By Monica Heger
In a step towards personalized cancer treatment, researchers have identified new drug targets in prostate cancer, gastric cancer, and melanoma that they said could be attacked by drugs currently in development.
Reporting in Nature Medicine this week, the authors used a paired-end transcriptome sequencing approach to identify gene fusions in the RAF kinase pathway, which they said could represent a new, rare, subtype of cancer. This week's paper is a follow-up of a proof-of-principle paper the authors published last year demonstrating that transcriptome sequencing could identify gene fusions (GWDN 6/24/2009).
Many prostate cancers contain a gene fusion in the ETS transcription factor gene, but the protein products for ETS fusions are considered undruggable because it is difficult to tailor drugs to act on transcription factors. So the researchers wanted to examine whether they could identify other gene fusions that would be more amenable to drugs.
"In this case we identified a kinase, and small-molecule drugs could target …kinases. So these are potentially targetable gene fusions," said Arul Chinnaiyan, director of the Michigan Center for Translational Pathology and senior author of the study.
He and his colleagues sequenced the transcriptomes of five prostate cancers that contained the ETS gene fusion, as well as 10 prostate cancers that did not have the gene fusion. The researchers used the standard mRNA-seq protocol provided by Illumina and a paired-end sequencing strategy on the Genome Analyzer with 40 base pair reads, generating 148 million mate pairs, about 69 percent of which were mappable.
In two of the ETS-negative samples, they identified mutations in the RAF kinase pathway that were not present in any of the ETS-positive samples, and confirmed the mutations with qPCR. One sample contained a fusion between SLC45A3 and BRAF, while the other mutation was a fusion between ESFRP1 and RAF1. Both the BRAF and RAF1 genes are involved in the RAF kinase pathway.
The researchers then analyzed additional prostate cancer cohorts using florescence in situ hybridization and found rearrangements in the RAF kinase pathway in 10 out of 799 samples. In those cases, the rearrangements involved either the BRAF or RAF1 gene, but not the SLC4A3 or ESRP1 loci, "suggesting fusions involving multiple 5' partners, similar to ETV1 fusions in prostate cancer," the authors wrote. However, they were unable to identify the specific genes due to a lack of available tissue. All of the cases were also negative for ETS rearrangements, suggesting that the mutations in the RAF kinase pathway represent a rare subtype of prostate cancer.
Clinical information was available for all of the cases and Chinnaiyan noted that the patients with mutations in the RAF kinase pathway tended to exhibit more aggressive and advanced forms of prostate cancer.
Additionally, the team analyzed the same pathway in gastric, breast, endometrial, melanoma, and liver cancers using florescence in situ hybridization probes. They found mutations in the pathway in about 1 percent to 2 percent of the gastric cancer and melanoma tumors.
Despite the fact that this subtype of prostate cancer seems to be rare, occurring in just 1 to 2 percent of cases, Chinnaiyan said that because the same pathway is implicated in other cancers — like gastric and melanoma — it was still a clinically relevant finding, and a worthy drug target. In fact, he said, there are currently drugs in development that target the RAF kinase pathway.
In the study, the researchers performed preliminary testing of these drugs in mouse cells. By creating the implicated gene fusions in a mouse cell line, they demonstrated that it was sensitive to the RAF kinase inhibitor Nexavar (sorafenib), a drug marketed by Bayer that has been approved for primary kidney cancer and advanced primary liver cancer.
Chinnaiyan added that eventually, cancer will likely be subtyped based on its mutational pathway rather than the organ in which it originally formed, and treatment will be targeted to that pathway.
"I think we're moving towards this era of personalized medicine where you're going to have these patients of rare subtypes, and the mutation might be rare, but the [pathway] is still important" across a range of cancers, Chinnaiyan said.
David Smith, professor of laboratory medicine and pathology at the Mayo Clinic who has used transcriptome sequencing to identify mutations in oral cancers, agreed that the fact that the mutations Chinnaiyan's group identified are rare does not mean that they do not have significance.
"The abundance of [the gene fusions] in any one population of cancer is quite small, but they cross a whole series of different cancers. So, it's a low percentage in this cancer, but lo and behold, it's mutated in three or four other cancers," he said. "It's an individualized medicine approach to cancer."
Smith was less convinced, though, that the method the researchers used would have clinical significance in the future. He said that while transcriptome sequencing may be more economical now, whole-genome sequencing costs are quickly falling and will likely become the standard approach. Also, he said, transcriptome sequencing could miss important rearrangements.
However, Chinnaiyan said that transcriptome sequencing's current lower cost and easier interpretation made it a more efficient approach for his study.
"When you sequence the DNA, it's hard to know what's important and what's not important." Transcriptome sequencing "focuses in on things that are potentially functional. You see the RNA being made. If you're doing whole-genome sequencing, you have rearrangements that won't necessarily create an mRNA," he said. "Our approach focuses in on mutations that could be driving."
Chinnaiyan said his group will continue to look for additional, targetable mutational pathways in prostate cancer.
"We're using the technology to scour additional prostate cancer samples to look for driving mutations… We're trying to fill out the molecular subtyping scheme for prostate cancer."
Chinnaiyan said that his team would also be using transcriptome sequencing, as well as exome sequencing, to sequence an additional 500 to 1,000 tumors from other carcinomas, including prostate, breast, and gastric cancers. The combination of transcriptome and exome sequencing yields a "pretty comprehensive picture outside of sequencing the entire genome," he said.
Additionally, he said that they will be testing existing RAF inhibitors on prostate cancer samples.