SAN FRANCISCO (GenomeWeb) – Startup Orbit Genomics plans to develop next-generation sequencing-based assays that analyze microsatellite sequences to diagnose cancer and other diseases.
The Boulder, Colorado-based firm, which was founded last year by CSO Harold Garner, CEO Dede Willis, and others, leverages research done by Garner on microsatellites as well as publicly available genomic databases, like the Cancer Genome Atlas to discover and validate microsatellite markers of disease.
Previously, Garner was co-founder and CSO of Genomeon, which also aimed to develop microsatellite-based diagnostics. In 2015, Genomeon and Cloud Pharmaceuticals agreed to partner to use Genomeon's microsatellite technology to find novel drug targets. However, Garner said that Genomeon has since been shuttered. Orbit, which now plans to develop that technology, was founded with a new management team and new business plan, Garner said. He said that he hopes to renew the agreement with Cloud Pharmaceuticals and form similar partnerships in the future.
Orbit's first test will be to diagnose lung cancer. Garner said it will be positioned as an early diagnostic for at-risk individuals who do not yet have symptoms. In addition, he said, Orbit is working on a pan-cancer test as well as companion diagnostic tests for cardiology drugs, and an autism test.
Garner said that the company is looking to conduct a Series A financing round this year and plans to launch its first test in a little more than two years. Orbit currently has just under 10 employees and is working with consulting firm Ventac Partners.
Recently, Garner led a study that was published in the journal Oncogene in which the researchers validated a set of microsatellite markers that were predictive of lung cancer.
"The reason why we focused on microsatellites is that they are more sensitive to the environment and so easily mutated," Garner said. "They are a sensitive readout of the overall stability of a cell or organism."
Nonetheless, sequencing microsatellites has been technologically challenging because they are repetitive and difficult to map to the correct location in the genome. Most variant-calling algorithms "do not call microsatellite sequences properly 80 percent of the time," Garner said. "So, discovery of conditions related to microsatellites are hard because of the noise and errors. "
Garner's team has developed a number of techniques to overcome these technological hurdles. The main problem with standard NGS approaches, he said, is that the algorithms perform one assembly for all the sequence data. But, he said, he has developed algorithms that perform a few million tiny assemblies, which enables more accurate mapping of the microsatellites. In addition, when doing the actual sequencing, the team will only include microsatellites for which the entire region and flanking region is sequenced, which helps improve the accuracy.
One bioinformatics approach, dubbed ReviSTER for Revise Simple Tandem repeat Error Reads, was described in Bioinformatics in 2013. Garner said that ReviSTER was developed primarily as an academic technique, and said that Orbit has improved and modified the technique significantly. In addition, the Orbit team has developed other proprietary bioinformatic tools that it uses in its assay development work.
In the most recent Oncogene study, the researchers first downloaded Cancer Genome Atlas data from 488 individuals with lung cancer as well as data from 390 control samples from the 1,000 Genomes Project.
For discovery purposes, the team first looked very broadly, identifying microsatellites that were enriched in the cancer samples, and found 119 different loci.
Next, the team designed a targeted sequencing panel for the 119 microsatellite sequences that were enriched in the lung cancer samples. But, they also wanted to see whether microsatellites they had discovered from previous studies of breast and ovarian cancer, lower-grade glioma, glioblastoma, melanoma, and medulloblastoma would also be relevant for lung cancer, so they included 144 additional microsatellite targets in the panel, plus 84 control microsatellites. The panel targeted both the microsatellite region and the unique flanking sequence for all 347 loci.
They then sequenced this targeted set on a different cohort of 30 lung cancer samples and 89 controls to validate the loci and further refine the markers. Sequencing identified 13 loci from the set of 119 lung-cancer specific microsatellites and eight from the set of 144 from other cancers that had different genotypes in the cancer versus control samples.
The 13 lung cancer-specific markers identified lung cancer samples from controls with a sensitivity of 90 percent and specificity of 94 percent. In addition, the researchers calculated that samples with eight or more of the lung cancer genotypes had an increased risk of non-small cell lung cancer. When all 21 markers were analyzed, sensitivity and specificity increased to 93 percent and 97 percent, respectively and the team determined that a threshold of 12 or more markers could classify someone as high risk.
Looking more closely at the 13 lung cancer-specific microsatellites, the researchers found that all 13 were in intronic regions of genes, including some that have previously been linked to lung cancer, such as REL and ARID1B, both of which are also potentially actionable.
Garner said the next step is to secure between $10 million and $20 million in funding to commercialize assays and set up a CLIA-certified laboratory. Aside from lung cancer, he said the company aims to develop a pan-cancer test; a test that distinguishes between high and low-risk brain tumors; and companion diagnostics for some cardiovascular diseases. The firm is also validating a test for autism.
For the lung cancer test, he said the goal is to position it as a way to predict risk of developing cancer. "We want to use it on a population that already has some risk factors — a family history or those who smoke," he said. The team is also continuing to improve on the sensitivity and specificity, he said, and "ultimately, we see this as something that could be a screening tool."
Aside from risk prediction, tests would also include markers that could guide treatment, Garner said. For instance, in a previous study in Oncotarget, researchers described microsatellite markers that could help determine whether a brain tumor was likely glioblastoma or a lower-grade glioma, information that helps determine whether a patient should be aggressively treated or monitored.
Orbit is not the only company to see the value in microsatellites. Promega offers a research-use only assay that measures microsatellite instability and plans to seek US Food and Drug Administration approval for it to help oncologists make treatment decisions for colorectal cancer patients. Its assay relies on fluorescent multiplex PCR to look at seven markers that are indicative of the microsatellite instability-high phenotype, which can help determine whether a tumor is deficient in mismatch repair. The FDA recently approved the Merck drug Keytruda to treat cancer patients who have the MSI-H phenotype or are mismatch repair deficient.
Orbit is pursuing a slightly different path, however, by analyzing specific variations in microsatellites. Garner added that the next steps for the company after securing funding would be to finish validation of the assay and conduct a clinical trial.