Predictive models for transient protein expression in tobacco (Nicotiana tabacum L.) can optimize process time, yield, and downstream costs.

The transient expression of recombinant biopharmaceutical proteins in plants can suffer inter-batch variation, which is considered a major drawback under the strict regulatory demands imposed by current good manufacturing practice (cGMP). However, we have achieved transient expression of the monoclonal antibody 2G12 and the fluorescent marker protein DsRed in tobacco leaves with ∼ 15% intra-batch coefficients of variation, which is within the range reported for transgenic plants. We developed models for the transient expression of both proteins that predicted quantitative expression levels based on five parameters: The OD(600 nm) of Agrobacterium tumefaciens (from 0.13 to 2.00), post-inoculation incubation temperature (15-30°C), plant age (harvest at 40 or 47 days after seeding), leaf age, and position within the leaf. The expression models were combined with a model of plant biomass distribution and extraction, generating a yield model for each target protein that could predict the amount of protein in specific leaf parts, individual leaves, groups of leaves, and whole plants. When the yield model was combined with a cost function for the production process, we were able to perform calculations to optimize process time, yield, or downstream costs. We illustrate this procedure by transferring the cost function from a production process using transgenic plants to a hypothetical process for the transient expression of 2G12. Our models allow the economic evaluation of new plant-based production processes and provide greater insight into the parameters that affect transient protein expression in plants.