Many of the genes that are mutated in human cancers are known.  These include both oncogenes like ras and tumour suppressor genes like p53 and p105Rb. While many of these mutations are common, no single mutation is sufficient to drive full-blown cancer. Instead, mutations in multiple genes cooperate.

  • What are the rules governing such interactions?
  • Why do certain combinations of mutations synergise strongly whereas others appear to have no additive effect whatsoever?

Many of the pathways that are mutated in cancer are highly conserved and our approach to understanding these problems is to use RNAi in the worm to screen through thousands of potentially interacting genes and thus identify which rare combinations of cancer-related genes interact functionally. In this way we hope to see patterns in these interactions and use these data to predict genetic interactions in cancer in humans.

RNAi screens for modifiers of ras signaling
Ras is mutated in a high proportion of human cancers; however, mutation of ras is not sufficient to drive full-blown carcinogenesis but other cooperating mutations are required. In the worm, the situation is similar. Gain-of function mutations in let-60, the worm ras orthologue, yield a multivulval (Muv) phenotype; however only ~60% of worms show this and the rest appear wild-type (see Fig1). We are carrying out large-scale RNAi screens in the worm to find genes that cooperate with ras mutations and whose loss changes increases the penetrance of this Muv phenotype. Once we identify candidates, we test in cell culture and in knockout models whether mouse orthologues also cooperate with ras-driven tumours. In this way we use the power of genome-scale RNAi screens in the worm as a discovery and follow-up all our hits systematically to determine their relevance in cancer.

Systematic mapping of genetic interactions in signaling pathways
Many of the genes mutated in cancer encode components of highly conserved signal transduction pathways.  In general, mutations in more than one gene combine in a simple additive way; however occasional combinations of mutations combine and highly syngergistic ways. Identifying these rare syngergistic combinations will shed crucial light on how similar mutations combine in cancers.

We have used RNAi to map genetic interactions in a systematic manner, examining ~65,000 possible interactions to date (see Fig2). We find that in worms (as in yeast) ~1% of genes interact, and that we can use these interactions to identify new components of these conserved pathways. Most genes interact only with a single pathway — e.g. sos-1 only interacts with mutations in the EGF signaling pathway. Intriguingly, however, we find that a small handful of chromatin regulating genes interact with every single pathway we have examined. The implications for genetic disease are clear: mutations in these chromatin regulators will affect the progression of every genetic disease. These conclusions were only possible through the analysis of systematic datasets rather through hypothesis-biased experiments.