The present study did not determine if the DNA was circular or integrated into the
chromosomes, as we intend to keep the sole surviving fish alive for further investigation.
Anderson et al. (1996b) showed that DNA was stable in an extrachromosomal
circular form up to at least 2 months after DNA immunisation in rainbow trout, and
this might also be the case after 2 years in glass catfish.
The present study provides data that is both positive and negative toward the
acceptance of commercial DNA vaccines for fish. That the site of expression did not
change during 2 years is a finding in favour of the acceptance of DNA vaccines, since
di#usion of the DNA throughout the body increases the chance of creating transgenic
fish. Further studies on expression in reproductive organs and o#spring are needed.
However, that expression after DNA injection can last for at least 2 years is a very
important negative finding for the acceptance of commercial DNA vaccines, since an
ideal vaccine would only be expressed for a brief period of time. Plasmid pCEP4-luc is
not representative of all DNA vaccines. For example, Heppel et al. (1998) found that
coinjection of plasmids expressing luciferase and the VHS-G gene significantly reduced
the period of luciferase expression. This di#erence might be caused by antibodydependent
cell mediated cytotoxicity, which VHS-G, as a membrane protein, is likely to
induce and luciferase, as an intracellular protein, is not.
The possibility of long-term expression should be analysed for each combination of
fish species and DNA vaccine before consideration for commercial usage. In addition,
research should be directed towards general approaches in creating DNA vaccines
with a limited expression period, possibly by adding sequences to the plasmid that
decrease its stability (Gordenin & Resnick, 1998) or by co-expression of sequences that
limit the lifespan of the expressing cells (Leitner et al., 2000). The glass catfish system
would be a convenient system to analyse such constructs since di#erent DNA
constructs can be injected at di#erent sites in a single fish and expression monitored
over time. In addition to luciferase, the gene for green fluorescent protein may also be
useful (Fernandez-Alonso et al., 1999). These in vivo systems would avoid misinterpretation
caused by di#erences in response by individual fish and significantly reduce the
number of fish required compared to the analysis of isolated tissues.