Cell therapy development is a long, layered process that turns a biological concept into a treatment that can be reliably manufactured, safely tested in people, and eventually delivered to patients. It spans cell sourcing, genetic engineering, preclinical safety studies, scaled-up manufacturing, quality testing, clinical trials, regulatory review, and ongoing monitoring that can last years after a patient is treated. Each stage introduces its own technical and logistical problems, and a failure at any point can send developers back to earlier steps. The field has grown rapidly since the first chimeric antigen receptor (CAR) T-cell therapy was approved in 2017, but the path from lab bench to clinic remains expensive, uncertain, and full of tradeoffs.
Choosing Where the Cells Come From
The very first decision in cell therapy development is the source of cells. The two main options are autologous cells, which come from the patient who will eventually receive the treatment, and allogeneic cells, which come from a healthy donor. Autologous products sidestep the immune rejection problem because the patient’s body recognizes its own cells. This is why the earliest approved cell therapies were autologous: each batch is made for one specific person.1PubMed. Progress and challenges in developing allogeneic cell therapies The downside is that every patient needs a custom manufacturing run, which makes production slow, costly, and hard to scale.
Allogeneic therapies, by contrast, could be produced in bulk from a single donor’s cells and stockpiled for many patients, much like a conventional drug. The catch is immune rejection. When donor cells are mismatched with the recipient’s immune system, the body mounts a strong inflammatory response. Research on organoid-based cell therapies has shown that fully mismatched donor cells trigger significantly higher levels of inflammatory signals and greater immune cell infiltration compared to partially matched or autologous cells.2Cell Reports Medicine. Immunogenicity of human primary organoid cellular therapies Even autologous cells can provoke a low-level immune response above baseline, a reminder that transplanting any living tissue is rarely immunologically silent. Overcoming rejection reliably remains one of the field’s central unsolved problems, and much of the engineering work described below is aimed at addressing it.
Engineering the Cells
Once cells are sourced, they usually need to be genetically modified to perform a therapeutic function. In CAR T-cell therapy, for example, a patient’s T cells are reprogrammed to recognize and kill cancer cells. The tools used to deliver genetic instructions into cells have evolved considerably, and the choice of tool affects everything from safety to manufacturing complexity.
Lentiviral vectors are one of the most widely used delivery systems. They can carry large and complex genetic payloads and sustain long-term gene expression in both dividing and non-dividing cells.3PubMed Central. Lentiviral Vectors for Delivery of Gene-Editing Systems Based on CRISPR/Cas: Current State and Perspectives Most approved CAR T products use lentiviral or retroviral vectors to permanently insert the CAR gene into T cells. An alternative approach uses CRISPR-Cas9 gene editing, sometimes paired with an adeno-associated virus donor template, to insert a gene precisely at a specific location in the genome rather than at a random spot. Researchers have compared these two strategies head-to-head for correcting the genetic defect in X-linked severe combined immunodeficiency, finding that each has distinct advantages in terms of expression levels and integration patterns.4PubMed Central. CRISPR-Cas9-AAV versus lentivector transduction for genome modification of X-linked severe combined immunodeficiency hematopoietic stem cells
A newer twist involves delivering CRISPR components via integrase-deficient lentiviral vectors, which enter the cell and express the editing machinery without permanently inserting themselves into the genome. This approach can produce rapid and sustained gene editing in dividing cells while reducing the risk of unwanted insertional effects.5Molecular Therapy. Efficient and Specific Gene Editing in Division-Arrested Cells via Transient Delivery of CRISPR/Cas9 from Integrase-Deficient Lentiviral Vectors The broader trend is toward more precise editing with fewer off-target effects, but the field is far from settling on a single best method.
Preclinical Safety Testing
Before any cell therapy reaches a patient, it must go through preclinical studies designed to flag potential dangers. This is harder than it sounds. A cancer-targeting cell therapy might hit its intended antigen on tumor cells perfectly well but also attack healthy tissues that express the same protein at low levels. This “on-target, off-tumor” toxicity is one of the biggest safety concerns in the field, and predicting it in animal models is genuinely difficult because human antigens often are not present in mice.
To get around this, researchers have developed knock-in mouse models where a human protein is expressed under control of the mouse’s own promoter, so it appears on the same tissues where it would appear in a person. One group created mice expressing human HER2 on liver tissue and used them to compare the toxicity of CAR T cells with different binding strengths.6PubMed Central. A rational mouse model to detect on-target, off-tumor CAR T cell toxicity A similar model for PSMA-targeted CAR T cells found that treatment could be lethal in mice within five days, with neurotoxicity and bone marrow changes that closely resembled the side effects seen in human clinical trials.7Journal for ImmunoTherapy of Cancer. 318 Developing a preclinical toxicity model to predict and prevent clinical CAR T cell toxicity in prostate cancer These models are valuable, but they come with an important caveat: they can both underpredict and overpredict what happens in humans. A study on claudin 18.2-targeted CAR T cells found that mouse models sometimes overcall the severity of toxicities, which could lead developers to abandon promising therapies prematurely.8PubMed Central. On-target off-tumor toxicity of claudin18.2-directed CAR-T cells in preclinical models
Safety testing also extends to the vectors used to modify cells. Retroviral vectors integrate into the genome, and where they land matters. Early gene therapy trials in the 2000s saw leukemia emerge in patients whose vectors inserted near cancer-promoting genes. Since then, studies have shown that self-inactivating vector designs dramatically reduce this genotoxic risk compared to older designs with active viral promoter regions.9PubMed Central. The genotoxic potential of retroviral vectors is strongly modulated by vector design and integration site selection in a mouse model of HSC gene therapy A machine-learning tool called SAGA can now classify vectors by their mutagenic risk based on the gene expression patterns they induce, achieving about 91% accuracy on benchmark vectors with known outcomes, and it reliably flags vectors that caused serious adverse events in clinical trials.10PubMed Central. Predicting genotoxicity of viral vectors for stem cell gene therapy using gene expression-based machine learning
Manufacturing and Scale-Up
Making a cell therapy in a research lab is one thing. Making it reliably, repeatedly, and at scale under strict cleanroom conditions is something else entirely. Cell expansion, the step where a small number of engineered cells is grown into a therapeutic dose, is highly sensitive to culture conditions. Temperature drifts, nutrient depletion, and timing mismatches can all alter the final product.
Automation is central to solving this. A comparison between a semi-manual culture platform and its automated counterpart for T-cell manufacturing found that both produced similar cell numbers after ten days of culture, with no statistically significant difference throughout the growth period, and both achieved viability above 85%.11PubMed Central. Transition from manual to automated processes for autologous T cell therapy manufacturing using bioreactor with expandable culture area This kind of head-to-head validation is essential because regulators need to see that switching from a manual to an automated process does not change what the patient receives. More broadly, automation has the potential to address the cost, consistency, and regulatory challenges that currently block large-scale commercialization of cell and gene therapies.12PubMed Central. Automation in cell and gene therapy manufacturing: from past to future
Quality Control and Release Testing
A cell therapy product is not a small molecule that can be fully characterized by its chemical structure. It is a living thing, and proving that each batch meets quality standards requires a suite of specialized assays. The four core attributes regulators look for are identity (confirming the cells are what they are supposed to be), potency (confirming they do what they are supposed to do), purity (confirming contaminants are absent or below acceptable levels), and safety (confirming the product will not introduce infections or other hazards).13PubMed. Developing assays to address identity, potency, purity and safety: cell characterization in cell therapy process development
Potency testing is often the hardest. Unlike a drug where you can measure milligrams, a cell product’s potency depends on biological activity that can vary between donors, between batches, and even between culture runs. The potency assay serves as the key quality gate for batch release and also provides the basis for comparability assessment when the manufacturing process changes, such as during scale-up or a switch to a new donor source.14PubMed. Potency assay development for cellular therapy products: an ISCT review of the requirements and experiences in the industry Despite their importance, these assays are not well harmonized across regulatory authorities. Different agencies use different terminology and sometimes different expectations for the same test categories, which complicates global development programs.15PubMed. Improved harmonization of critical characterization assays across cell therapies
Freezing, Shipping, and Keeping Cells Alive
Most cell therapies need to be cryopreserved at some point during manufacturing or distribution. Freezing and thawing living cells is inherently destructive, and evaluating how well a cryopreservation protocol works is trickier than it might seem. Simply measuring viability right after thawing can give misleading results: some cells appear alive immediately post-thaw but go on to die through delayed apoptosis. Research has shown that several cryoprotective systems gave apparently high viability scores but very low total cell recovery, meaning the “surviving” cells were not actually usable.16PubMed Central. Post-Thaw Culture and Measurement of Total Cell Recovery Is Crucial in the Evaluation of New Macromolecular Cryoprotectants Post-thaw culture for an extended period, giving apoptosis time to set in, is now considered essential for honest assessment.
Logistics add another layer of complexity. Unlike conventional drugs that can sit in a warehouse, many cell therapies are patient-specific, time-sensitive, and must be maintained within a narrow temperature window from the moment cells leave the patient’s body until the finished product is infused back. This “vein-to-vein” supply chain requires strict chain-of-identity and chain-of-custody tracking, and any deviation in temperature, timing, or labeling can compromise patient safety.17West Kazakhstan Regenerative Medicine Journal. Cold Chain Digitalization in Regenerative Pharmacology: Ensuring Integrity, Traceability, and Quality of Biologics and Cell Therapies – A Narrative Review Digital cold-chain monitoring tools are increasingly used to synchronize these workflows, but the logistics remain one of the most underappreciated bottlenecks in the field.
Clinical Trial Design for Cell Therapies
Running a clinical trial for a cell therapy is not the same as running one for a pill. Dose-finding, for instance, is complicated by the fact that the “drug” is alive and expanding inside the patient. Standard dose-escalation frameworks can struggle with late-onset toxicities that show up well after the typical observation window, and they rarely account for manufacturing feasibility. Not every planned dose level may be achievable for every patient if the cell expansion process fails or produces too few cells. Newer statistical designs have been developed specifically for these situations, incorporating both late-onset toxicity monitoring and feasibility constraints to identify a maximum tolerated dose that is also actually manufacturable.18PubMed Central. Dose-finding designs for cell therapy cancer clinical trials evaluating drug-combinations
Safety monitoring during and after infusion is intense. CAR T-cell therapies in particular can trigger cytokine release syndrome and a neurotoxicity known as ICANS. Biomarker monitoring has become a major focus for early detection. Elevated pre-infusion fibrinogen and lactate dehydrogenase levels have been associated with ICANS in lymphoma patients.19PubMed Central. Immune effector cell–associated neurotoxicity syndrome after chimeric antigen receptor T-cell therapy for lymphoma: predictive biomarkers and clinical outcomes Cytokine monitoring after infusion is also informative: rising IL-6 levels have been linked to a roughly 76% increase in the odds of developing ICANS the following day, independent of how severe cytokine release syndrome is at that point.20Cytotherapy. Early Detection of CAR-T-Associated Neurotoxicity via Cytokine Monitoring in Serum Real-time monitoring like this is pushing clinical teams toward earlier intervention.
The Solid Tumor Problem
CAR T-cell therapy has shown remarkable efficacy in blood cancers. As of early 2023, six CAR T-cell therapies had been approved, all for B-cell malignancies or multiple myeloma.21PubMed Central. From bench to bedside: the history and progress of CAR T cell therapy Solid tumors have been far more resistant. A key reason is the tumor microenvironment: the tissue surrounding a solid tumor actively suppresses immune cells, causing CAR T cells to lose function and fail to infiltrate the tumor effectively.22PubMed Central. Tumor Microenvironment Immunosuppression: A Roadblock to CAR T-Cell Advancement in Solid Tumors On top of that, persistent antigen stimulation inside the tumor drives T-cell exhaustion, a state where the cells become increasingly dysfunctional over time.23PubMed Central. Improving CAR-T immunotherapy: Overcoming the challenges of T cell exhaustion Overcoming exhaustion and the hostile tumor environment is now one of the most active areas of cell therapy research.
Managing Process Changes Along the Way
Cell therapy manufacturing processes rarely stay the same from early clinical trials through commercialization. A developer might switch from one cell culture supplement to another because the original is no longer available, scale up from flasks to bioreactors, or move production to a new facility. Each change triggers a comparability exercise: the developer must demonstrate that the product made with the new process is equivalent in identity, purity, potency, and safety to the product made with the old one.24PubMed Central. The comparability tales: A phase-appropriate roadmap for CGT drug product development
These exercises are more demanding than they might appear. A comparability study for switching from autologous serum to human platelet lysate in a clinically used cell therapy required defining critical attributes, choosing appropriate test methods, and setting acceptance criteria through a formal risk-based approach.25PubMed. Good Manufacturing Practice-compliant change of raw material in the manufacturing process of a clinically used advanced therapy medicinal product-a comparability study The statistical methods themselves matter: choosing the wrong test, or confusing a paired analysis with an unpaired one, can lead to false conclusions about whether the product has actually changed. Regulators expect these comparability narratives to be grounded in the product’s mechanism of action, not just a checklist of standard assays.
Regulatory Pathways and Expedited Programs
Cell therapies often target serious or life-threatening diseases with few existing treatments, which makes them candidates for expedited regulatory programs. Both the FDA and the European Medicines Agency offer mechanisms to speed up development of products that address unmet medical needs, including Breakthrough Therapy and Regenerative Medicine Advanced Therapy designations in the United States and the PRIME scheme in Europe.26PubMed Central. Considering Global Development? Insights from Applications for FDA Breakthrough Therapy and EMA PRIME Designations These designations provide benefits like more frequent interactions with regulators and rolling review of applications, but they do not lower the evidentiary bar for safety or efficacy. Products that use gene-integrating vectors face additional requirements, including FDA guidance specifically addressing long-term follow-up for therapies with genome integration potential.27PubMed Central. Food and Drug Administration Guidance on Design of Clinical Trials for Gene Therapy Products with Potential for Genome Integration or Genome Editing and Associated Long-Term Follow-Up of Research Subjects Long-term follow-up of patients who received gene-modified stem cells for cerebral adrenoleukodystrophy showed durable polyclonal distribution of corrected cells with no adverse effects, illustrating the kind of monitoring data regulators want to see.28PubMed. Long-Term Follow-Up of Hematopoietic Stem-Cell Gene Therapy for Cerebral Adrenoleukodystrophy
Why Commercialization Keeps Stumbling
The science of cell therapy has advanced faster than the business model. Early cell-based companies in the 1990s managed to create products that worked clinically, but small margins combined with small initial patient populations made it impossible to generate returns on the large upfront investments required.29PubMed. Cell-based therapeutics from an economic perspective: primed for a commercial success or a research sinkhole? That pattern has not entirely gone away. A recent stakeholder analysis of cell therapy access in Europe identified donor-to-donor and batch-to-batch variability, fragmented manufacturing infrastructure, inconsistent national access pathways, and chronic underinvestment as persistent barriers to reaching patients.30PubMed Central. Why promising ATMPs fail to reach patients: a qualitative stakeholder-informed analysis of cell therapy access in Europe Several approved cell therapies in Europe have been withdrawn from the market not because they did not work, but because the economics did not.
Decentralized Manufacturing as a Potential Fix
One response to the logistical and cost burdens of centralized production is point-of-care manufacturing, where cells are produced at or near the hospital where the patient is treated. This eliminates much of the shipping and cold-chain complexity and could make therapies accessible in regions where centralized facilities do not exist.31PubMed Central. Promises and challenges of a decentralized CAR T-cell manufacturing model But decentralized manufacturing introduces its own challenges. Every participating site must follow the same protocol, use comparable assays, and maintain equivalent quality programs. Coordinating all of this under a single regulatory application is, in many ways, harder than producing everything in one place.32PubMed Central. Point-of-care cell therapy manufacturing; it’s not for everyone The field is still working out which products and which settings are best suited to each model.
Engineering CAR T Cells Directly Inside the Body
A more radical shift would eliminate the ex vivo manufacturing step altogether. In vivo cell engineering aims to reprogram a patient’s own immune cells without ever removing them from the body. Recent work has demonstrated that targeted lipid nanoparticles carrying messenger RNA can reprogram specific T-cell subsets directly in living animals, producing CAR T cells that control tumors in humanized mice and deplete B cells in monkeys.33PubMed. In vivo CAR T cell generation to treat cancer and autoimmune disease When paired with transposase technology for stable gene insertion, the same nanoparticle platform has generated durable antitumor responses without any cell collection or manufacturing step.34PubMed Central. T cell-specific non-viral DNA delivery and in vivo CAR-T generation using targeted lipid nanoparticles This approach is still in early stages, but if it works in humans, it would fundamentally change the development process by collapsing cell collection, engineering, expansion, and infusion into a single injection.
Informed Consent and the Ethics of Emerging Therapies
Cell and gene therapies pose distinctive challenges for informed consent. The treatments are complex, the risks are unfamiliar, and the long-term consequences are often unknown. Research on consent for genome-editing trials in sickle cell disease found that patients and community members want clear information about side effects, how the editing works, trial eligibility, and the impact on quality of life. Physicians tended to underestimate their patients’ genetic literacy, while patients demonstrated higher understanding than expected.35PubMed Central. The Meaning of Informed Consent: Genome Editing Clinical Trials for Sickle Cell Disease In sub-Saharan Africa, where sickle cell disease is most prevalent, the consent process faces additional layers of difficulty: language barriers, cultural norms around family decision-making, low health literacy, and stigma around genetic conditions all shape how information must be delivered.36BMC Medical Ethics. Informed consent for somatic gene editing for sickle cell disease in sub-Saharan Africa: a scoping review Designing consent processes that are both scientifically accurate and culturally appropriate is not a box-checking exercise. It requires sustained engagement with the communities most affected by these diseases, well before a trial begins enrolling.