NEW YORK (GenomeWeb News) – By sequencing the protein-coding repertoire or expressed RNA transcripts from more than a dozen rare ovarian tumors, a Canadian team has uncovered recurrent mutations in the microRNA-processing gene DICER1 — changes also found in a subset of other tumors involving primitive tissue types.
As they reported online today in the New England Journal of Medicine, researchers from the British Columbia Cancer Agency sequenced the whole exomes and/or transcriptomes of 14 non-epithelial ovarian cancers, identifying missense mutations affecting the same residue of DICER1in four of these samples. Mutations in and around the same mutation hotspot in the genes turned up in a nearly a third of the 102 non-epithelial ovarian tumors subsequently screened by targeted sequencing.
The team saw similar changes in DICER1 in a subset of other cancers, too, including a testicular germ cell tumor, two embryonal rhabdomyosarcomas, and a single epithelial ovarian cancer sample.
"It may become an excellent example of how, for post-surgical treatment, we have to start thinking about tumors and classifying them by what they are rather than where they occur, because we may have a new group of tumors emerging, which are tumors of Dicer malfunction," co-corresponding author David Huntsman, a genome sciences researcher with the British Columbia Cancer Agency and director of its ovarian cancer research program, told GenomeWeb Daily News.
The findings hint that DICER mutations may occur early on, leading to developmental field quirks that can eventually lead to cancer at various locations throughout the body, he added, though more research is needed to test that hypothesis.
Non-epithelial forms of ovarian cancer include cancers that originate from germ cells, as well as so-called sex cord-stromal tumors involving tissues formed during early embryonic development.
To look at these relatively rare forms of ovarian cancer in more detail, researchers did transcriptome or exome sequencing on 14 non-epithelial ovarian cancer samples collected by members of the Ovarian Cancer Research Program tissue bank in Vancouver and by the Children's Oncology Group at the Nationwide Children's Hospital in Columbus, Ohio.
These tumors included two Sertoli-Leydig tumors, four juvenile granulose-cell tumors, and eight yolk-sac type primary germ-cell tumors.
"Rarer tumors represent a discrete biologic space and, therefore, smaller studies are more likely to yield knowledge, as opposed to studies of more common cancers, which are more heterogeneous," Huntsman explained.
When they sorted through this sequence data, investigators found a few mutations scattered at different sites in the DICER1 gene, including some truncating mutations.
Intriguingly, though, four of the samples shared non-synonymous missense mutations at the same spot in DICER1: the sequence coding for an aspartic acid residue at position 1709. This amino acid, which falls in the RNase IIIb domain, binds magnesium or manganese metals and contributes miRNA processing by facilitating interactions with pre-miRNAs, Huntsman explained.
In contrast, researchers did not see mutations in sequences coding for another Dicer domain, RNase IIIa, which chops up pre-miRNA to make partnering miRNA strands during the formation of short miRNA duplexes.
"What we expected was that each of these tumors would have its own mutation type," Huntsman said, "but what we found was a smattering of the same mutation across these different tumor types. The mutation was instantly interesting because it was the first recurrent missense mutation in DICER1."
When they sequenced the RNase IIIb domain of DICER1 in 101 more non-epithelial ovarian cancers and in 39 testicular cancer samples, researchers found hotspot mutations in 29 percent of the ovarian tumors tested, including around 60 percent of the tumors classified as Sertoli-Leydig cell tumors.
Mutations in the same part of the DICER1 domain were not limited to ovarian cancer, study authors explained, but were also detected in one of the 14 non-seminomatous testicular germ-cell tumors tested.
In researchers' follow-up experiments, they also found mutations around the same site in DICER1's RNase IIIb domain in two of five embryonal rhabdomyosarcomas tested and in one of 266 epithelial ovarian and endometrial carcinomas tested.
Of the 33 tumors that were found to have mutations within the RNase IIIb hotspot, researchers reported, 17 specifically affected residue 1709. On the other hand, they did not detect DICER1 hotspot mutations in matched normal tissues, when available, suggesting such changes arise somatically rather than being inherited.
In seven samples containing both RNase IIIb mutations and other somatic or germline mutations in DICER1, the team's analyses suggest that somatic hotspot mutations fell on a different allele than other DICER1 alterations.
Based on their findings, for example, the researchers believe individuals who are born with truncating germline mutations in one allele of DICER1 will have somatic hotspot mutations that land in the other allele of the gene — a pattern that Huntsman said might help explain why such germline mutations have been found in previous studies in the absence of obvious changes to the gene's other allele.
"The reason that you can't find loss of heterozygosity in individuals born with a truncating mutation in this gene is that the second hit is always one of these missense mutations," he said.
From their initial follow-up studies, researchers believe the protein that results from versions of DICER1 harboring hotspot mutations are still somewhat functional, though the repertoire of miRNAs they handle may shift.
"We presume that the differences which are pathogenic are in the discrete miRNAs or families of miRNAs as opposed to the total community of miRNAs," Huntsman said.
In the future, the researchers intend to look more closely at the functional impacts of these hotspot mutations, including studies focused on finding out which pre-miRNAs are inappropriately processed in the presence of these changes. Ultimately, the group hopes such work will reveal downstream targets with ties to tumor formation.
"If we can discover what those [pre-miRNAs] are, we will either have a direct or indirect [treatment] target at that point," Huntsman explained, noting that such a target "could potentially cut across several different tumor types which no one ever would have linked together in any clinical trial."
The team is also planning to look at additional ovarian and testicular germ cell tumors as well as whole-genome sequencing studies of the rare tumor types.