This blog contains information originally published as part of the Business of Biotech podcast
Pillar, M. (Host). (2021, May 3). Digital Transformation In Cell Therapy with Organicell Drs. Mari Mitrani & Michael Bellio [Audio podcast episode]. The Business of Biotech. BioProcess Online.
Biopharmaceutical innovators are adopting machine learning, bioinformatics, and automation to speed discovery and development and create production process efficiencies.
On a recent episode of The Business of Biotech, we sat down with Organicell’s Dr. Mari Mitrani, co-founder and chief science officer, and Dr. Michael Bellio, laboratory director, to discuss how the clinical-stage regenerative medicine company has prioritized machine learning since its inception, and what they and other biopharmas are aiming to do with it in years to come.
An Early Foray Into Data-Driven Discovery
Machine learning was in its nascency when Organicell was founded in 2008, but the company was an early adopter. “We were very invested in autologous stem cells and transitioned from a regular puncture to extract those stem cells from the bone marrow to an apheresis device,” says Dr. Mitrani. “We were doing sternum punctures, which made the process less invasive, and even with that protocol, machine learning enabled us to jump into a more data-driven and precise collection of stem cells.”
That was early days for the technology, but machine learning applications—such as imaging analysis for non-invasive measurement of morphological cell dynamics—were just the kind of commitment to innovation that attracted Dr. Bellio to the company. “It was this goal of developing these new therapeutics with exosomes, in facilities where we could complete all of our clinical trials and basic science research in one place with innovative digital technology,” that appealed to him, he says.
Machine Learning, Bioinformatics Speed Time To Clinic
Bioinformatics is another key tool in Organicell’s tech arsenal, aiding the company in its recent work sequencing micro-RNA and protein cargo within perinatal exosomes. “That was really the most important part for us, to be able to characterize micro RNAs from high expression to low expression. We use bioinformatics to help guide the writing of a lot of our clinical applications to the FDA and demonstrate the product and why it’s effective, and it all kind of trickles down from there,” says Dr. Bellio. “It's almost a requirement at this point, when you're developing these new therapies. You need to use bioinformatics to help guide your initial thought process about that mechanism of action that the FDA wants to know.”
That “trickle down” effect extends all the way to the fill/finish line, where the efficiencies gained by machine learning and bioinformatics move the company more quickly to the tech-enabled automation it’s deployed in its aseptic filling operations. “It's never too late to start thinking about automation,” says Dr Mitrani. There are many avenues for it in cell and gene therapy and significant benefit can come from implementation, such as reduced error, improved consistency, automated product handling, and management of other manufacturing functions. “One of the fastest, easiest deployments of automation is implementation of a system to fill, cap, and crimp vials.”
Digitization Drives Future Machine Learning Advances
Dr. Mitrani points out that the digitization of processes and paperwork — from donor qualification to batch records and manufacturing documentation — represents an ongoing opportunity.
Organicell’s digitization strategy relies on an array of data collection points deployed at many stations across its workflow, “It’s not just one device, or two,” Dr. Mitrani says. “We’re looking for the biomarkers outside the membrane, counting the proteins, measuring the concentration of hyaluronic acid, and even tracking endotoxins. We run a procedure throughout the day to ensure our final products don’t have any endotoxin, without relying on private testing.”
Drs. Bellio and Mitrani stress that this increased adoption of bioinformatics, sequencing, and proteomics will help meet the developmental needs of new exosomic therapies. Unlike in more established biopharmaceutical development pathways where developers track a molecule from absorption to metabolism and excretion, exosome-based therapies will increasingly rely on new strategies that include bioinformatics and AI to deal with added complexity. “The entire field is moving this way,” Dr. Mitrani says.
Organicell certainly is moving that way, and it’s reaping the rewards. In January, the company announced it had received FDA approval to pursue clinical trials for Zofin, an acellular, biologic therapeutic derived from perinatal sources and used to treat chronic obstructive pulmonary disease (COPD). In July, the FDA approved its IND application for Zofin in the treatment of COVID “long haulers” — people who suffer from symptoms related to COVID-19 well after being infected with the virus.
Though COVID-19 vaccine efforts are ongoing, Dr. Mitrani says the population of long haulers has yet to be addressed. The company’s flurry of recent successes, say Drs. Mitrani and Bellio, are directly related to its early and ongoing adoption of machine learning, bioinformatics, and automation technologies.
Download the Business of Biotech podcast series to hear from guests who turned biotherapy ideas into clinical realities.