We work to identify and understand the genomic events that cause human cancers, concentrating on lung cancer. We are also seeking to discover infectious causes for diseases of unknown origin.
Somatic genetic alterations in cancer: We use genome-scale approaches to discover chromosomal alterations and cancer-causing mutations. I am a principal investigator for NCI's 'The Cancer Genome Atlas' project, or TCGA, for comprehensive cancer genome characterization. I serve as co-chair of the executive committee for TCGA and co-chair of the lung cancer disease working group with Drs. Govindan and Baylin. In squamous cell lung carcinoma, we identified loss-of-function HLA-A mutations, the first somatic evidence for immune evasion in cancer (TCGA, Nature, 2012).
We developed the use of single-nucleotide polymorphism (SNP) arrays for human cancer genome analysis (Lindblad-Toh et al., Nature Biotech, 2000). We have now defined both lineage-specific and cancer-universal regions of amplification and deletion by SNP array analysis of over 7,000 cancer DNAs. Using SNP arrays, we identified the most common DNA amplification in lung adenocarcinoma, which targets the NKX2-1 pneumocyte-specifying transcription factor (Weir et al., Nature, 2007), common amplification of the SOX2 transcription factor in squamous cell carcinomas (Bass et al., Nature Genetics, 2009), and amplification of anti-apoptotic genes including MCL1, across multiple human cancers (Beroukhim et al., Nature, 2010). Recently, we have collaborated with the Beroukhim on pan-cancer copy number analysis in TCGA, identifying patterns of whole genome doubling and frequent amplification of epigenetic regulators (Zack et al., Nature Genetics, 2013).
Our cancer sequencing projects identified mutations in the epidermal growth factor receptor tyrosine kinase gene, EGFR, in lung adenocarcinomas, associated with clinical response to gefitinib and erlotinib (Paez et al., Science, 2004), and in glioblastoma (Lee et al., PLoS Med, 2006). We also identified activating mutations of FGFR2 in multiple cancers (Dutt et al., PNAS, 2008), of ALK in neuroblastoma (George et al., Nature, 2008), and of the ERBB2 extracellular doain in lung adenocarcinoma (Greulich et al., PNAS, 2012), and mutations of APC, NF1, and ATM in lung adenocarcinoma (Ding et al., Nature, 2008).
Our group pioneered the use of next-generation sequencing in cancer genome analysis (Thomas et al., Nature Medicine, 2006), which we are now applying widely. Recent projects have identified mutation of MAP3K1 and CBFB in breast carcinoma (Banerji et al., Nature, 2012), translocations of the TCF7L2 gene in colon carcinoma (Bass et al., Nature Genetics, 2011), and mutations of multiple genes including U2AF1 and RBM10 in lung adenocarcinoma (Imielinski et al., Cell, 2012). We have also identified translocations of STAT6 in solitary fibrous tumor (Chmielecki et al., Nature Genetics, 2013), and mutations of DDX3X in medulloblastoma (Pugh et al., Nature, 2012).
Functional analysis of lung cancer genes: We study oncogenic transformation by the major oncogenes that cause lung cancer, including EGFR and NKX2-1. For EGFR, we demonstrated the concept of mutation-selective therapy: distinct mutations are differentially sensitive or resistant to inhibitors (Greulich et al., PLoS Medicine, 2005). For NKX2-1, we have now identified LMO3 as a downstream target (Watanabe et al., Genes & Development, 2013). We have also recently identified multiple drug-sensitive alterations in the squamous cell lung cancer genome: mutations of DDR2 (Hammerman et al., Cancer Discovery, 2011), FGFR2 and FGFR3 (Liao et al., Cancer Research, 2013) and amplifications of FGFR1 (Dutt et al., PLoS One, 2011).
Tumor suppressor proteins and chromatin modification: We showed that several endocrine tumor suppressor proteins, including menin and parafibromin, are associated with histone methyltransferases (Hughes et al., , Molecular Cell, 2004; Rozenblatt-Rosen et al., Mol Cell Biol, 2005). We are now working to modulate histone methylation activity for treating endocrine tumors and leukemias, and have showed that deletion of the Rbp2 histone demethylase can reverse tumorigenesis by loss of the menin tumor suppressor (Lin et al., PNAS, 2011).
Discovery of pathogenic microbes: We developed a genomic approach to discover microbial sequences in human disease, by sequencing nucleic acids from diseased tissues followed by computational subtraction of human sequences (Weber et al., Nature Genetics, 2002). We developed a new software approach for identifying pathogens using next-generation sequencing data (Kostic et al., Nature Biotech, 2011), have identified an association of Fusobacterium species with colorectal carcinoma (Kostic et al., Genome Research, 2012) and have demonstrated that Fusobacterium infection potentiates lung cancer in mouse models (Kostic et al., Cell Host & Microbe, 2013). Recently, we identified a novel bacterium, Bradyrhizobium enterica, associated with the transplant-associated cord colitis syndrome (Bhatt et al., New England J Med, 2013).