SAN FRANCISCO (GenomeWeb) – Building on previous work to develop a next-generation sequencing assay to assess microsatellite instability, researchers at the University of Washington have developed an approach that harnesses single-molecule molecular inversion probes to target loci that can inform microsatellite instability status across multiple cancer types.
The researchers described the new assay in Clinical Chemistry this week, validating it in colorectal, prostate, and endometrial cancers. Stephen Salipante, senior author of the study, said that the goal is to validate it across other cancer types so that it can be run as a primary diagnostic test for microsatellite instability in cancer patients.
Microsatellite instability has become an important biomarker for cancer diagnostics over the last several years with research showing that tumors classified as MSI-high are more likely to respond to certain immunotherapies. Initially, the biomarker was validated only in colorectal cancers, but more recently, researchers have demonstrated that other tumor types exhibit the phenotype. Last year, the US Food and Drug Administration approved Merck's Keytruda (pembrolizumab), an anti-PD-1 immunotherapy, as a tissue-agnostic drug for MSI-H tumors in advanced cancer patients who have failed other drugs.
Laboratories are now looking to develop tests that can accurately classify tumors as MSI-H. The current gold standard is a PCR test that was validated for colorectal cancers and typically analyzes just five loci. Those five loci are robust for determining MSI status in colorectal cancers, but they "have significant losses outside of that," Salipante said.
Salipante's lab already offers a targeted NGS-based MSI clinical test called MSIPlus. The UW team developed MSIPlus originally for colorectal cancer, and the assay evaluates not only MSI status, but also mutational hotspots in the cancer genes KRAS, NRAS, and BRAF.
Salipante said the new MSI assay has two major differences. First, the assay looks at significantly more microsatellite loci — around 110 sites, up from 10 to 15 sites. Second, the team switched from a PCR approach for targeting the loci to molecular inversion probes.
UW researchers including Salipante and led by Jay Shendure initially described the single-molecule molecular inversion probe (smMIP) capture in Genome Research. The strategy enables multiplexed targeted sequencing and makes use of molecular barcodes to do error correction, which "dramatically increases the sensitivity of this approach in heterogeneous material," Salipante said.
For the recent study, Salipante's team chose the set of microsatellite loci based on a previous study where they evaluated data from the Cancer Genome Atlas to identify loci that were relevant across cancer types.
The team first tested the smMIP capture strategy on synthetic DNA mixtures. They combined DNA from a known MSI-H colorectal cancer cell line with a normal cell line at various concentrations to simulate tumor heterogeneity. In a mixture containing 1 percent of the MSI-H cell line they were able to distinguish the MSI-H loci from the control.
Next, the team evaluated the assay in known samples, testing 33 MSI-H and 35 stable colorectal cancer samples, 24 MSI-H and 19 stable endometrial cancers, and 11 MSI-H and 22 stable prostate cancer samples. They compared the smMIP assay with the gold standard PCR test for MSI and also evaluated the samples using a more comprehensive NGS strategy.
In total, of the 144 specimens tested, only one sample was falsely diagnosed by the smMIP strategy, an endometrial cancer sample. The assay had 100 percent sensitivity and specificity for all colorectal and prostate tumors. Overall, the assay had 98.5 percent sensitivity and 100 percent specificity.
Comparing the smMIP assay with the PCR test, the researchers found that the smMIP assay was more accurate. Two samples could not be analyzed via PCR because they failed to amplify. The PCR assay correctly classified all 32 colorectal samples as MSI-H, nine out of 11 prostate cancer samples, and 18 out of 24 endometrial samples. Its overall sensitivity was 88 percent.
The PCR assay was developed specifically for colorectal cancer, and has "significant performance losses" for other cancer types Salipante said, due to the fact that it only evaluates a handful of loci that are optimized for colorectal cancer.
Salipante said that the next step is to continue to validate the assay across other cancer types. He said that the goal is to develop this into a clinical test, but declined to provide a timeframe for that.
The smMIP test has a performance advantage over PCR, but it also has an advantage over other NGS techniques, including UW's MSIPlus test, in that it is low-cost, Salipante said. The researchers estimated that the reagents cost just $80 per sample for smMIP.
Although a clinical test would be more expensive, since it would include labor costs, Salipante said it would still be less expensive than other NGS approaches. Currently, most available NGS approaches for evaluating MSI status are built into large gene panels that also evaluate other molecular features. For instance, Caris Life Sciences' and Foundation Medicine's panels each include more than 300 genes.
Nonetheless, Salipante does not anticipate that the smMIP MSI test would compete with these types of tumor mutational profiling tests, since those larger panels provide other information.
He added that the assay could potentially replace UW's MSIPlus test, but likely would not replace its larger gene panel offerings. While Salipante said the initial goal is to offer it clinically at UW, there is also the potential to commercialize the test more broadly, but the group does not yet have plans to do so.
"There's a lot of interest in MSI as an actionable phenotype for immunotherapy, but the only technology that many labs have access to is the five-marker PCR panel, which doesn't work well outside of colorectal cancer," he said. So, the goal with the smMIP assay is to provide a more accurate option that is quick and cost-effective.