This article was edited to clarify the types of molecules being investigated for therapeutic potential.
NEW YORK – Empress Therapeutics sees the next wave of small-molecule medicines reflected in the drug-like behavior of metabolites produced by the human microbiome.
The Watertown, Massachusetts-based firm is mining the human microbiome –– more than 170 million genes, by some estimates –– to identify genes encoding proteins that confer disease protective effects, focusing on immune disorders, metabolic disorders, and cancer.
"You've got around 20,000 genes in the human genome, but inside of every person on the planet, we've actually got 170 million," said Jason Park, CEO and cofounder of Empress. "These genes are part of your physiology, part of your health, [and] a huge number of those genes are dedicated to the synthesis of compounds."
Empress' general approach to drug discovery is to compare the metagenomic profiles of healthy individuals to those with disease to find molecules associated with healthy states. By applying machine-learning algorithms to stool and saliva samples from healthy and affected people, the firm has so far identified 15 potentially therapeutic molecules produced by human microbes and plans to file its first investigational new drug (IND) with the US Food and Drug Administration sometime in the next year.
In particular, the company relies on its Chemilogics platform, which uses natural language processing and other machine-learning methods to analyze metagenomic sequences encoding enzymes that produce small molecules with disease-protective associations. Promising sequences are examined in greater detail in engineered cells, and the firm resolves the structures of any molecules of interest through mass spectrometry and nuclear magnetic resonance (NMR) spectroscopy, allowing Empress insight into these molecules' precise targets and mechanisms of action. The goal, Park said, is to then improve upon those molecules' therapeutic capabilities.
Park said that Empress' genetics-first approach sets the company apart from others in the field.
"What we're doing is completely orthogonal to the way people used to think about small molecule drugs," Park said. "We ask the question: 'What looks like and acts like a drug inside the body, and then what does it interact with?'"
According to Park, "improving upon nature" and optimizing molecules that already play therapeutic roles within our bodies may streamline the drug discovery process. As molecules produced by our microbiomes are already more likely to be safe and well tolerated, this approach could, he argued, simplify much of the traditional preclinical toxicology work and improve the odds of clinical trial success.
"We have every reason to believe that we're going to have a higher probability of success in the clinic," Park said. "The beauty is because we start with what's already happening inside of humans, there's a lot of reason to believe in the translatability of our findings."
So far, Empress has identified hundreds of compounds, 15 of which are of particular interest to the firm. It is in the process of determining how to prioritize those compounds for further development, though Park added that approximately 50 percent of the molecules that the company has decided to test in vivo to date have shown potentially useful activity, which he calls an "unprecedented" success rate.
Park said that he hopes to file the company's first IND sometime next year for a drug candidate in the autoimmune space.
He added that the company's ability to generate so many leads with a staff of approximately 45 people also speaks to Empress' greater efficiency and, potentially, lower costs of drug discovery. "We are going from data to drug candidates in about two years," Park said, whereas some estimates put average drug discovery time frames between three and five years or even four and seven years.
The company has also been partnering with drug developers to advance its lead molecules. However, Park said that those partnerships are not yet ready to be disclosed.
Empress is operating in a small niche of the greater drug discovery field. Although the microbiome has attracted considerable interest as a source of potential diagnostic and prognostic biomarkers for a wide array of conditions and as a therapeutic target itself, relatively few companies are mining it for drug discovery purposes, particularly in Empress' genetics-first manner.
Biosortia Microbiomics, for instance, also mines microbiomes for therapeutic molecules, but the Dublin, Ohio-based company focuses on microbes found in the environment, particularly those found in bodies of water. Biosortia's microbiome prospecting platform is designed to collect microbes from large water volumes in situ while preserving their physical and chemical integrity, after which the company mines the samples using a combination of genomic, proteomic, and metabolomic approaches to identify natural drug-like small molecules that serve as starting points for preclinical drug discovery research.
Another actor in this space is Viome Life Sciences, which has recently branched out from its core microbiome-based wellness products business into diagnostics and therapeutics research. The company's ViOS platform applies machine learning-driven metatranscriptomic analyses to discover biomarkers and identify potential therapeutic targets for further research, as well as to construct predictive models of various disorders.
Empress was founded in 2020 by Flagship Pioneering, a life science venture capital firm, with an initial commitment of $50 million. While Park declined to discuss the company's cash flow at this time, he said that Empress is "well funded."
"This is a totally different way to think about discovering chemistry-based medicines," Park said.