically-relevant gene-expression and genomic subtypes of prostate cancer. Our recent efforts have focused on characterizing “5q/6q” prostate cancer, a TMPRSS2-ERG fusion-negative molecular subtype characterized by deletions on 5q (CHD1) and 6q, together with SPOP mutation. Ongoing studies, in collaboration with Jim Brooks (Urology) and Rob West (Pathology), seek to define intratumor heterogeneity and genome evolution, with respect to distinguishing lethal from indolent cancers.

Gene-expression signatures and biomarkers

DNA microarrays, and now RNAseq, enable the profiling of cancer transcriptomes, to identify signatures and biomarkers for improved diagnosis and outcome prediction. Our lab previously identified a prognostic gene-expression signature in acute myeloid leukemia, and novel prognostic biomarkers in prostate cancer, including AZGP1. The AZGP1 biomarker has since been validated by multiple labs, and has been incorporated into a commercialized multi-gene assay. We continue to discover and characterize biomarkers from transcriptomic data, including non-coding RNAs.

Recurrent structural alterations in cancer genomes

Our lab was among the early pioneers of array-based comparative genomic hybridization. Array CGH is used to map recurrent DNA amplifications and deletions in cancer genomes, thereby pinpointing novel cancer genes. Using this approach, we have discovered new amplified oncogenes, including CAMK1D and GAB2 in breast cancer, NKX2-1 and CRKL in lung cancer, GATA6 in pancreatic cancer, and CDX2 in colorectal cancer. Our recent efforts include strategies to identify recurrent breakpoints underlying cancer gene fusions. Ongoing studies rely on whole-genome sequencing, with a focus on understanding cancer genome evolution.

Amplified “lineage-survival” transcription factors

A major theme emerging from our array CGH studies has been the discovery of “lineage-survival” oncogenes. These are master transcriptional regulators of cell lineages that we have found to be amplified and overexpressed in a subset of tumors derived from those lineages. Examples include NKX2-1 amplification in lung cancer, GATA6 amplification in pancreatic cancer, and CDX2 amplification in colorectal cancer. Our efforts focus on understanding the altered transcriptional programs that drive lineage specification and differentiation during normal development, but abnormal cell proliferation and survival in cancer.

Altered chromatin remodeling in cancer

Another major theme to emerge has been the discovery of altered chromatin remodeling complexes, including CHD1 deletion in prostate cancer, and widespread SWI/SNF subunit deletion/mutation in pancreatic cancer. Current efforts center on understanding the role of altered chromatin remodeling in tumorigenesis, and developing therapeutic strategies that selectively target aberrant complexes.

Mechanisms of genomic instability

Our prior array CGH studies revealed different patterns of genomic alterations associated with the known gene-expression subtypes of breast cancer, suggesting distinct underlying mechanisms of genomic instability. Recent efforts have sought to uncover the mechanisms underlying the making and shaping of structural alterations in cancer genomes.

Less common neoplasms

While large genome consortia (e.g. TCGA) have now analyzed major cancer types, fewer systematic studies have addressed less common neoplasms that are of no less biologic interest and clinical importance to those patients affected. In collaboration with Rob West (Pathology), we have been carrying out genomic studies of ameloblastoma, an uncommon tumor arising from the ameloblast-lineage, the cells that deposit enamel during tooth development. Our recent studies have identified actionable driver mutations in most ameloblastomas, including mutations in the Hedgehog (SMO) and MAP-kinase (BRAF) pathways. Current efforts focus on further characterizing tumor pathobiology, as well as conducting preclinical and pilot clinical studies with molecularly-directed therapies.

Benign diseases

Recently, we have begun applying genomic approaches to the study of benign disease processes. For example, with Jim Brooks (Urology) and Rob West (Pathology), we have been characterizing the transcriptomes of benign prostatic hyperplasia. BPH is thought not to be a precursor of prostate cancer, but nonetheless represents a cellular proliferation with substantive impact on men’s health.

Patient-derived cell culture models

Not all cancer subtypes and genotypes are well represented by existing cell line models. Adopting the “Georgetown” method of conditional reprogramming (ROCK inhibitor plus irradiated fibroblast conditioned media), we have begun generating patient-derived cell culture models of varied cancer molecular subtypes.

Single-cell genomics

Cancer is marked by intratumor cell heterogeneity, with implications for disease progression and therapy resistance. We have begun to address cellular heterogeneity in cancer and other human diseases by single-cell genomics (including transcriptomes).