Automating chimeric antigen receptor T-cell manufacturing across distinct platforms delivers similar cell quality, phenotype and function

Authors:
Lee EH, Safford M, Staab T, Park MS, Ryan M, Pleitez D, Sargent A
In:
Source: Cytotherapy
Publication Date: (2026)
Issue: (12): 102978
Research Area:
Cancer Research/Cell Biology
Immunotherapy / Hematology
Cells used in publication:
T cell, human stim.
Species: human
Tissue Origin: blood
Culture Media:
Experiment
TheraPEAK™ T-VIVO™ medium was used on two different automated manufacturing platforms (CliniMACS Prodigy and Sefia) to isolate platform effects from media effects.
The cells were activated with  TransAct® and cultured TheraPEAK™ T-VIVO™ medium supplemented with IL-7 and IL-15.  TheraPEAK™ T-VIVO™ supported a complete 9-day automated CD19 CAR-T manufacturing process on both the CliniMACS Prodigy and Sefia platforms. The resulting CAR-T products displayed comparable expansion, phenotype, metabolism, transduction efficiency, and anti-tumor function across both manufacturing systems, with 95% CD3+ T-cell purity after enrichment and 
CAR transduction efficiencies  of approximately 62-78%
Abstract

Background aims: Manufacturing chimeric antigen receptor (CAR) T cell therapies is complex and costly, and automation holds tremendous promise to improve access to these life-saving treatments.

Methods: Here, we compared two end-to-end automated CAR T-cell manufacturing platforms, the CliniMACS Prodigy and the Sefia system, using identical reagents, stimulation conditions, culture duration, perfusion rate and donor starting material to assess how platform design affects critical product attributes.

Results: With identical CD4 and CD8 magnetic isolation reagents, both systems produced similar T-cell yields, with exceptionally high purity and comparable phenotype. During culture, total cell yield differed between platforms, reflecting differences in culture volume; however, expansion rate, population doubling time, transduction efficiency and CAR T-cell yield were not significantly different. The resulting CAR T cells also showed comparable phenotype, exhaustion profiles and in vitro tumor-killing activity. These similar outcomes may be explained by shared media perfusion rates and convergent metabolic profiles, because cell-specific glucose consumption and cell-specific lactate production were similar by the end of culture.

Conclusions: Collectively, these findings show that distinct automated platforms can generate CAR T-cell products with remarkably similar quality and function when process conditions are aligned. More broadly, they highlight the potential of automation and in-line analytics to standardize CAR T cell manufacturing and improve product consistency from cell collection to patient infusion.