Our laboratory is pursuing combined genomic and functional approaches aimed to the characterization of human solid tumors, with an emphasis on melanoma.

A major thrust of our work in recent years has been to perform comprehensive genomic characterization of a large collection of patient-derived melanoma “short-term” cultures and cell lines, and to apply both computational and experimental approaches to identify the critical target genes and pathways enacted by these alterations. This approach has proved successful in several instances, including the identification of MITF as a “lineage survival” oncogene, and the discovery of the oncogenic role of the ETS transcription factor, ETV1, in melanoma. Presently, ongoing work involves the application of new generation sequencing approaches to identify novel chromosomal rearrangements, as well as systematic pooled RNAi screening to undertake a systematic functional interrogation of the melanoma genome.

With this detailed genomic information in hand, a parallel effort in the lab involves mechanistic studies of hallmark tumor pathways that are perturbed by the most prevalent genomic alterations in human solid tumors. Two examples include the MAP kinase pathway, which is activated by oncogene mutations involving BRAF or NRAS in the vast majority of melanomas, and the PI3 kinase pathway, which is activated by mutation in many human cancers. Our laboratory has identified mechanisms by which MAP kinase-dependent melanomas may become resistant to the targeted inhibitors in clinical development, as well as discovered novel effector genes activated downstream of MAP kinase in melanoma. Work in our laboratory has also nominated a novel oncogenic effector mechanism within the PI3 kinase pathway. Unlike PTEN-null cancers, which almost invariably show a strong dependence on AKT signaling, our laboratory has found that many cancer cells that contain activating PIK3CA gene mutations (PIK3CA encodes a catalytic subunit of PI3 kinase) may elaborate an AKT-independent signaling mechanism that engages PDK1 and SGK3.

The discovery of alternative oncogenic effectors has important implications for the development of rational therapeutics against cancer pathways, and they underscore the importance of linking specific tumor genetic alterations to key “druggable” cellular mechanisms, one of the overarching goals of our research.

Finally, with implementation of personalized medicine in mind, our group has applied several genomic technologies to create avenues by which critical tumor genetic alterations might be identified rapidly in the clinical or translational setting.

Altogether, these integrative genomic and experimental approaches applied to melanoma and other tumor types offer considerable promise to aid investigators and patients alike along the path to rational cancer therapeutics.


Systematic genomic studies in melanoma

It is now well established that cancer is a genetic disease. Genomic alterations such as point mutations, amplifications, deletions, insertions and translocations, are some of the ways in which oncogenes, tumor suppressors and stability genes are altered in a cancerous cell. Knowledge of these alterations can lead to a better understanding of the genesis of, and the possible mechanisms operating in, a tumor, and to novel therapuetic avenues. Therefore, one of the main focuses of our laboratory is to perform a comprehensive genomic characterization of a large collection of patient-derived melanoma tumor samples and cell lines to identify these alterations, and to identify the critical target genes and pathways enacted by these alterations. To this end, we are applying high-through put, cutting-edge sequencing and genomic analysis technology.

Study of resistance to BRAF/MEK inhibition in melanoma

Mutations in the BRAF kinase are one of the most common alterations in melanoma, with 50-70% of melanomas containing mutated BRAF. This makes BRAF and downstream effectors of the MAP kinase pathway, promising therapeutic targets. Recent phase I trials of a selective BRAF inhibitor showed promising clinical success in patients harboring BRAFV600E-mutant tumors. Unfortunately, it is very likely that tumors in patients treated with targeted therapeutics will develop resistance to MAPK kinase pathway inhibition. Therefore, discovery of emerging resistance mechanisms will provide valuable information to design effective and relapse-free treatments.

Our laboratory is taking several approaches to identify and “predict” resistance mechanisms to BRAF and MEK inhibition in melanoma. Recently, our group identified mutations within MEK that promoted in vitro resistance to a MEK kinase inhibitor (Emery et al). Excitingly, one of these in vitro-identified-mutations was later found in a tumor from a patient who had developed resistance to MEK-inhibition therapy. In parallel, our laboratory is also taking a high-throughput experimental approach to identify genes that will substitute for BRAF function upon pharmacologic inhibition of this kinase, thus identifying novel resistance mechanisms, and, very importantly, novel “druggable” therapeutic targets.

Functional characterization of melanoma tumor dependency

Genomic analysis of melanoma tumor samples reveal the presence of multiple amplifications and deletions events. Many of the identified chromosomal aberrations, however, consist of large, low-amplitude gains and losses that encompass dozens of genes, thus making the identification of the relevant effector mechanisms very challenging. An approach that our laboratory has taken to address this problem, is to perform high-throughput RNAi screens using lentivirally-delivered shRNAs. This approach is based on the assumption that silencing of essential genes will result in drastic reduction of cell populations containing the shRNAs targeting such genes. This reduction can easily be detected thanks to the presence of identifying barcodes added to each shRNA. Our current shRNA library consist of 55000 shRNAs, targeting 11000 genes. Thus, systematic pooled RNAi screening, offers an avenue for the systematic functional characterization of essential genes and cellular dependencies linked to chromosomal alterations on a genome scale.

AKT independent signaling in the PI3K pathway

Genetic alterations that activate the PI3 kinase pathway are highly prevalent in many human cancers. One such case is gain-of-function mutations in the PIK3CA gene, a gene that encodes a key enzymatic subunit of PI3 kinase. These activating mutations are thought to promote tumorigenesis primarily through downstream activation of the AKT kinase. However, we have recently discovered that many PIK3CA-mutant cancer cell lines employ an AKT-independent oncogenic signaling mechanism, relying, instead, on PDK1-SGK3 for viability (Vasudevan et al).Thus, deregulated PI3K may contribute to cancer through both AKT-dependent and AKT-independent mechanisms. These findings may have important implications for PI3 kinase signaling and the development of rational therapeutics against this key cancer pathway. Consequently, one of the main interests of our laboratory is to decipher the oncogenic properties of SGK3 kinase.

Human Cancer Cell Line Encyclopedia

In the current post-genomic era, tumor samples are being subjected to an extensive characterization of their genomes. This type of efforts has expanded our knowledge about cancer genetic alterations. The availability of comprehensive catalogues of such alterations, has shifted the emphasis towards figuring out which genes are targeted and which is their impact on cancer biology. The notion is that this exercise might reveal tumor targets that can be exploited therapeutically, as well as potential biomarkers of drug response, enabling the translation of all this information into therapeutic benefit for cancer patients.

Our group anticipated the need for a companion resource to systematically probe tumor biology armed with the cancer genomics knowledge being generated. In January of 2008, we launched a collaborative project, the Cancer Cell Line Encyclopedia (CCLE), between the Broad Institute, the Novartis Institutes of Biomedical Research (NIBR), and the Genomic Institute of the Novartis Foundation (GNF). The goal of this initiative is to conduct a detailed genetic and pharmacologic characterization of a large panel of human cancer models (~1000 human cancer cell lines), to develop integrated computational analyses that link distinct pharmacologic vulnerabilities to genomic patterns, and to translate cell line integrative genomics into cancer patient stratification.

High-throughput cancer mutation profiling

For several years now, cancer treatment has been moving towards a more personalized cancer medicine, were treatment decisions are based on the genetic makeup of individual cancers. Discovery of new therapeutic targets has lead to a rapid proliferation of targeted agents in development, and this in turn has called increasing attention to the importance of molecular profiling approaches that pinpoint in situ the tumors most likely to respond. Knowledge of tumor genetic alterations in the clinical and translational settings –including mutations, insertions and deletions, copy number alterations, and polymorphisms– should predict patient outcomes, inform treatment options, facilitate rational clinical trial design, and ultimately facilitate individualized approaches to cancer treatment. With all of this in mind, we are working to develop a platform which utilizes high-throughput and innovative sequencing technology to easily and quickly detect somatic genetic alterations in human tumors.

Identification of small molecules that perturb oncogenic transcription factors

One of the main goals of cancer research is to identify suitable drug targets for targeted therapy. Unfortunately, the majority of oncogenes, including transcription factors, are deemed undruggable by conventional means. In our laboratory we are trying to overcome this obstacle by using Small Molecule Microarray (SMM) to discover putative lead compounds able to modulate the biological activity of relevant oncogenic transcription factors. Small molecules, either derived from natural products or from synthesized compounds, are chemically printed on a chip. Soluble purified protein is added to the chip and the binding affinity for each compound is measured. The promising hits are confirmed by detailed Surface Plasmon Resonance (SPR) analysis and followed up with various functional assays.