Plasmid DNA malaria vaccine: the potential for genomic integration after intramuscular injection

Plasmid-based (naked DNA) genetic vaccines are now entering clinical trials to test their safety and efficacy in healthy human volunteers. A safety concern unique to this new class of vaccines is the potential risk of deleterious integration into host cell genomic DNA following direct intramuscular injection. To address this issue experimentally, a preclinical safety study was conducted in mice to determine the structural nature of plasmid DNA sequences persisting in total muscle DNA at both 30 and 60 days following a single intramuscular injection of a plasmid expressing the Plasmodium falciparum circumsporozoite protein. In a protocol described for the first time, total DNA was extracted from muscle tissue and was subsequently linearized with a restriction endonuclease to enable agarose gel size fractionation of all extrachromosomal plasmid DNAs from high molecular weight mouse genomic DNA. Using PCR assays to quantitate plasmid-specific sequences, it was found that the amount of plasmid DNA persisting in muscle tissue varied but averaged about 10 fg per microgram of genomic DNA (in the range of 1500 copies per 150,000 genomes). In two of four separate experimental injections of mouse muscle, PCR assays of genomic DNA fractions indicated that agarose gel purification removed plasmid DNA down to a level of < or =3 copies per 150,000 mouse genomes. In the two other experimental samples, 3-30 copies of plasmid DNA remained associated with purified genomic DNA. The time following injection (i.e., 30 or 60 days) was not a factor in the number of copies of plasmid associating with genomic DNA and it was not possible to conclude if such sequences were covalently linked to genomic DNA or simply adventitiously associated with the genomic DNA. However, if an assumption is made that the highest level plasmid DNA found associated with genomic DNA (i.e., 30 copies) represented covalently integrated plasmid inserts and that each insert resulted in a mutational event, the calculated rate of mutation would be 3000 times less than the spontaneous mutation rate for mammalian genomes. This level of integration, if it should occur, was not considered to pose a significant safety concern.