Search

Although effective gene therapy vectors have been developed for organ systems such as the liver, an effective delivery vector to the pancreas in vivo has remained elusive. Of the currently available viral vectors, adenovirus and adeno-associated virus (AAV) are two of the most efficient at transducing nondividing cells. We have constructed recombinant adenovirus (AdVLacZ), adeno-associated virus serotype 2 (AAV2LacZ), and pseudotyped adeno-associated virus serotype 5 and 8 (AAV5LacZ, AAV8LacZ) carrying the LacZ reporter, and compared the transduction efficiency of these four vectors in the pancreas of mice in vivo. We showed that adenovirus, AAV2, and AAV8 are capable of transducing the pancreas in vivo, but with different expression kinetics, efficiencies of transduction, and persistence. AdVLacZ-transduced pancreas exhibited maximum LacZ expression at 1 week postdelivery, with greater than 90% of expression lost at 4 weeks. AAV2LacZ-transduced pancreas displayed peak LacZ levels at 4 weeks postdelivery, with no significant decrease in expression for up to 8 weeks. AAV8LacZ was at least 10-fold more efficient than AAV2LacZ in transducing the pancreas in vivo, with significant levels of expression detectable at 1 week, whereas AAV5LacZ did not result in any detectable transgene expression at all tested time points. All three vectors primarily transduced pancreatic acinar cell types, with limited transduction of pancreatic endocrine cells. AdVLacZ elicited a significant leukocyte infiltration early after delivery into the pancreas, whereas none of the AAV vectors elicited a significant leukocyte response. None of the tested vectors caused significant changes in serum amylase or blood glucose levels, suggesting that they do not significantly alter pancreatic function. These vectors will be useful for studying novel gene delivery based treatments in animal models for diabetes and other pancreatic disorders.

While naked DNA gene transfer in vivo usually results in transient gene expression, in some cases long-term transgene expression can be achieved. Here we demonstrate that cis-acting DNA elements flanking the transgene expression cassette and components in the plasmid backbone can significantly influence expression levels from nonviral vectors. To demonstrate this, we administered our most robust human coagulation factor IX (hFIX) expression cassette placed in two different plasmid backbones, into the livers of mice, by hydrodynamic transfection. We found that placing the expression cassette within a minimal plasmid vector pHM5, a modified version of pUC19, resulted in 10 times higher serum hFIX expression levels (up to 20,000 ng/ml, 400% of normal hFIX serum levels), compared to a pBluescript backbone. To optimally increase expression levels from a nonviral vector, we added matrix attachment regions (MARs) as cis-acting DNA elements flanking the hFIX expression cassette. We detected five fold higher hFIX expression levels in vivo for up to 1-year posttransfection from a vector that contained the chicken MAR from the lysozyme locus. Together, the present work demonstrates that in addition to the transgene expression cassette, cis-acting DNA elements within and outside of the plasmid backbone need to be evaluated to achieve optimal expression levels in a nonviral gene therapy approach.