wall as the balloon contacts the arterial wall coincident with balloon inflation. Experiments with 
radiolabeled DNA established that 97% of DNA applied in aqueous solution to the hydrogel-coated balloon 
was still present on the balloon after drying of the gel. Autoradiograms of the arterial wall demonstrated 
that inflation of the hydrogel balloon results in DNA uptake which is distributed across the full thickness 
of the arterial wall. DNA was shown to penetrate the intact internal elastic lamina and was distributed 
intracellularly as well as extracellularly. The same approach has more recently been used to directly 
delivery anti- thrombotic drugs to the arterial wall 40,41 
Gene transfer in this study was achieved without the use of transfection vehicles such as 
liposomes or viral vectors. The feasibility of arterial gene transfer with DNA alone has been previously 
demonstrated by Chapman et al. 42 They observed no differences in luciferase expression when 
surgically exposed canine femoral arteries were transfected with a 1:1 DNA-liposome mixture versus DNA 
alone. Transfection of skeletal muscle and myocardium have been routinely accomplished by direct 
injection of DNA without facilitating transfection vehicles 43-48. The use 0 f DNA alone clearly simplifies 
the transfection protocol. 
It is noteworthy that despite elimination of accessory transfection vehicles, both the frequency 
of successful transfection and the magnitude of reporter gene expression achieved were superior to that 
previously reported from our laboratory 49 and comparable to the results achieved by others 42,50 us i n g 
alternative delivery schemes. The success rate of transfection in our rabbit model as measured by 
expression of the luciferase transgene was 100% (37 of 37 artery segments), even in those cases in which 
the inflation time was reduced to one minute. The duration of inflation within a range from 1 to 30 minutes 
did not have a significant impact on transfection efficiency, a feature which would be expected to facilitate 
human arterial, particulary coronary, gene transfer. 
While the levels and extent of reporter gene expression achieved in this case with the hydrogel-coated 
balloon are not equivalent to those recendy reported with viral vectors 51-54^ resu lts here indicate that 
the hydrogel catheter may represent an attractive alternative to other delivery devices used to transfer naked 
DNA or DNA-liposome mixtures. 
I. D. Arterial Gene Transfer of cDNA Encoding for Secreted Protein may result in 
Meaningful Biological Outcomes Despite Low Transfection Efficiency. 
In experiments which have relied exclusively on the use of nonsecreted gene products, examination by 
histochemical staining, in situ hybridization, and/or polymerase chain reaction has suggested that the 
transfection efficiency of direct gene transfer to vascular smooth muscle cells within the arterial wall was 
considerably less than 1% and might therefore preclude a meaningful biological response. In contrast, 
genes encoding for a secreted protein may overcome the handicap of inefficient transfection by a paracrine 
effect, secreting adequate protein to achieve local levels that may be physiologically meaningful. Nabel et 
al 55 demonstrated that despite similarly low efficiencies, cell surface protein expression resulting from 
percutaneous transfection of vascular smooth muscle cells with the histocompatibility gene HLA-B7 may 
be adequate to induce a biological response, namely, focal vasculitis. Necropsy evidence of a 
pathobiological response following arterial gene transfer was reported by the same group in the case of 
transgenes encoding for the secreted proteins PDGF-B 9 and FGF-1 56; i n the former study, only 0.1 to 
1 % of cells in the artery segment were estimated to contain plasmid DNA by PCR approximation. 
To more specifically determine the relation between a secreted gene product and transfection efficiency 
after in vivo arterial gene transfer, we devised in vitro 5 and in vivo 6 models to serially monitor 
expression of a gene encoding for a secreted protein. In vivo analyses were performed using the central 
artery of the rabbit ear. Liposome-mediated transfection of plasmid DNA containing the gene for human 
growth hormone (hGH) was successfully performed in 18 of 23 arteries. Serum hGH levels measured 5 
days after transfection ranged from 0.1 to 3.8 ng/mL (mean, 0.97 ng/mL); in contrast, serum drawn from 
the control arteries demonstrated no evidence of hGH production. Serial measurement of hGH from 
transfected arteries demonstrated maximum hGH secretion 5 days after transfection and no detectable 
hormone after 20 days. Despite these levels of secreted gene product documented in vivo, 
immunohistochemical staining of sections taken from the rabbit ear artery at necropsy disclosed evidence 
of successful transfection in <0.1% of cells in the transfected segment. Thus, low-efficiency transfection 
with a gene encoding for a secreted protein may achieve therapeutic effects not realized by transfection 
with genes encoding for proteins which remain intracellular. 
I. E. Pre-clinical Animal Studies have Established that Percutaneous Arterial Gene 
Transfer may be Utilized to Successfully Accomplish Therapeutic Angiogenesis. 
Recombinant DNA Research, Volume 20 
[321] 
