NEW YORK (GenomeWeb News) – A search for low-frequency cancer driver genes suggests truncating mutations to a transcription factor gene called CUX1 may be the impetus for cancer formation in a small but significant subset of cases across several cancer types.
A Wellcome Trust Sanger Institute-led team, which included members of the International Cancer Genome Consortium's chronic myeloid disorders working group, sorted through exome or genome sequence data for more than 7,650 tumors profiled by the ICGC, the Cancer Genome Atlas, or other groups. The search focused on particularly prone to nonsense mutations — alterations that introduce stop codons into a gene and truncate its protein product.
As they reported online yesterday in Nature Genetics, the researchers picked up numerous known driver genes using this approach. But they found new candidate drivers as well, including the cut-like homeobox 1 transcription factor-coding gene CUX1. Within the various cancer types tested, between 1 percent and 5 percent of tumors contained inactivating CUX1 mutations, which appeared to have a tumor suppressor role in the mouse and fruit fly experiments.
In particular, inactivation of CUX1 appears to spur on the activity of the PI3-kinase pathway, the group found, causing amped up signaling that seemed to be susceptible to the types of PI3K inhibitors already being developed and/or tested.
As such, the work is "a prime example of how understanding the genetic code of cancers can drive the search for targeted cancer therapies that work more effectively and efficiently," according to senior author David Adams, with the Wellcome Trust Sanger Institute.
"This could improve the lives of thousands of people suffering from cancer," Adams said in a statement.
Despite the progress that's been made characterizing cancer genomes and identifying genes frequently mutated within and across cancer types, the researchers reasoned that extremely large sample sets would be needed to narrow in on undetected driver genes acting at relatively low frequency.
With that in mind, they assessed available sequence data across 28 tumor types, bringing together exome sequence data for 7,299 tumors and whole-genome sequences for another 352 tumors.
Using computational methods to track down genes containing nonsense mutations in a significant subset of the tumors, the team found more than four-dozen potential cancer drivers.
Some are already notorious for their role in cancer, the researchers noted, including TP53, PTEN, and CDKN2A. Other candidate drivers were new. For example, they saw hints that genes such as LARP4B, AMOT, and CUX1 may also act as tumor suppressor genes, driving cancer formation when sufficiently mutated.
To explore that possibility further, the team homed in on one of these genes: the chromosome 7 transcription factor gene CUX1.
Across the tumor genomes considered, the CUX1 gene contained nonsense mutations more than three times as often in tumors than predicted by typical variation patterns, the researchers noted.
Other types of non-synonymous alterations in the gene were recurrent as well, including frameshift alterations and missense mutations predicted to change the amino acid sequence of the CUX1 product.
All told, the team saw nonsense and frameshift glitches expected to inactivate the resulting transcription factor protein between 1 and 5 percent of the time, depending on the tumor type. Such changes were most common in endometrial cancers, which appeared apt to contain missense and synonymous CUX1 mutations as well.
A potential tumor suppressor role for the gene required some scrutiny, though, since some studies linking CUX1 to cancer in the past pointed to a potential cancer causing, or oncogenic, role.
Such findings may partly reflect the multiple transcript isoforms and functions possible for CUX1, researchers explained. For instance, they noted that a short CUX1 isoform has been implicated in some breast cancers, though prior research also revealed a dramatic dip in CUX1 levels in some acute myeloid leukemias, consistent with a tumor suppressor function for other versions of the transcription factor.
Through a deeper dive into existing data for more than 2,500 individuals with AML and other myeloid cancers, the team found that CUX1 mutations are quite recurrent in that group of cancers, often corresponding with bleak survival outcomes.
"CUX1 defects are particularly common in myeloid blood cancers, either through mutation or acquired loss of chromosome 7q," first author Chi Wong, a hematologist and researcher affiliated with the Sanger Institute and the University of Cambridge, said in a statement. "As these patients have a dismal prognosis currently, novel targeted therapies are urgently needed."
One potential avenue for treating the CUX1-containing cancers came out of the group's functional experiments.
In addition to finding further evidence for a tumor suppressor role in mouse and fruit fly models, the researchers used RNA interference and gene expression profiling to show that human cell lines with lower-than-usual CUX1 levels are subject to rampant signaling through the PI3K pathway.
A closer look at such cell lines suggests that CUX1 normally keeps such signaling in check by boosting the transcription of PIK3IP1, a gene involved in inhibiting excessive activity by members of that pathway such as PI3K, AKT, and mTOR.
Consistent with that model, results from their preliminary experiments suggest cells with CUX1 mutation or depletion are especially vulnerable to drugs that inhibit PI3K pathway components.
"[T]he susceptibility of CUX1-mutant cells to PI3K-AKT-mTOR pathway inhibition provides a potential therapeutic avenue," Adams, Wong, and their co-authors wrote, though they cautioned that CUX1's broader role in suppressing tumor formation might involve other pathways as well.