Massive
efforts are being invested worldwide in cataloguing the mutations and
transcriptomic changes characterizing various cancers, in order to provide a
starting point for identification of the cellular pathways involved in cancer. Yet
the typical individual analysis of the identified mutations and transcriptomic
changes enables only limited understanding: the functions of many mutations and
the pathways upstream the transcriptomic changes often remain hidden. Here we report the
application of a novel integrative approach that takes these exciting new data
into the next step of identifying cellular pathways involved in cancer. Specifically,
we integrated data of 20
melanoma-associated mutations and 1,500 transcripts that were differentially
expressed in melanoma cell lines, with extensive data of over 157,000 known
molecular interactions among human proteins, genes and micro-RNAs (miRs). Using
a network optimization technique, we identified a melanoma sub-network containing
around 170 proteins and miRs that connects melanoma-associated mutations and
transcripts. The network included well known cancer-related proteins,
such as MYC and E2F1, and was
highly enriched for pathways already associated with cancer and melanoma, such
as the MAPK pathway and response to UV. Importantly,
the network unraveled several intriguing relationships among melanoma
related genes, such as the functional relationship between the transcription
factor E2F1 and the tumor suppressor protein PTEN. E2F1, which is typically associated
with poor survival prognosis in patients, was found to be over-expressed in melanoma
and other cancers. PTEN was down-regulated in the majority of the melanoma cell
lines, however only a small fraction of the cell lines contained PTEN mutation.
Our network analysis revealed that the oncogenic mir-19a that regulates PTEN
levels is under transcriptional regulation of E2F1, thus providing a molecular mechanism
linking E2F1 over-expression with PTEN down-regulation. This mechanism and
others we uncovered demonstrate
the value of our integrative network approach and its applicability to the
accumulating state-of-the-art cancer catalogs.
