- Key takeaways
- Three modalities, three manufacturing paradigms
- What makes viral vector manufacturing so difficult to scale?
- Autologous cell therapy: when the patient is the batch
- How is mRNA manufacturing different from conventional biologic production?
- Plasmid DNA: the overlooked upstream bottleneck
- What does the regulatory landscape for CGT manufacturing look like?
Cell therapies, gene therapies, and mRNA therapeutics share one inconvenient truth with the biologics they are replacing: the biology of the drug and the engineering of the manufacturing process are both problems that must be solved at the same time and at scale. Each modality brings a distinct manufacturing challenge, and none of them can be addressed by adapting a facility designed for conventional biologics. This article examines the core production bottlenecks across all three categories and the engineering approaches the industry is deploying to overcome them.
Key takeaways |
|
|
|
|
|
For the foundational bioprocessing context in which advanced therapy manufacturing sits within the broader drug development and scale-up landscape, see the ultimate guide to bioprocessing scale-up and Pharma 4.0.
Three modalities, three manufacturing paradigms
The defining characteristic of advanced therapy manufacturing is that the three major modalities in the category share almost nothing at the process level. A viral vector production facility optimized for adeno-associated virus uses adherent or suspension HEK293 cell culture, triple transfection with plasmid DNA, and a downstream purification train built around ion exchange and affinity chromatography. An autologous CAR-T manufacturing suite starts with a patient leukapheresis product, performs T cell activation and lentiviral transduction in a closed small-scale bioreactor, and formulates a patient-specific infusion product with a shelf life measured in hours. An mRNA facility runs E. coli fermentation for plasmid DNA amplification, enzymatic in vitro transcription, and microfluidic lipid nanoparticle encapsulation at continuous flow rates measured in liters per hour.
The consequence is that the conventional pharmaceutical strategy of building one facility and qualifying it for multiple products does not transfer straightforwardly to the CGT space. Platform manufacturing, in which the same basic process equipment and methodology handles a range of molecules in the same product class, has been transformative for monoclonal antibody production, where Protein A capture, ion exchange polishing, and UF/DF formulation apply broadly across most IgG-class molecules. No such platform exists for the three advanced therapy modalities together, and even within each modality the process-to-process variation is substantially higher than in the mAb world.
Modality | Production system | Scale model | Primary manufacturing challenge | Cost driver |
AAV gene therapy | HEK293 suspension or adherent; baculovirus/insect cell; stable producer lines under development | Clinical: 10-200 L bioreactors. Commercial: 200-2,000 L required; capacity severely constrained | Downstream purification: yields below 50%; empty-to-full capsid ratio; potency characterization | Capital intensity of GMP vector manufacturing suites; CRO/CDMO pricing for constrained capacity |
Lentiviral vector (CAR-T ex vivo) | Transient transfection in HEK293; stable packaging cell lines increasing | Batch sizes limited by transient transfection scale; continuous manufacturing approaches emerging | Transient transfection is not scalable indefinitely; plasmid costs; titer and particle quality variability | Plasmid DNA supply; GMP-grade production costs; downstream VSV-G inactivation requirements |
Autologous CAR-T cell therapy | Patient-derived T cells; lentiviral or retroviral transduction ex vivo; closed-system bioreactors | Single patient dose per manufacturing run; decentralized or centralized processing; manufacturing time bounded by patient timeline | Logistics chain from leukapheresis to infusion; QC release within patient timeline; closed-system contamination prevention | $250,000-$450,000 per infusion driven by individualized labor, specialized equipment, and QC overhead per batch |
Allogeneic cell therapy (off-the-shelf) | Allogeneic donor-derived T cells, NK cells, or iPSC-derived; engineered to avoid host rejection | Larger production batches from single donor run; same batch used for multiple patients; more scalable than autologous | Immunogenicity management; HLA-matching or editing to prevent rejection; cryopreservation stability across patients | Higher upfront development cost for cell line engineering; distribution and cold chain for multi-patient batch |
mRNA therapeutic (LNP-formulated) | E. coli fermentation for pDNA; enzymatic IVT; microfluidic or impingement jet LNP encapsulation | Continuous or semi-continuous at clinical scale; highly scalable upstream; formulation scale-up more challenging | LNP size, polydispersity, and encapsulation efficiency consistency at scale; pDNA quality and yield | Low raw material cost relative to biologics; formulation equipment capital; pDNA quality requirements at commercial volumes |
What makes viral vector manufacturing so difficult to scale?
The viral vector manufacturing challenge is not a shortage of bioreactor capacity in the conventional sense. It is a combination of process immaturity, downstream yield limitations, and analytical complexity that prevents the industry from producing sufficient quantities of quality-characterized vector to advance the number of programs currently in development.
Downstream purification is the primary yield constraint. Technical analysis of the viral vector industry consistently identifies low chromatographic recovery as a structural problem: yields for AAV and lentiviral vectors from downstream purification are typically below 50%, compared to greater-than-90% recovery achieved routinely in monoclonal antibody platform processes. Each chromatographic step that a vector product passes through removes a fraction of the product, and the lack of standardized platform processes means that every new vector product requires process development from near-scratch.
The empty-to-full capsid ratio is an additional analytical and process challenge specific to AAV. Full capsids, which contain the therapeutic gene of interest, are the active drug substance. Empty capsids, which form in parallel during production and package no genetic payload, are process-related impurities that are physically nearly identical to full capsids and extremely difficult to remove by standard chromatographic means. Ultracentrifugation or analytical ultracentrifugation can separate them, but neither method is practically scalable to commercial volumes without substantial process development investment.
The longer-term context for the viral vector capacity challenge, including the transition from artisan-scale to industrialized manufacturing and the role of stable producer cell lines in resolving the transient transfection scalability ceiling, is covered in the related article on the viral vector bottleneck and the road from the artisan workshop to the factory floor.
Autologous cell therapy: when the patient is the batch
Autologous CAR-T cell therapy is manufactured from a blood product drawn from the same patient who will receive the finished therapy. The manufacturing process receives the patient's T cells from a leukapheresis procedure, activates and expands them in culture, introduces the chimeric antigen receptor gene via lentiviral or retroviral transduction, tests the modified cells for identity, potency, and sterility, and formulates them for re-infusion into the patient. The total manufacturing time is typically two to five weeks, bounded not by production efficiency targets but by the patient's disease progression.
This manufacturing model creates logistics requirements that have no parallel in conventional pharmaceutical production. The leukapheresis product must be transported under controlled temperature conditions from the collection site to the manufacturing facility. The finished drug product must be transported back to the clinical site, with a shelf life that may be as short as a few days for non-cryopreserved products. The entire chain must maintain chain-of-custody and patient identity documentation at every step to prevent a patient receiving another patient's therapy. Every manufacturing deviation, quality failure, or logistical delay directly affects a specific patient waiting for that dose.
The operational and financial implications of this patient-specific manufacturing model, including the cost structure that produces autologous therapy price points in the range of $250,000 to $450,000 per infusion, and the strategies for managing the vein-to-vein logistics chain at scale, are explored in the related DDN coverage of the manufacturing and logistics realities of CAR-T and cell therapy at scale.
How is mRNA manufacturing different from conventional biologic production?
mRNA therapeutics are produced through a fundamentally different manufacturing chemistry than conventional biologics. Where proteins are produced by living cells in bioreactors and purified through a multi-step downstream train, mRNA is synthesized enzymatically in vitro from a DNA template. The in vitro transcription (IVT) reaction uses a bacteriophage RNA polymerase to transcribe the plasmid DNA template into mRNA at room temperature in a matter of hours. The resulting mRNA is then purified by chromatography and formulated into lipid nanoparticles that protect it from enzymatic degradation and deliver it into target cells.
The speed advantage of this manufacturing chemistry was demonstrated definitively in the COVID-19 vaccine response, where mRNA vaccines reached clinical trials within months of sequence publication. The scalability challenge lies primarily in the formulation step. Lipid nanoparticle encapsulation requires precise mixing of an ethanolic lipid solution with the aqueous mRNA in milliseconds at controlled flow rate ratios, and the particle size, polydispersity, and encapsulation efficiency of the resulting LNPs are sensitive functions of mixing dynamics, flow rates, and buffer conditions. Maintaining these parameters consistently from development through commercial scale is the primary mRNA manufacturing engineering challenge.
The IVT process science and LNP mixing parameters that define mRNA drug product quality are covered in the related DDN article on LNP formulation for mRNA vaccines. The engineering of the single-use mixing systems that enable consistent LNP formation at scale, including the hollow fiber and microfluidic mixing architectures, is covered from a fluid path and membrane engineering perspective at Separation Science, in the article on single-use mixers in formulating lipid nanoparticles.
Plasmid DNA: the overlooked upstream bottleneck
Every advanced therapy manufacturing process runs upstream of one critical raw material: plasmid DNA. Plasmid DNA is the IVT template for mRNA manufacturing. It is the transfer plasmid carrying the therapeutic gene for AAV and lentiviral vector production. It is the vector backbone for non-viral gene delivery approaches currently in development. The quality and consistency of the pDNA used in each of these processes directly affects the quality and batch-to-batch consistency of the downstream drug product.
Global pDNA manufacturing capacity has become a structural bottleneck because demand has expanded faster than production capacity. AAV triple transfection processes require large quantities of high-purity, supercoiled plasmid DNA at clinical and commercial scale. mRNA manufacturing requires pDNA templates that meet exacting quality specifications for residual host cell protein, RNA, and nicking to ensure consistent IVT reaction performance. Neither requirement is met by the plasmid preparation quality typically produced in research environments.
GMP-grade pDNA production in E. coli requires cell banking, fermentation scale-up, alkaline lysis at large volumes, and multi-step purification by anion exchange and size exclusion chromatography to meet the quality specifications for pharmaceutical use. The global shortfall in qualified pDNA supply directly limits the number of gene therapy and mRNA programs that can advance through clinical development simultaneously, making pDNA production one of the most commercially valuable and undersupplied capabilities in the advanced therapy manufacturing ecosystem.
What does the regulatory landscape for CGT manufacturing look like?
Cell and gene therapies are regulated as biological products in the United States and as Advanced Therapy Medicinal Products (ATMPs) in the European Union, applying regulatory frameworks that were designed primarily with conventional biologics in mind. The FDA's Center for Biologics Evaluation and Research (CBER) manages CGT product submissions through its Office of Tissues and Advanced Therapies (OTAT), which processed record numbers of IND applications in 2024 and 2025. A systematic review of regulatory approvals for cell and gene therapies through December 2024, published in Therapeutic Innovation and Regulatory Science in 2026, found that the most frequent quality objections in marketing authorization submissions concerned manufacturing comparability, potency assay validation, specifications, and stability data, precisely the areas where CGT manufacturing process maturity lags furthest behind conventional biologics.
The manufacturing comparability challenge arises from the nature of CGT manufacturing: process changes during development inevitably occur, and demonstrating that the product after a process change is comparable to the product before it requires analytical methods that can characterize a complex biological product with sufficient sensitivity to detect clinically meaningful differences. For an autologous CAR-T product where the starting material varies with every patient, the concept of comparability has additional complexity beyond the manufacturing process itself.
For detailed analysis of the regulatory failures that have slowed cell and gene therapy programs at the FDA, including the potency assay and manufacturing characterization gaps most commonly cited in complete response letters, see the related DDN coverage of why gene and cell therapies are stalling at the FDA.
This article was produced under Drug Discovery News' AI Editorial Guidelines.
