NEW YORK (GenomeWeb) – The Wellcome Trust Sanger Institute in collaboration with Astex Pharmaceuticals has launched a resource based on its COSMIC (Catalogue of Somatic Mutations in Cancer) database that provides 3D structures of proteins featuring cancer mutations.
Called COSMIC-3D, the database combines COSMIC's cancer mutation data with protein sequence data from the European Bioinformatics Institute' UniProt database and structural information from the Worldwide Protein Data Bank (wwPDB) to present a "human structural proteome of oncology," said Harry Jubb, a staff scientist and postdoctoral researcher at Sanger and one of the leaders of the project.
Jubb, whose postdoctoral fellowship is funded in part through Astex's Sustaining Innovation program, said that he believes that the resource, in addition to providing insights into basic cancer biology, would prove particularly powerful for identifying cancer drug targets and working out drug designs.
By integrating data from COSMIC, UniProt, and the wwPDB "we are putting mutations in their protein structural context, and we can use that to generate hypotheses about how the protein structural nature of these mutations impacts cancer biology and drug design," he said.
"We've done some structure-based predictions of [protein] drug pockets to try to understand how the mutations affect them," he added. "We have done some very basic modeling of missense mutations that is available in COSMIC-3D and that is very powerful for trying to understand how a mutation affects the shape of a binding site and how we can design a drug in response."
Jubb also noted that the database could help researchers better understand how cancer mutations affect protein-protein interactions. The tool lets users look at specific cancer mutation hotspots in the context of protein structures, which will allow researchers to "see exactly where the most important mutations are in [a given protein-protein] interaction," Jubb said."So it's a powerful tool for understanding the underlying biology."
He added that this could also aid drug-discovery efforts focused on disrupting protein-protein interactions, as opposed to inhibiting single proteins. Structural information like that provided by COSMIC-3D "is hugely important in drug design for protein-protein interactions where you are trying to find a drug to fit into a space that is not particularly well-defined," he said. "If any kind of mutation can impact that binding, then that is hugely important information in terms of elucidating drug discovery in [protein-protein interactions]."
While protein-protein interactions are still a relatively new target for drug development, such approaches hold promise, particularly in diseases like cancer where drug resistance complicates treatment.
As George Mason University research Emanuel Petricoin noted in an interview last year, cancer drugs like kinase inhibitors that target enzymatic activity have proven susceptible to resistance due to the ability of tumors to develop mutations that allow, for instance, the targeted ATP-binding domains to get around the competitive inhibition used by such drugs.
"But if we can identify inhibitors that can disrupt protein-protein interactions, those will be much more difficult [for the tumor] to develop resistance to, because evolutionarily it is harder to wire around," he said.
The resource currently provides structural models of more than 8,000 proteins with cancer-linked mutations and will be updated every three months with new mutational, structural, and computational data, said Simon Forbes, head of the COSMIC database.
He noted that, much like the original COSMIC database, the goal is to provide a comprehensive look at the cancer mutational landscape.
"We are coming up on around 1,000 genes that are properly implicated [in cancer] with known drivers," Forbes said, but in addition to mutations occurring in these genes, the database "catalogues all the mutations observed in all cancers across all the genome," he said.
"So what that is doing is not necessarily telling you that every gene is a cancer target, but it gives you a lot of genetic data to enable [researchers] to explore novel targets," he added. "There are the [roughly] 1,000 genes we know cause cancer, and all of those are targetable for drug design using the COSMIC-3D system. But beyond that there are genome-wide swaths of additional data out there that are available for investigation."
The structural models are generated computationally, Jubb said, adding that the Sanger researchers had not done experimental validation of these models, but considered them a starting point for other researchers.
"The idea is more to empower researchers to go out and be able to do those experimental validations," he said. "It's a powerful platform for exploring the biology, exploring where the mutations occur and then using that as a starting point for drug discovery. It is trying to translate from basic cancer genomics research to more of an early-stage drug design perspective, where you can start to investigate the roles [of particular mutations] based on a more knowledge-driven approach."
Astex has provided funding for the project, but Forbes said that having a drugmaker on board is also important in terms of ensuring that the system is "as valuable and relevant as possible for drug discovery and design."
Jubb said that the resource has thus far been well received internally at Astex, where structure-based drug design is a point of focus.
"The real impact has been [the ability] to take this complex genomic data that is hugely available, but underutilized, I think, in pharmaceutical [firms], where there is a desperate need for new targets," he said. "Being able to have the whole of the structural human proteome mapped with rich [cancer mutation] annotation from COSMIC, and the understanding of mutation hotspots and which are important, gives [drugmakers] a much more impactful perspective."