We describe a design of experiments (DoE) response surface modeling strategy to optimize the concentration of basal variables underpinning polyethylenimine (PEI) mediated transfection of different CHO-K1 derived parental cell populations in a chemically defined medium, specifically the relative concentration of linear 25 kD PEI, host CHO cells and plasmid DNA. Using recombinant secreted alkaline phosphatase (SEAP) reporter activity as the modeled response, a discrete simple maximum was predicted for each CHO host cell population. Differences between the modeled optima derived from host cell specific differences in PEI cytotoxicity, such that the PEI:cell interaction effectively limited PEI-DNA polyplex load at a relatively constant PEI:DNA ratio. However, across the three CHO host cell populations, SEAP reporter production was not proportional to plasmid DNA input at the host cell specific predicted basal variable optima. A 10-fold variation in SEAP reporter output per mass of plasmid DNA delivered was observed. To determine the cellular basis of this difference in transient productivity, host CHO cells were transfected with fluorescently labeled polyplexes followed by flow cytometric analysis. Each CHO host cell population exhibited a distinct functional phenotype, varying in the extent of PEI-DNA polyplex binding to the cell surface and degree of polyplex internalization. SEAP production was directly proportional to the level of polyplex internalization and heparan sulfate proteoglycan level. Taken together, these data show that choice of host CHO cell line is a critical parameter, which should rationally precede cell line specific transient production platform design using DoE methodology.
Copyright © 2011 American Institute of Chemical Engineers (AIChE).