CHICAGO – Polish-American startup MNM Bioscience began as a bioinformatics company that curates data and then builds artificial intelligence-based algorithms, but it is now looking to branch out into oncology drug target discovery.
"We've realized that if you do very detailed analysis of thousands of genomes, suddenly you see cohorts of patients which are not addressed by any existing therapies," CEO Pawel Zawadzki said.
About six months ago, management had what he called a "revelation" that the firm could work in this new realm because MNM discovered that about 30 percent of patients with triple-negative breast cancer had tumors with a unique genomic signature.
"We started to dig deeper, and then we realized we think we know the origin of this particular signature. There is a particular defect in this cohort," Zawadzki said. While this data has not been published yet, the firm is now working with its laboratory partners to validate its preliminary findings on cell lines, a form of reverse translational research.
"I think in 12 or 18 months we are going to a full-blown drug company with our own proprietary strategy to generate targets, validate targets, and validate and define cohorts where you use [a] particular drug," Zawadzki said.
Clinical validation is underway. MNM said last week that it has received an $800,000 Eurostars grant to develop a software tool for the stratification of patients with ovarian cancer. A few days later, the company announced that it had filed a provisional patent application for its algorithm for predicting therapeutic response of cancerous tumors to compounds including PARP and CDK4/6 inhibitors.
Under the two-year grant, the firm will work with partner Macrogen Europe to develop a product called PARPiNDx, which will apply artificial intelligence to whole-genome sequencing data to find patients likely to respond to PARP inhibitors, as well as those with intrinsic resistance to the drug family. The Eurostars grant will also help MNM clinically validate the algorithm developed for PARP inhibitors.
The grant is limited to ovarian cancer, but Zawadzki said that the technology is applicable to any kind of cancer.
A second algorithm that MNM has developed is to assess the effectiveness of palbociclib and other CDK4/6 inhibitors, an area that nobody has found any predictive biomarkers for yet, according to Zawadzki. This algorithm, which Zawadzki called the firm's biggest success to date, has been licensed to a major US-based molecular lab firm that MNM is contractually prohibited from naming.
Now, MNM is also working with its lab partners — companies Zawadzki declined to name — to verify insights generated from whole-genome sequencing, sometimes comparing its findings to panel-based data that labs have.
Zawadzki said that MNM decided to focus on whole genomes because the company wants to derive knowledge from noncoding regions of the genome. "The insights come from the parts where people have been neglecting for years," he said. "That was actually a driving force to start a company in the first place."
The firm grew out of Adam Mickiewicz University in Poznan, Poland, in 2018, where Zawadzki had established a research group focused on DNA repair a year earlier. Zawadzki previously spent seven years as a postdoctoral researcher in DNA technology at the University of Oxford in the UK.
For its first two years, MNM exclusively dedicated its time to curating data. The firm now has a dataset of more than 5,000 whole tumor genomes. About half of the genomes come from open databases, while the rest are proprietary, generated by partners, according to Zawadzki, since MNM does not have its own molecular laboratory.
The company recently relocated its headquarters to Cambridge, Massachusetts. Zawadzki called the Boston area the "epicenter for biotech," one that was too appealing to ignore. He also said that most of the company's current clients are in the US and Canada.
MNM will keep its engineering staff in Poland while running its business development and scientific operations from the US.
The company name stands for Mutations No More. "When we started, the idea was that tumors are full of mutations," Zawadzki explained. "When we realized that if you understand these mutations, actually, you can turn it into a weapon against the tumors, so we thought that was an appropriate name for a company that is trying to understand the patterns of mutations and turn [that knowledge] into diagnostic tools and new therapies."
Zawadzki said that Cofounder and Chief Operating Officer Katarzyna Zawadzka developed similar algorithms at the University of Cambridge. The company's third cofounder, Dominik Glodzik, a Harvard University biomedical informatician, built an AI-based classifier using whole-genome sequencing data, and detailed his work in a 2017 Nature Medicine paper.
MNM focuses on characterizing cohorts of patients. Though the technology can support research, Zawadzki said that the business case for MNM is in the diagnostic realm.
While insights derived from MNM analysis can inform drug development, Zawadzki said that "the most immediate way to use it is to actually stratify patients for a given therapy."
MNM has begun to look beyond genomics by incorporating transcriptomics and medical imaging into its analysis. Eventually, the firm wants to include epigenomic and methylomic data, according to Zawadzki.
Zawadzki said that the company is working to get its first two algorithms into clinical practice following validation. Preliminary validation of its algorithm via high-throughput screening should be done by midsummer. There are about five or six more algorithms in MNM's development pipeline now, according to the CEO.
For its drug program, MNM has no designs on starting its own drug discovery operation, but instead will partner with medium-sized pharmaceutical companies and contract research organizations.
"We do not have experience in lead optimization," Zawadzki said. "We have experience with cell lines and preclinical validation of a target, and this is what we intend to do partially in-house, partially via CRO." Lead optimization and actual clinical research will happen at partner labs.
Aside from the Eurostars grant, MNM is essentially bootstrapping based on revenues. Zawadzki said that ramping up the drug program will be expensive. Expect MNM to look at venture capital as target development gets off the ground, but not before the firm has enough high-quality data to attract top investors.
Zawadzki said that the company's target is synthetic lethality with BRCA 1 and 2. "If we manage to develop an inhibitor, we think we're going to be in a comfortable position to actually run fundraising to support this program," he said.
As is typical of emerging bioinformatics companies, MNM is approximately doubling its workforce every year or so. Zawadzki said that the company current employs nearly 40 people and is heavily recruiting scientific and business development people in a hypercompetitive labor market.