Proceedings of a Workshop—in Brief

Emerging Technologies and Innovation in Manufacturing Regenerative Medicine Therapies

Proceedings of a Workshop—in Brief

In 2017 the National Academies of Sciences, Engineering, and Medicine’s Forum on Regenerative Medicine held a workshop on manufacturing of early-generation regenerative medicine products and measuring their quality (NASEM, 2017). This workshop raised issues associated with product critical quality attributes and emerging technologies capable of meeting the manufacturing needs and regulatory standards of these products. Since 2017 the Food and Drug Administration (FDA) has approved 32 cell and gene therapies, with more approvals anticipated,¹ said Claudia Zylberberg, founder and former chief executive officer of Akron Biotech. Since 2018 the number of regenerative medicine therapy product developers has increased from 900 to at least 2,700 globally, the number of gene and cell therapy clinical trials has increased from around 1,000 to 1,600, and the number of patients treated with chimeric antigen receptor T cell (CAR-T) therapies alone has increased from at least 180 to 20,000 (Mitra et al., 2023).² While manufacturing advances have accompanied some of this growth, the field is also starting to realize the potential of machine learning (ML) and artificial intelligence (AI), said Zylberberg. Access to therapies, cost, and supply chain challenges remain a concern.

As the field evolves to accommodate a higher volume of products and to increase the capacity for delivering these therapies to people who need them, the Forum on Regenerative Medicine convened experts from industry, government, academia, and patient advocacy groups for a public workshop on October 17, 2023, to discuss the current challenges and opportunities in emerging technologies in manufacturing regenerative medicine therapies. The discussions focused on the role of emerging technologies, evolving manufacturing approaches, regulatory considerations, and diverse partnerships needed to deliver these therapies to patients, said Scott Steele, senior advisor with the Center for Biologics Evaluation and Research (CBER) at the FDA. Furthermore, the overall goals for the workshop were to (1) explore gaps and opportunities for new biomanufacturing technologies to change the future of cell and gene therapy, (2) highlight available models of decentralized or distributed manufacturing for regenerative medicine, (3) explore the role of automation and ML and AI in biomanufacturing, (4) discuss ways to ensure quality control (QC) and improve

__________________

manufacturing reliability, and (5) consider strategies to build partnerships and promote technology transfer.

OPENING KEYNOTE AND REFLECTIONS ON REGULATORY AND MANUFACTURING CONSIDERATIONS

With the growing use of CRISPR-Cas9, the regenerative medicine field is likely to see an increased development of cell-based gene therapies, said Peter Marks, director of CBER at FDA.³ Such changes are evident in the CAR-T cell therapy space as CRISPR-Cas9 genome editing has transformed the task of developing allogenic CAR-T cells into a relatively easy process, and the shift from autologous to allogenic CAR-T cells could be transformational in this cell therapy field, said Marks. For example, using allogenic CAR-T cells to treat cancer, particularly hematologic malignancies, could result in greater efficacy compared to autologous CAR-T cells derived from chemotherapy-treated individuals. Likewise, allogenic CAR-T cell therapies may benefit from improved manufacturing consistency, immediate availability, and decreased cost by facilitating the production of off-the-shelf products. Genome editing might also allow developers to overcome some challenges that have previously hindered the development of CAR-T cell therapies for solid tumors.

A key regulatory consideration for regenerative medicine manufacturing is stage-appropriate chemistry, manufacturing, and controls (CMC), said Marks. While FDA allows more leeway regarding CMC in the early phases of human clinical trials, the agency becomes more stringent as a product moves closer to approval. One challenge FDA has seen with CAR-T cell therapies is that they tend to progress rapidly from early- to late-phase development, compressing the time developers have to solidify CMC specifications. Another manufacturing consideration is product characterization. This is especially true for non-genetically modified cell therapies, which generally need more clearly defined critical quality attributes and potency assays to ensure better product reproducibility, said Marks. Developers’ reliance on contract manufacturers can exacerbate these reproducibility problems as even using the same protocol, product manufacturer in different facilities may result in products with different quality attributes. Likewise, the distributed manufacturing model can introduce further quality control challenges.⁴

The hope is that automated and closed system manufacturing of cellular therapy products can address some of these current manufacturing obstacles, said Marks. In his view, the field has underutilized AI to continually control important manufacturing parameters. For example, with CAR-T cells, one manufacturing issue is the need for lentiviral vectors, which is often the rate-limiting and most costly step in the production process. Automating vector production could help address these problems (Guan et al., 2022). CMC technology progress, though, is not keeping pace with the current advancements in science, said Marks. In response, he called for academia to better reward researchers for working on what some might consider the not-so-interesting aspects of product manufacturing, such as purification and characterization methods. To further promote the growth of regenerative therapy CMC, the manufacturing process could be treated like open-source code, with the intellectual property value placed in the constructs themselves rather than how they are made.

FDA has issued guidance for cell and gene therapy developers, including a July 2023 draft guidance document on managing manufacturing changes and product consistency, said Marks.⁵ The agency has also prepared a draft guidance on developing CAR-T cell therapies and human gene therapy products incorporating human genome editing.⁶ For novel products intended to address serious or life-threatening disease or conditions, FDA can designate them as a regenerative medicine advanced therapy. Such therapies are eligible for priority review and accelerated approval

__________________

and can fulfill post-approval requirements using clinical evidence, larger confirmatory datasets, and extended post-approval monitoring of all patients treated prior to approval.

Marks encouraged developers to take advantage of early, nonbinding regulatory meetings with FDA to discuss nonclinical studies, manufacturing, and clinical development plans. Opportunities to do so include Initial Target Engagement for Regulatory Advice on CBER/CDER (Center for Drug Evaluation and Research) Products (INTERACT) meetings for novel products with unique challenges during early development and CBER Advanced Technology Team meetings for advanced technologies or platforms needed to develop CBER-regulated products. These meetings benefit both the developer, by providing nonbinding feedback, and FDA, by giving the agency an early look at new technologies. When considering the regulation of early-stage regenerative medicine technologies, FDA must carefully balance the benefit of approving these products with the risk that adverse effects might take years to become apparent, much as they did with retroviral gene therapies, said Marks.

When the National Organization for Rare Disorders (NORD) surveyed its community about cell and gene therapies, respondents noted concerns about safety, efficacy, response durability, and availability, said Karin Hoelzer, director of policy and regulatory affairs at NORD, in reaction to Mark’s presentation. Yet, since many in the community are unaware of the importance of CMC, manufacturing was not considered a current concern. She identified three key issues from the survey in which manufacturing plays a significant role: (1) navigating the FDA approval process in a timely manner, (2) ensuring affordability under innovative payment models, and (3) providing timely, equitable, and local access to therapies. Because of their powerful role in advocacy, said Hoelzer, the patient community needs to understand CMC and be included in these discussions, which could help both highlight the importance of CMC to the patient community and motivate improvements upon some of these manufacturing challenges.

CMC is important in the large and continuously growing regenerative medicine field, said Sarah Nikiforow, medical director of the cell manipulation core facility at Dana-Farber Cancer Institute. Academic centers around the country serve as de facto contract manufacturing organizations, producing cellular therapy products for multiple sites, because of their ability to rapidly modify and revalidate production processes during phase I clinical trials, said Nikiforow. However, she voiced concerns about the lack of communication between the different independent entities invested in therapy manufacturing and the lack of training available at earlier stages. Professional societies and regulatory and accreditation bodies can help provide such education and address other CMC challenges.

QUALITY CONTROL AND REGULATORY CONSIDERATIONS

Raw Material Quality as a Factor in Reproducible Manufacturing

While regenerative medicines account for approximately 2 percent of all NIH listed clinical trials, they are responsible for about 40 percent of all clinical holds (Wills et al., 2023),⁷ said Sadik Kassim, chief technology officer of genetic medicines at the Danaher Corporation. Of these clinical holds, about 25 percent are related to issues with CMC (Wills et al., 2023). The raw materials used in regenerative medicine therapies are often more complex than traditional pharmaceutical products. Thus, these clinical holds could likely be the consequence of insufficient characterization of the complex input materials and their critical quality attributes (Tsokas et al., 2019). With the sourcing, quality, and analytical characteristics of raw materials playing a crucial role in all stages of product development, the overall manufacturing of a therapy is dependent on the reliability of these materials, said Kassim, as highlighted by two critical raw materials at different stages of development: viral vectors and guide RNAs.

Historically, viral vectors have been viewed with extreme caution, but with their current clinical safety record, developmental evolution, and extensive characterization,

__________________

they are now more broadly available and can be easily manufactured using a decentralized system. In contrast, guide RNAs for CRISPR-based gene editing are in the earlier stages of development, so much is still unknown about their safety and efficacy. Guide RNA variants, for example, have significant differences in target editing efficiencies, which could potentially result in off-target profile changes that impact the safety of these therapies, said Kassim. Similarly, the delivery method, purity, and downstream purification method of guide RNA and Cas9 complexes can also affect the overall viability of the edited cell. Cell viability could theoretically be a critical attribute for these cells, yet this metric can often be misleading. For instance, many cell-based therapies have failed the industry-standard 70 percent viability threshold, but still produce sufficient clinical outcomes when administered under a compassionate use exemption, said Kassim. “If you measure cell viability as an in-process parameter and then prematurely terminate that process, is [that patient denied] access based on something that makes sense intrinsically but, in the end, does not determine that much in terms of complete response?” asked Kassim.

With early technology such as CRISPR nucleases, guide RNAs, and CRISPR gene-edited cell therapies, the synthesis may be easier than with viral vectors, but the unknown design space and minimal manufacturing knowledge may limit the field’s ability to accelerate translation or decentralize manufacturing. A federated learning model, like the Machine Learning Ledger Orchestration for Drug DiscoverY (MELLODDY) project, may be a solution to accelerating adoption and distributing novel technologies, said Kassim.

Meeting Demand for Cell-Based Therapies

As more cell therapies are being commercially approved in the United States and other jurisdictions, contract development and manufacturing organizations (CDMOs) are transitioning to support both clinical and commercial manufacturing, said Matthew Hewitt, vice president and technical officer of cell and gene therapies and biologics at Charles River Laboratories. In addition to rapid increases in investigational new drug applications (INDs) being submitted to FDA, therapeutic demand for commercial CAR-T cell therapies continues to outstrip supply. As of 2022 the median waitlist time to receive a commercial CAR-T cell therapy was six months (Kourelis et al., 2023). The issue is unlikely patient adoption but an inability to produce enough therapies to meet patient demand, said Hewitt.

“We are currently in a situation where we have to learn and innovate on the way we are manufacturing to properly address the markets,” said Hewitt. Ideally, he envisions a future where centralized CDMO facilities continue to manufacture clinical and commercial products while also overseeing decentralized commercial manufacturing sites located near or connected to medical centers in a franchise-like model. Such a design could help streamline manufacturing logistics and reduce vein-to-vein time but would require automated closed systems and regulatory guidance. Once these manufacturing challenges are resolved, though, QC testing, quality assurance’s ability to release products, and supply chain logistics for decentralized facilities will still remain a problem. Decentralization could be a solution to some of these issues, said Hewitt. Furthermore, CDMOs often have trouble manufacturing a therapeutic developer’s product because of an incomplete understanding of nuances that only the developer understands. “If you have nuance, you do not have a GMP (good manufacturing practice) process,” said Hewitt.

Regulatory Perspective on Decentralized Manufacturing

FDA/CBER has not yet established regulatory guidance or policy for cell and gene therapies manufactured using point-of-care or distributed manufacturing models, said Heather Lombardi, director of CBER’s Office of Cellular Therapy and Human Tissue. “We are in the process of trying to understand all of these different approaches and formulate policy accordingly,” said Lombardi. The agency considers distributed manufacturing to be a subset of decentralized manufacturing and defines decentralized manufacturing as production that occurs at multiple geographically dispersed locations. As a subset of decentralized manufacturing, distributed manufacturing likewise involves production at geographically dispersed sites, but these facilities are identical in design and process and are under

the oversight of a single unified quality system. This form of manufacturing can shorten supply chains and increase supply reliability. Alternatively, with point-of-care manufacturing, production occurs at host sites close to patient care. From a regulatory perspective, one challenge with decentralized and point-of-care distributed manufacturing is maintaining comparability between sites. Other concerns, particularly with point-of-care manufacturing in a hospital setting, include proper training for staff and maintaining aseptic environments.

For cellular therapies, FDA approval requires a biologics license application and data from clinical trials conducted under an IND. Any instruments used as part of the manufacturing process are considered manufacturing equipment and are also reviewed under an IND and biologics license application. FDA’s current thinking is that the IND sponsor must ensure that all facilities are following the same manufacturing process, including in-process and final product testing, and that the manufacturing process at each facility follows all regulations and current GMP requirements, said Lombardi. Each manufacturing site will need to be registered and inspected before licensure. Moreover, from a regulatory perspective, understanding how cells behave throughout manufacturing helps instill confidence about the process. Since the agency receives many questions about qualification and process validation, Lombardi suggested that sponsors meet with FDA to discuss their unique situation.

FDA has ongoing initiatives to support advanced manufacturing, including the framework for regulatory advanced manufacturing evaluation (FRAME), said Lombardi. With FRAME, the agency aims to engage with relevant parties to ensure their regulatory framework for advanced manufacturing technologies is science- and risk-based and compatible with future advanced manufacturing technologies. This space provides clarification of manufacturers’ expectations and ensures global regulatory practices are clear. FDA also recently participated in a distributed and point-of-care public manufacturing workshop where the challenges and needs of relevant parties were discussed. With respect to harmonization across global regulatory bodies, Lombardi said that FDA/CBER needs more experience with distributed and point-of-care manufacturing approaches and encouraged sponsors to bring their proposals directly to FDA to facilitate discussions earlier in the process and help make advances in the field.

Challenges of Using QC Testing to Accelerate the Production Process

In the discussion, speakers considered where in the production timeline more analytics might be needed to promote faster regulatory approval. Kassim noted that talking about enhancing process speed is important, but knowing the directionality of the analytics is likewise necessary. Developers need to recognize that capturing the relevant analytics is difficult, as the field has yet to establish what attributes determine potency, and that there is still much to learn about mechanisms of action, said Kassim. Sharing information between developers could help overcome this hurdle and better define acceptance criteria. Capturing these analytics with in-process release is possible and efficient, said Hewitt, but requires a close coordination of everyone involved in the QC testing and a short window of opportunity. Automation could help alleviate this burden, though, said Hewitt, and Lombardi further emphasized the need for modernization of these testing methods. Additionally, greater regulatory flexibility is needed to promote strategies for building in earlier testing timepoints and the adoption of alternative compendial tests, including rapid tests, rather than relying on post-release results, said Lombardi.

DECENTRALIZED MANUFACTURING AS A POSSIBLE STRATEGY TO ADDRESS PRODUCTION AND SUPPLY CHAIN CHALLENGES

A Vertically Integrated Development and Manufacturing Approach

As a new manufacturing model, integrated development and manufacturing organizations (IDMOs) are distinctly different than conventional CDMOs through their implementation of vertical integration and technology integration, said Fabian Gerlinghaus, cofounder and chief executive officer of Cellares. Instead of relying on third-party equipment, Cellares, who defined the term IDMO, has built a cell manufacturing therapy system (Cell Shuttle) to provide cost-efficient end-to-end automation

and high-throughput production of regenerative medicine therapies. This technology can then be utilized at different Cellares IDMO facilities worldwide, irrespective of whether the facility is centralized or decentralized, said Gerlinghaus.

The end-to-end automation of the IDMO-supported process allows for a tenfold increase in productivity compared to a CDMO, said Gerlinghaus. This closed and automated manufacturing technology also eliminates opportunities for contamination and decreases the amount of labor required. By automating the manufacturing process, the opportunity for operator error likewise decreases. Together, these two improvements reduce the failure rate by some 75 percent, resulting in increased product quality, said Gerlinghaus. Specifically, throughout 2023, Cellares has been manufacturing CAR-T cell therapies representative of those currently FDA approved, with critical quality attributes far exceeding the release specifications for commercial CAR-T products. By reducing manufacturing costs, the Cell Shuttle system can produce products at a 50 percent lower cost compared with those manufactured by traditional CDMOs that rely on traditional benchtop equipment, said Gerlinghaus. With decreased costs, an unlimited capacity, and broad applicability, this system provides an opportunity for increasing patient access to much-needed therapies. Minor modifications to the standard process template would be needed to translate the process to different constructs and genes of interest. Such standardization, though, could reduce the work and effort required by regulatory bodies like the FDA to review the CMC package and accelerate the time to filing the IND, he said.

End-to-End Manufacturing Platform for Autologous Cell Therapies

Autologous cell therapies are not reliant on economy of scale, said Rahul Singhvi, cofounder and chief executive officer of Resilience. Therefore, to reduce the unit cost of producing such products, the three main drivers of cost could be considered: processing cost, reagent costs, and QC. Industry is doing a good job of reducing the cost of reagents, and systems such as the one Gerlinghaus discussed could dramatically reduce processing costs, said Singhvi. Resilience’s priority is to attempt to reduce QC costs by creating a multi-omicsbased system that can automatically and rapidly process cells at a lower cost and perform the required QC testing in a centralized or decentralized platform, said Singhvi. This QC box could then be used in conjunction with any preestablished cell processing system and could broaden the understanding of therapy biology.

To promote transparency and allow customers insight into the production of their therapies, this entire QC process is digitized, providing constant monitoring of the entire process through an online portal. “That is a key enabler to achieve our vision of decentralized manufacturing,” said Singhvi. Creating such a system requires developing new approaches to analyzing cell therapy products. By developing a different set of analytical tools, Resilience may both reduce the amount of time required to release a product and gain more insights into critical quality attributes of a product, which could correlate with improved patient outcomes, he said.

Point-of-Care Manufacturing at Academic Centers

In contrast, Nikiforow said that her facility at the Dana-Farber Cancer Institute is not automated and currently relies on 70 people supporting 80 clinical trials and 25 investigational new drug applications with complex manufacturing. One advantage of manufacturing at a point-of-care academic center is that the logistics are simpler. There are no silos between manufacturing, QC testing, and quality assurance groups, so the process is faster. Another benefit comes from having the production facility closer to the patient, enabling more flexibility in manufacturing and more direct communication with clinicians.

When considering an automated manufacturing system, challenges arise upon dealing with the heterogeneity of cells in normal donor and patient blood draws, said Nikiforow. Even within the same patient population and vector, there can be large variations in transduction efficiencies across and between protocols and automated systems, she added. Furthermore, with decentralization, dealing with out-of-specification products promptly may be difficult. Additional concerns are that individual sites may not have the capability to conduct in-process and release testing, and that sending samples for central

analysis would add time and cost to the process, she said.

Considerations for Selecting a Manufacturing Model

During the discussion, speakers explored the main considerations for selecting different models of manufacturing. The choice comes down to patient access in terms of scalability and immediate local access, said Nikiforow. Alternatively, physicians are often uncomfortable with releasing control of the manufacturing process to a centralized facility, Gerlinghaus said. The solution could be developing a good chain of custody dashboard that provides step-by-step information about a product during processing, he added. Moreover, distributed manufacturing also shortens the vein-to-vein time and eliminates the need to freeze cells, said Gerlinghaus. Decentralization with automation, coupled with cloud-based data and a new approach to in-process and final product QC testing, can shorten the time to release and increase the number of patients that can be served, said Singhvi.

AUTOMATION AND ALGORITHMS IN REGENERATIVE MEDICINE MANUFACTURING

A Scalable Biomanufacturing Platform for Personalized Regenerative Therapies

Induced pluripotent stem cells (iPSCs) offer a powerful source of stem cells for personalized regenerative therapies, said Nabiha Saklayen, cofounder and chief executive officer of Cellino. Autologous iPSC-derived therapies avoid the necessity of immunosuppression or donor matching, making cell therapies more accessible for immunocompromised and genetically diverse populations. Currently, manufacturing iPSCs is difficult and time-consuming, requiring highly skilled scientists to repeatedly evaluate and passage cells manually in a clean room. However, recent advances in AI-powered control algorithms, optics for bioprocessing, and fluidics have enabled Cellino to develop an autonomous, closed, and modular platform for manufacturing these cells. Such a cassette-based manufacturing platform is important as manufacturers consider how to provide local accessibility to patients, said Saklayen.

The Cellino bioprocessing platform relies on time-lapse imaging that visualizes cells repeatedly and feeds data into an ML-based prediction algorithm. This control algorithm detects high-quality clones and make decisions that affect the biomanufacturing process, said Saklayen. Unwanted cells can then be removed using a proprietary method involving laser-generated bubbles, thus eliminating the need for passaging. The entire process can be monitored by a web-based platform that tracks the experiments and provides visibility and transparency to the bioprocess. Currently, QC is the biggest bottleneck, as the platform does not yet replace existing cell assays but develops control algorithms to make better in-process decisions over time.

Robust collaborations between experts in academia, industry, regulatory affairs, and biomanufacturing are likely necessary to overcome the current QC challenges and support the rapidly advancing field, especially as more products transition to clinical trials. Saklayen posed the following questions to be considered as the regenerative medicine field thinks about the next steps in biomanufacturing innovation and iPSC-derived therapies:

• How does the field establish a new standard approach compared to manual biomanufacturing?
• How does the field define a “good” iPSC?
• How does the field accelerate the path of standardization to meet clinical trial timelines?
• How does the field implement good ML practices for AI-based biomanufacturing processes?

Many different aspects of the field need these innovations, said Saklayen, but bringing all of the required pieces together without collaboration will be quite difficult.

Addressing Biology through Mathematics to Enable Cell Manufacturing

The biology behind biomanufacturing processes can be controlled if the mathematics dictating the events in these processes are better understood, said Jan Jensen, founder, chief executive officer, and chief scientific officer at Trailhead Biosystems. For example, combining large-dimensioned experiments testing combinatorial inputs in a cell system with multivariate data analyses can produce mathematically defined equational

representations of the system’s behavior. Trailhead Biosystems seeks to accomplish this by developing perturbational and computational methods that evaluate which signaling input combinations may control cells as they undergo differentiation or attain homeostatic stability. This process is not reliant on AI but instead is a means of empirically testing a process through combinations.

Through their mathematical methods, the complex interactions occurring within a biological system can be recognized rather than just identifying the primary effects or gene expression. “We do not want to guess our way through differentiation, we need numbers underneath it,” said Jensen. Therefore, the process-development work moves away from hypothesis-driven approaches to a system that can better identify critical process parameters and operate with more statistical confidence. Computerized, robotically executed experiments can further improve the speed, precision, and scalability of interrogating the process space and generate novel empirical data. With this approach, Trailhead Biosystems can build new protocols for multiple cell types, driven by their own data and compliant with quality-by-design standards. The versatility of such methods results in cost-effective protocol development that can be universally applied, said Jensen, and has thus far produced high-purity and functional cells from individual donors.

The Path to Automation of Manufacturing: A Case Study

A case study in which an automated platform was used to manufacture an implant for retinal pigment epithelial (RPE) cell and retina degeneration due to dry age-related macular degeneration was presented by Jane Lebkowski, president of Regenerative Patch Technologies. This tissue-engineered implant contained pluripotent stem-cell-derived RPE cells embedded on an ultrathin parylene membrane. Together, these components could replace the degenerated RPE and Bruch’s membrane in areas of geographic atrophy and address an unmet medical need of the hundred thousands of patients who develop dry age-related macular degeneration annually, said Lebkowski. The first step to developing such a product is to identify the critical quality attributes and characteristics of the cells, investigating both their phenotype and function, said Lebkowski, and the next step is then manufacturing the implant, which involves developing two processes: cell production and implant production.

When developing an automated process for cell production, the required cell number, the current size of the master and cell banks, and the variability of yields, purity, and function across different banks should be considered, said Lebkowski. Furthermore, AI could be beneficial during this stage of process development by helping characterize the cell phenotype, function, composition, and uniformity between batches. Evaluating their process manually gave Regenerative Patch Technologies more confidence that they had enough reproducibility and control over the process to automate it, said Lebkowski. To incorporate the RPE cells with the engineered implant, they are now developing ML- and AI-powered tools for in-process testing, rejection of incomplete or damaged implants, and eventual release of the implants. Key considerations for automating this process, said Lebkowski, include maintaining a sterile, closed system; ensuring the compatibility of materials with cell viability and growth; processing times; imaging during production and for in-process specifications; and using AI to develop structural and functional specifications.

Evolution of ML and AI Technologies and Their Regulatory Impacts

Considering the recent evolutions of ML and AI and their effect on manufacturing regenerative medicine therapies, Saklayen acknowledged the role of the self-driving car industry in developing ML and AI algorithms for image analysis. The regenerative medicine industry is now “starting to see innovation happening where [the field is] connecting the imaging space with the genomics data space,” said Saklayen. The workforce is also changing in response to changes in ML and AI technologies with a growing number of ML specialists and multidisciplinary experts entering the biomanufacturing field, said Saklayen and Lebkowski. There is also a growing potential for ML and AI to assist in quantitating and characterizing various cell parameters, said Lebkowski.

When asked to consider how regulators might adopt these new technologies as part of in-process and release testing, Saklayen was optimistic about the progress in that space, commenting on recent successes by her collaborators at the National Eye Institute. Their experiences have highlighted the need to show that these advanced technologies increase consistency and yield. It will also be important, though, to demonstrate that an algorithm is fit for the purpose and how it affects manufacturing decisions, said Lebkowski. Meanwhile, Jensen highlighted the need for continued communication between agencies and product developers to establish new guidelines.

IMPLEMENTING NEW MANUFACTURING STRATEGIES THROUGH PARTNERSHIPS AND INNOVATIVE TECHNOLOGY TRANSFER

Industrialization of Cell Therapies through Shared Manufacturing Platforms

While cellular and gene therapies are more complex than traditional pharmaceuticals, the regenerative medicine field can learn much from the industrialization of these large molecule therapies, said Jens Vogel, senior vice president and head of Bayer Biotech’s pharmaceutical product supply operations. What allowed this field to industrialize traditional pharmaceuticals more quickly and efficiently was the convergence upon common processing methods such as standard bioreactors and purification protocols, said Vogel. This manufacturing platform was developed using a quality-by-design-based CMC approach that relied on a deep fundamental understanding of the product and process and a related control strategy.

Translating the lessons learned from industrializing traditional pharmaceuticals to cell therapies, said Vogel, requires five elements: software, hardware, an operating system, people and culture, and flexible facilities. Specifically, software refers to process platform and the critical quality attributes needed to control the process. The hardware is the equipment platform, and operating system correlates to the CMC ecosystems, including the quality-by-design-based approach and development strategies. Vogel envisions the future of industrialized cell-based therapy hardware platforms to be modular, flexible, highly integrated, and automated. Such a model could provide an end-to-end single-use flow path capable of producing diverse allogenic cell therapies with every module running on the same operating system in a plug-and-play manner, thus providing a means to accelerate technology transfers and scaling, improve the reliability of scaling, and ultimately bring products to patients more quickly and reliably, said Vogel.

Meeting the Clinical Needs of People with Rare Diseases

Despite having already solved the technical challenges of editing blood in the clinic, there are no clinical trials ongoing for gene editing and only one gene therapy product has been approved, said Fyodor Urnov, professor of molecular therapeutics and scientific director of the Innovative Genomics Institute at the University of California–Berkeley. One problem confronting the field is the large spectrum of variants within a single condition, making a disorder appear rare and requiring more tailored gene therapies. For example, at least 114,189 individuals globally have been diagnosed with a specific primary immunodeficiency defect such as the 450 known single-gene inborn errors of immunity (Quinn et al., 2022). Producing a gene therapy for these patients would take an estimated four years and cost $7 million at an academic center, leaving a costly and time-consuming path for these patients to receive essential treatments, said Urnov. The thousands of U.S. newborns with inborn errors of metabolism in the liver are similarly left without treatment options, despite the rapid progresses in liver genome editing (Waters et al., 2018).

To address these unmet medical needs for such rare diseases, nonclinical pharmacology and toxicology, CMC, and regulatory frameworks should be upgraded to better accommodate the established CRISPR-Cas genome editing platform, said Urnov. Changing the regulatory framework requires addressing the “CRISPR Catch-23”: the only way to evaluate the safety and efficacy of genome editing is to edit more people, yet there is currently not enough of this data necessary to test the product in the clinic with the current system. Moreover, the path to IND approval is long and expensive. One solution is to work in alignment with FDA to promote streamlined approval of bespoke therapies, in which only the gene of interest is changed while the rest of the manufacturing process is maintained.

The Role of Public–Private Partnerships

The goal of the Foundation for the National Institutes of Health (FNIH) is to build bridges toward breakthroughs and accelerate cures by connecting the National Institutes of Health with the private sector in a focused, precompetitive manner, said Courtney Silverthorn, associate vice president for science partnerships at FNIH. One such program at the FNIH is the Bespoke Gene Therapy Consortium (BGTC), the largest and most complex public–private partnership in the Accelerating Medicines Partnership® portfolio and the only project in the program focused on a platform rather than a specific disease or disease area. In particular, BGTC focuses on increasing the efficiency of the preclinical phase and regulator processes and the transition toward an IND and clinical trial. To accomplish this, BGTC has prioritized two critical pathways and eight gene therapies, said Silverthorn. One pathway is rooted in the optimization of vector generation and gene expression for adeno-associated virus gene therapies. The second pathway focuses on establishing a publicly available playbook with streamlined manufacturing, preclinical, and regulatory processes to enable safe acceleration of adeno-associated virus gene therapies. The aim is that these pathways can then be leveraged for other diseases, and the consortium plans to release this model publicly, said Silverthorn.

High-Risk, High-Reward Projects for Accelerating Better Health Outcomes for Everyone

The Advanced Research Projects Agency for Health (ARPA-H) is a recently established U.S. government funding agency with a mandate to make big, bold, yet well-considered investments in research and development that could have a transformative effect on health and health care, said Ravi Basavappa, senior advisor at ARPA-H. Through contracts and mechanisms other than grants, ARPA-H is capable of providing a more expedient review of proposals and disbursement of awards, enabling the organization to anticipate and react to changing circumstances and opportunities quickly. The success of such programs is evaluated based on their ability to develop transformative health solutions that are easily accessible to everyone, regardless of location, social demographics, or community in a cost-effective manner.

The scientific portfolio of ARPA-H is driven by program managers who identify important high-risk, high-reward problems in the health field and establish teams from academia, nonprofits, and industry to innovatively address these issues within a limited time frame. Funding announcements for these projects include requirements for a commercialization plan and intellectual property protection. To support awardees not well versed in commercialization pathways, ARPA-H has established the Project Accelerator Transition Innovation Office (PATIO), which helps strategize and map out commercialization paths. Furthermore, technology transfer is intentionally built into each program, said Basavappa. Through the ARPA-H Health Innovation Network program, the agency aims to engage the entire health care ecosystem in the development process, including patient groups, clinicians, researchers, investors, and regulators.

Enabling Affordable and Accessible Advanced Medicines

The current pricing of approved cell and gene therapies is unsustainable for insurance companies and health systems and unfeasible for low- and middle-income countries, and it restricts access particularly for underserved populations, said Boro Dropulić, executive director of Caring Cross. The high prices for these products result from complex processing logistics and the high infrastructure costs associated with establishing a manufacturing facility. Two pathways could be explored to reduce the cost of and enable affordable access to these therapies: establishing point-of-care manufacturing for autologous therapies and developing simple and robust manufacturing processes. Point-of-care manufacturing, building on the autologous bone marrow transplantation model, is economical and would reduce vein-to-vein times, said Dropulić. Other advantages include improved scheduling for patients and the opportunity to deliver freshly manufactured cells to the patient. “We know that large hospitals are competent in making cell products, and there is really no need for expensive company manufacturing facilities for single products, which drives up the cost,” said Dropulić. Point-of-care manufacturing of CAR-T cell therapies are already producing excellent and reproducible clinical results in patients with leukemia and lymphoma, said Dropulić.

To functionally address these pathways, Caring Cross has initiated a public benefit corporation to provide low-cost vector manufacturing and improve ancillary technologies for activation and surface programming of T-cells. They are also developing a network of clinical collaborators at various hospitals as part of a robust and economical point-of-care paradigm, said Dropulić. Caring Cross is also working on improving vector technology and is collaborating with the National Institute for Innovation in Manufacturing to develop small-scale, open-process methods that will be publicly available.

Obstacles to Forging Partnerships that Enable Technology Transfer

When asked about barriers to forming more partnerships within the field, Vogel, Silverthorn, and Dropulić all cited the lack of standards and standardized platforms. Vogel called for the community to come together in the precompetitive space to settle on standards that will accelerate technology transfer. For Urnov’s organization, money is the biggest obstacle, and he called for the community to share regulatory learnings. Basavappa said a major challenge for ARPA-H’s model is building trust among members of the vast network it is building and with the communities the organization serves.

Working with Patient Advocates

Highlighting the need for patient advocacy, Urnov said that engagement with relevant parties is critical since gene therapies are built not just for patients but with them. Patient advocates play a significant role in capturing regulatory, donor, and manufacturing attention and support. Silverthorn said FNIH has a commitment to include patient voices in every new public–private partnership it develops and in the preclinical and clinical trials it organizes. Patients and caregivers are a critical component of the BGTC, she added, especially those willing to participate in natural history studies of which they may never see the benefit.

FINAL REFLECTIONS ON THE WORKSHOP

During the workshop, Wade Forbes, cofounder and graphic recorder at RedTale Communications, captured the key concepts and themes that were presented by individual speakers and discussed during the day.⁸ He summarized what he heard with an eye toward the future of regenerative medicine manufacturing, noting the gains in the field since the forum’s 2017 workshop on manufacturing and the potential to further excel in this space (Box 1).

__________________

This workshop was a day full of enlightenment, information, learning, and opportunities, said Zylberberg. To promote patient access to the innovative therapies that have come from the regenerative medicine field and to offer additional therapies for diseases that need cures, the field could work harder and think outside the box, she said (see Box 2). “This is just the beginning of the conversation. The more that we brainstorm together as a community, the more we will learn from each other, and most likely we will bring more solutions for our patients and families,” said Zylberberg.

REFERENCES

DISCLAIMER This Proceedings of a Workshop—in Brief has been prepared by Michelle Drewry, Joe Alper, and Sarah H. Beachy as a factual summary of what occurred at the meeting. The statements made are those of the rapporteurs or individual workshop participants and do not necessarily represent the views of all workshop participants; the planning committee; or the National Academies of Sciences, Engineering, and Medicine.

REVIEWERS To ensure that it meets institutional standards for quality and objectivity, this Proceedings of a Workshop—in Brief was reviewed by Nabiha Saklayen, Cellino. Leslie Sim, National Academies of Sciences, Engineering, and Medicine, served as the review coordinator.

SPONSORS This workshop was partially supported by contracts between the National Academy of Sciences and Advanced Regenerative Manufacturing Institute; Akron Biotech; Alliance for Regenerative Medicine; American Society of Gene & Cell Therapy; Burroughs Wellcome Fund (Grant No. 1200542); California Institute for Regenerative Medicine; Centre for Commercialization of Regenerative Medicine; Department of Veterans Affairs (Contract No. VA268-16-C-0051); Food and Drug Administration: Center for Biologics Evaluation and Research (Contract No. R13FD006614); International Society for Cellular Therapy; International Society for Stem Cell Research; Johnson & Johnson; National Institute of Standards and Technology; National Institutes of Health (Contract No. HHSN263201800029I, Task Order No. 75N98019F0084; including National Center for Advancing Translational Sciences; National Eye Institute; National Heart, Lung, and Blood Institute; National Institute on Aging; National Institute of Biomedical Imaging and Bioengineering; and National Institute of Diabetes and Digestive and Kidney Diseases); New York Stem Cell Foundation; and United Therapeutics Corporation. Any opinions, findings, conclusions, or recommendations expressed in this publication do not necessarily reflect the views of any organization or agency that provided support for the project.

STAFF Sarah H. Beachy, Senior Program Officer, Samantha N. Schumm, Program Officer, Michelle Drewry, Associate Program Officer, Kathryn Asalone, Associate Program Officer, Lydia Teferra, Research Assistant, and Ashley Pitt, Senior Program Assistant.

For additional information regarding the workshop, visit https://www.nationalacademies.org/event/40391_10-2023_emerging-technologies-and-innovation-in-manufacturing-regenerative-medicine-therapies-a-workshop.

Suggested citation: National Academies of Sciences, Engineering, and Medicine. 2024. Emerging technologies and innovation in manufacturing regenerative medicine therapies: Proceedings of a workshop—in brief. Washington, DC: The National Academies Press. https://doi.org/10.17226/27483.

Health and Medicine Division Copyright 2024 by the National Academy of Sciences. All rights reserved.