Reviewed by Sarah Alter, Ph.D. — Scientific Affairs, OrganaBio. 15 years of immunology research spanning autoimmunity, cancer, and infectious disease. University of Miami Miller School of Medicine. Registered Patent Agent.
The Default That Costs Programs Their Translational Value
Healthy donor PBMCs are the default in cell therapy research for one reason: they’re available, characterized, reproducible, and cheap per lot. That’s a procurement argument, not a scientific one. The question that matters is whether the biology of your experimental model matches the biology your therapy will encounter in patients.
For oncology programs — CAR-T targeting solid tumors, NK cell therapies, checkpoint combination approaches — healthy donor starting material is often appropriate. The target is the tumor, the immune environment is the obstacle, and the question is whether your construct functions under T cell exhaustion or immunosuppression conditions that can be modeled independently. Healthy donor cells are the right chassis for that.
For autoimmune and inflammatory disease programs, the argument inverts. The disease immune environment is not the obstacle — it’s the therapeutic target. Your therapy must function inside a Th17-dominant synovial joint, a Th2-skewed colonic mucosa, or a CNS infiltrated by autoreactive T cells. Those environments cannot be added to healthy donor cells as conditions. They have to come from donors who carry them.
This piece maps the decision across program types. Not every program needs disease-state material. But the programs that do need it — and switch to it late — consistently report the same outcome: unexpected behavior in Phase I that clean preclinical data didn’t predict.
When Healthy Donor PBMCs Are the Right Call
Healthy donor PBMCs remain the right starting material when:
The therapeutic target is a defined molecule, not a disease-specific immune environment. CD19-targeted CAR-T for B-ALL, BCMA-targeted constructs for myeloma, HER2-targeted NK cells — the construct targets a specific antigen, and efficacy is determined by binding, signaling, cytotoxicity, and persistence. The disease immune suppression (T cell exhaustion, checkpoint upregulation, Treg infiltration) can be modeled by stimulation conditions, and the manufacturing process is the primary variable. Healthy donor cells provide the clean chassis for construct optimization.
The research question is about the cell product, not the disease environment. Process development, manufacturing scalability, cryopreservation protocol optimization, genetic modification efficiency, expansion kinetics — these are questions about what happens to the cells during manufacturing, not what happens when they encounter patients. Healthy donors are appropriate controls for manufacturing characterization.
The therapy is intended for use in healthy or functionally normal immune contexts. Adoptive T cell therapies for infectious disease, prophylactic NK cell products, some tolerogenic constructs for transplant — where the disease hasn’t fundamentally altered the immune landscape the therapy encounters.
Early discovery where the hypothesis is being defined, not validated. Initial target identification, construct screening across multiple candidates, preliminary cytotoxicity assays — healthy donor cells provide a reproducible, variable-controlled model for early hypothesis testing before the experimental design commits to disease-state material.
When Disease-State Material Becomes Required
The switch from optional to required happens when your experimental question depends on cellular properties that only exist in disease-state donors.
The therapy targets an immune mechanism that is disease-specifically dysregulated. Treg therapies for inflammatory bowel disease, tolerance-inducing constructs for celiac, Th17-suppressing approaches for RA or MS — these therapies must demonstrate activity against the specific dysregulated immune cell populations they’ll encounter in patients. That demonstration requires cells from patients.
Specifically:
- Ulcerative colitis: Th2/IL-13 skewing, ILC2 expansion, Treg dysfunction in the mucosal context. Healthy donor controls don’t carry any of these in the concentrations that define UC pathology.
- Crohn’s disease: Th1/Th17 dominance, M1-polarized macrophages, NOD2 pathway involvement, granuloma biology. Distinct from UC, not interchangeable as an IBD model.
- Celiac disease: HLA-DQ2/DQ8-restricted gluten-reactive T cells, IEL expansion, anti-tTG2 B cells — none of which exist in general healthy donor populations.
- Rheumatoid arthritis: Shared epitope HLA, anti-CCP immune repertoires, TNF-primed monocytes, unstable Tregs under inflammatory conditions.
- Multiple sclerosis: HLA-DRB1*15:01-enriched donor pools, CNS-homing T cell phenotypes, B cell pro-inflammatory cytokine profiles, disease subtype-specific immune signatures.
Manufacturing will eventually use patient-derived cells. For autologous programs, or allogeneic programs where the manufacturing process must be validated against patient-derived starting material, developing with only healthy donor cells introduces a comparability gap that becomes a regulatory and scientific problem. The earlier you incorporate disease-state starting material into process development, the earlier you characterize how patient cell biology interacts with your manufacturing conditions.
Potency assays need to be calibrated against disease-relevant biology. The FDA and EMA expect potency assays to measure the activity relevant to the mechanism of action. A Treg potency assay for an autoimmune therapy that measures suppression of healthy donor effector T cells in a non-inflammatory in vitro system is not measuring what the therapy needs to do. Disease-state donor cells provide the biologically relevant assay components.
Alloreactivity data for allogeneic programs needs to reflect the HLA architecture of patients, not general population donors. Many autoimmune diseases — RA, MS, celiac, T1D — are enriched for specific HLA alleles in the patient population. Alloreactivity testing against a general healthy donor pool will not represent the actual HLA landscape your allogeneic therapy will encounter.
The Mixed Model: When to Run Both in Parallel
The lowest-risk approach for programs targeting autoimmune or inflammatory indications is to run healthy donor and disease-state material in parallel from the earliest stages where the experimental design allows it.
This serves three functions:
First, it establishes the delta between healthy and disease-state biology for your specific experimental system. That delta is data — it tells you whether the disease creates a meaningful challenge for your construct, and where the challenge is located.
Second, it builds the comparison dataset for regulatory submissions. When reviewers ask whether your preclinical data was generated in a disease-relevant model, parallel-run data with both healthy donor controls and disease-state donors is the strongest possible answer.
Third, it catches surprises early. Programs that exclusively used healthy donor material through IND submission and then discovered the comparison gap during first-in-human data collection had the worst outcomes. The surprise was expensive and late.
OrganaBio can provide matched healthy donor controls from the same collection period as disease-state material, supporting direct within-experiment comparisons that control for processing and collection variables.
Practical Decision Criteria
Before specifying starting material for your next experimental run, apply these four questions:
One: Does your therapeutic mechanism depend on a cellular property that is specifically dysregulated in the disease? If the answer involves Treg instability, exhaustion phenotypes, disease-specific HLA presentation, or pathological cytokine environments — disease-state material.
Two: Will your therapy eventually be manufactured from or tested against patient-derived cells? If yes, the earlier you introduce disease-state material into your development program, the fewer surprises.
Three: Does your potency assay need to reflect a disease-relevant suppressive or effector challenge? If yes, healthy donor calibration is not sufficient.
Four: Are you targeting an indication with strong HLA associations — celiac (DQ2/DQ8), RA (DR4 shared epitope), MS (DRB1*15:01), T1D (DQ2/DQ8/DR4)? If yes, alloreactivity testing and HLA-matched donor selection require disease-state donor pools with confirmed allele frequencies.
If any of these answers is yes, contact OrganaBio to discuss disease-state donor selection. Our scientific team can advise on indication-specific donor selection criteria, matched healthy donor controls, and integration into your existing research workflow.
