A modeling study of monoclonal antibody manufacturing suggests that the production facility, rather than the bioreactor, may be the largest source of energy demand. Heating, ventilation and air conditioning accounted for 46% of modeled energy use, while lighting contributed another 21%. Overall, facility infrastructure made up about 85% of demand, according to findings published in New Biotechnology.
Benjamin Smith of Amgen and colleagues from the NIIMBL Process Intensification Sustainability Workgroup compared three production approaches: stainless-steel Fed-batch, single-use fed-batch and end-to-end single-use continuous processing. Their models assessed energy, water, materials and greenhouse-gas emissions at hypothetical facilities producing about 120 kilograms of antibody annually.
The two fed-batch scenarios used 2,000-liter bioreactors. The continuous process used a 500-liter vessel over a longer production run and achieved comparable annual output with a smaller modeled facility: 12,387 cubic meters, compared with 14,089 cubic meters for single-use fed-batch and 14,937 for stainless-steel fed-batch.
The continuous scenario also had the lowest energy use per kilogram of antibody, at 13,239 kilowatt-hours. The figures were 17,017 kilowatt-hours for single-use fed-batch and 19,120 for stainless-steel fed-batch. The researchers concluded that, when output is equivalent, facility footprint is the main determinant of energy intensity.
The comparison did not show a single process to be best on every sustainability measure. Stainless-steel manufacturing used considerably more water for cleaning but required fewer disposable materials. Single-use fed-batch had the highest consumables intensity, while continuous processing used more culture medium and some buffers.
The team also modeled an enhanced continuous process combining a 50% increase in antibody titer with two flow-through polishing steps, more concentrated medium and reuse of media and buffer bags. Against baseline continuous processing, the changes reduced water use per kilogram by 26% and consumables intensity by 58%. Modeled carbon dioxide-equivalent emissions were 66% lower than in baseline continuous processing and 78% lower than in the stainless-steel fed-batch scenario.
Facility changes could also make a substantial difference. Measures such as adjusting cleanroom classifications where appropriate, improving airflow and lighting, increasing equipment density and optimizing utilities reduced modeled facility energy consumption and emissions by up to about 35%. The models also indicated that temperate locations generally needed less heating or dehumidification than colder or more humid sites.
The analysis covered hypothetical plants with one bioreactor, rather than larger commercial facilities with multiple production trains, and excluded some lifecycle impacts, including raw-material production and transportation. The workgroup plans to expand the research with lifecycle assessment and test-bed data.