Size exclusion PEGylation reaction chromatography was investigated using a model
developed by Fee (2005). Column dispersion was neglected and the PEGylation
reaction was modelled as second order. The model allowed up to four PEG groups to
be attached to a protein and accounted for succinic acid hydrolysis from activated
PEG. The model was adapted to simulate a-lactalbumin PEGylation and succinic
acid hydrolysis from activated PEG in a batch stirred tank so rate parameters from
stirred tank kinetic experiments could be obtained and the model verified. The model
was solved using finite differences and simulations run in Matlab. The effect of
reaction parameters such as timing, length and concentration of PEG and protein
injection, reaction rates, and model resolution on model simulation results was
explored.
In the size exclusion PEGylation simulations it was found that increasing protein
concentration increased MonoPEG concentrations and increased the ratio of
MonoPEG to starting protein feed concentration. Increasing PEG pulse length and
starting PEG concentration initially increased MonoPEG concentration and product
ratio until all protein had been PEGylated at which point MonoPEG concentration the
product ratio levelled out. Increasing PEG hydrolysis rates did not affect the amount
of MonoPEG produced but reduced the activated PEG concentration and increased
succinic acid concentration. Optimal conditions for producing MonoPEG were found
to be equal concentrations of PEG and protein, with the PEG injection length twice as
long as the protein injection, and the PEG injection done immediately after the protein
injection.
