Cell and gene therapy (CGT) manufacturing is rapidly transitioning from scientific curiosity to clinical reality. But the manufacturing complexity of CGTs far outpaces traditional biologics production, presenting a multifaceted challenge that is part scientific, part technological. “The manufacturing process for cell therapies and gene therapies is infinitely more complex than it is for let’s say, monoclonal antibodies or recombinant proteins,” said Betty Woo, vice president and general manager of cell, gene, & advanced therapies at Thermo Fisher Scientific.

Despite the complexity, much of the current cell therapy manufacturing process relies on manual procedures. “Manual intervention introduces a level of human variability,” Woo said. “This not only leads to inconsistencies, but also increases the risk of human error.”

While manufacturing complexity can pose hurdles, the field as a whole shows tremendous growth potential. In 2020, McKinsey emphasized the growth in the field, noting that while CGT accounted for 1% of launched products in major markets, it makes up 12% of the industry’s clinical pipeline and at least 16% of the preclinical pipeline. These numbers signal a field gaining rapid momentum. According to In Vivo, there are more than 2,000 CGT clinical trials underway. The international cell therapy market was worth less than $8 billion in 2020, but could surpass $60 billion by 2030, according to projections from Grand View Research.

From monogenic to polygenic: The expanding scope of CGTs

“Historically, gene therapy and cell therapy have focused on monogenic diseases,” said Betty Woo, vice president and general manager of cell, gene, and advanced therapies at Thermo Fisher Scientific.“Correcting mutational errors in the gene sequence has the potential to cure disease. We’re seeing evidence of that with some of the early commercialized therapies,” Woo said.

Examples of gene therapy targeting monogenic diseases include an AAV DNA technology for spinal muscular atrophy (SMA) and β-thalassemia, a rare blood disorder. Onasemnogene abeparvovec-xioi, can treat SMA by replacing the mutated SMN1 gene in motor neurons. The mutated SMN1 gene causes SMA. In addition, U.S. and E.U. regulators have also approved betibeglogene autotemcel for β-thalassemia, which is caused by HBB gene mutations.

More commonly, diseases involve multiple genes (polygenic diseases) with mutations dispersed throughout a gene or genome. Examples include common conditions like heart disease, diabetes and many forms of cancer.

Unleashing the power of AI and ML in CGT therapies

The challenges developing therapies for mono- and polygenic diseases demands sophisticated approaches. Machine learning is helping tame this complexity in the domain of basic research and beyond. “From an R&D perspective, AI and ML could be applied in predicting the most stable nuclease binding sites for gene editing that would minimize off-target effects. AI could also be used to identify optimum CAR antigens and binding sites, again, yielding cell therapies with higher therapeutic indices.”

“Ideally, automation would serve to control and monitor the manufacturing process, while AI and ML could be applied to adapt the process to produce safe and effective therapies with a well-defined quality standard. Automation is even more important with CGTs than for more traditional biologics because many steps in manufacturing of these new modalities are manual, and therefore subject to human variability and even error. To add, in the manufacturing of autologous cell therapies, the input material from patients is highly variable in quantity and quality. Inherent biological variability along with the health of the patient’s cells may require adaptation to the manufacturing process.” The ideal scenario, she suggests, is to achieve a “consistent, predictable process every single time.”

Beyond being a mere manufacturing tool, automation, when combined with AI and machine learning, provides vital process control. This fusion allows for dynamic adjustment based on real-time process monitoring. Woo described this seamless integration as the “Holy Grail” of cell and gene therapy manufacturing.

ML can also predict binding site stability, optimize CAR-T cell manufacturing by analyzing the structure of CAR molecules. The technique can also help identify tumor targets for immunotherapies like mRNA vaccines or CAR-T therapies. ML can also help identify genome editing sites to maximize on-target effects and minimize off-target effects for CRISPR-based viral therapies.

Manufacturers are also using ML to optimize cell and gene therapy production across diverse instruments. By improving processes like cell culture, purification and quality control, ML enhances efficiency, reproducibility and scale.

Riding the wave of progress in cell and gene therapy manufacturing

In addition to the control and monitoring functions, AI and ML-based software solutions can also help to ensure compliance with regulatory standards such as 21 CFR Part 11.

“So it is definitely a crawl, walk run type of progression cell and gene therapy, and we’re getting ever better in directing these therapies to the right tumor sites with the most efficacious drugs,” Woo concluded.

“We are moving so quickly, technology is moving so quickly in the field, that basically if you don’t innovate, you die,” she said.

---

Filed Under: Cell & gene therapy, Women in Pharma and Biotech
Tagged With: AI in Pharma, automation in manufacturing, biopharmaceutical manufacturing, cell and gene therapy, gene editing, monogenic diseases, polygenic diseases
 

---