Bioprocess development for plasmid-based vaccine production

Plasmid DNA (pDNA) vaccine is a promising vaccine technology, with better safety profile, more economical production and transport logistics, than conventional viral vaccines. Most importantly, pDNA vaccines elicit different immune responses including antibody-mediated, CD4 T-cell-mediated and CD8 T-cell-mediated immune responses, for defending against viral infections and cancer. The increasing number of preclinical and clinical trials on plasmid vaccines has triggered the need to make more in less time. Recent developments in the production of plasmid therapeutics involve the establishment of innovative and cost effective methods as well as simplified operations. This dissertation reports fundamental studies essential to the development of a rapid economically-viable plasmid production system which is cGMP-compatible. Optimisation of upstream bacterial fermentation and continuous downstream purification of the plasmid vaccine fraction are the main aspects considered in the project of this dissertation. Process variables required to improve the volumetric and specific yields of a model plasmid-based measles vaccine (pcDNA3F) harboured in E. coli DH5α were investigated. A cGMP-compatible method offering the capacity to continuously produce homogeneous supercoiled pDNA from clarified bacterial lysate using a monolithic adsorbent was developed. The method involved optimisation of the adsorbent characteristics, ligand functionalisation and chromatographic process conditions. The feasibility of using free metal ions to preferentially precipitate endotoxins (LPS) from a clarified plasmid DNAcontaining bacterial lysate was investigated. Screening of various free metal ions for effective endotoxin removal and optimisation of process conditions, such as pH, ion concentration, temperature and incubation time, using central composite design experiments were performed. The potential and advantages of using Zn2+-induced LPS aggregation as a secondary pDNA purification method was validated by studying the interaction of Zn2+ with LPS and pDNA. A comparative economic analysis on the basis of vaccine cost per dose for influenza vaccine produced via pDNA vaccine technology and fertilised egg-based technology was also studied. Experimental results from growth medium optimisation in 500 mL culture showed a maximum volumetric yield of 13.65 mg/L, twice the amount generated using a standard medium (PDM). Fed-batch fermentation in combination with exponential glycerol feeding strategy resulted in a significant increase of 110 mg/L pcDNA3F volumetric yield and a specific yield of 14 mg/g. In addition, growth pH variation (6 to 8.5) and temperature fluctuation (35 oC to 45 oC) also resulted in improved plasmid yield. Chromatographic purification of pDNA using a triethylamine-activated conical monolithic adsorbent resulted in preferential pcDNA3F adsorption with optimum resolution achieved under the conditions of 400 nm pore size of monolith, 0.7 M NaCl (pH 6) of binding buffer and 3 % B/min of gradient elution up to 1 M NaCl. Plasmid volumetric yield and recovery of ~3g/L and ~90% were obtained. Contaminant levels recorded were protein (0.01 mg/L), LPS (0.12 EU/mg) with no detectable gDNA and RNA. Results from endotoxin removal and analysis showed that ZnSO4 displayed the highest endotoxin removal efficiency (~91%) and plasmid recovery (~100%). It was found that selective endotoxin precipitation (< 0.05 EU/Mg) could effectively be carried out during neutralisation in alkaline cell lysis at a pH condition similar to that of clarified cell lysate, a low ZnSO4 concentration (0.5 M), a minimum incubation time (30 min) and a temperature of 15 oC. Apparently, the lipopolysaccharide (LPS) showed a decreased aggregate size at the start of the ZnSO4 addition before increasing gradually. Results from the LPS aggregation analysis drew a hypothesis that cationic close range encounter and interaction with LPS monomers may contribute to LPS self-aggregation whilst bridging of LPS monomers may increase the LPS aggregate size to a greater extent compared to that of self-aggregation. Specifically, addition of Zn2+ resulted in the largest number of LPS particles per aggregate and the value of aggregation constant (Km) for LPS-Zn2+ was substantially low (0.28 M) and considerably large (>2 M) for pDNA-Zn2+, indicating its preferential ability to remove LPS from pDNA-containing solutions. The economic studies suggested that pDNA-based influenza vaccine production was highly dependent on the selling price and production volume. A similar cost per dose of about $2 was calculated although most of the manufacturing costs for plasmid DNA vaccine were lower than inactivated virus vaccine. This dissertation has developed a simple bioprocess framework to successfully improve production specification of plasmid vaccines using pcDNA3F as a model. The method offers ease of plasmid DNA purification due to reduced bulk impurities, cost-efficiency and most importantly high endotoxin removal (> 80%) and plasmid recovery (> 90% ). The technology will have a great impact on overall plasmid production and in particular on the development of axial flow monolithic purification in combination with selective endotoxin precipitation