The success and validity of gene therapy and DNA vaccination in in vivo experiments and human clinical trials depend on the ability to produce large amounts of plasmid DNA according to defined specifications. A new method is described for the purification of a cystic fibrosis plasmid vector (pCF1-CFTR) of clinical grade, which includes an ammonium sulfate precipitation followed by hydrophobic interaction chromatography (HIC) using a Sepharose gel derivatized with 1,4-butanediol-diglycidylether. The use of HIC took advantage of the more hydrophobic character of single-stranded nucleic acid impurities as compared with double-stranded plasmid DNA. RNA, denatured genomic and plasmid DNAs, with large stretches of single strands, and lipopolysaccharides (LPS) that are more hydrophobic than supercoiled plasmid, were retained and separated from nonbinding plasmid DNA in a 14-cm HIC column. Anion-exchange HPLC analysis proved that >70% of the loaded plasmid was recovered after HIC. RNA and denatured plasmid in the final plasmid preparation were undetectable by agarose electrophoresis. Other impurities, such as host genomic DNA and LPS, were reduced to residual values with the HIC column (<6 ng/microg pDNA and 0.048 EU/microg pDNA, respectively). The total reduction in LPS load in the combined ammonium acetate precipitation and HIC was 400,000-fold. Host proteins were not detected in the final preparation by bicinchoninic acid (BCA) assay and sodium dodecylsulfate-polyacrylamide gel electrophoresis (SDS-PAGE) with silver staining. Plasmid identity was confirmed by restriction analysis and biological activity by transformation experiments. The process presented constitutes an advance over existing methodologies, is scaleable, and meets quality standards because it does not require the use of additives that usually pose a challenge to validation and raise regulatory concerns.

Two important issues in the downstream processing of plasmids for gene therapy are the stability of plasmids in the process streams, and the presence of contaminating host RNA. Results with a 4.8-kb plasmid harbored in a non-nuclease-deficient strain of Escherichia coli show that, in spite of the harsh conditions during alkaline lysis, a fraction of endogenous nucleases remains active, degrading both RNA and genomic and plasmid DNA. Although it is possible to minimize plasmid degradation by decreasing temperature and reducing processing times, the presence of endogenous nucleases can be used advantageously to purify the plasmid streams. The kinetics of nucleic acid degradation showed that, by controlling the incubation at 37 degrees C, it was possible to degrade RNA selectively, while maintaining plasmid integrity. A reduction of 40% in RNA content was obtained, corresponding to a 1.5-fold increase in plasmid purity using high-performance liquid chromatography (HPLC). This strategy is simple and straightforward, and the increase in processing time and the associated plasmid loss (9%) are fully justified by the purity increase. Furthermore, the use of endogenous RNase activity is clearly advantageous over alternative procedures, such as the addition of external RNase, in terms of cost, validation, and compliance with guidelines from regulatory agencies.