IV. Overall Assessment of Risks vs. Benefits 
A. Potential Efficacy of Gene therapy 
Critical to an assessment of the potential efficacy of this procedure is an estimate of the 
expected level of genetic reconstitution. We demonstrated in preliminary studies of cultured 
human hepatocytes, retrovirus mediated gene transfer into a large proportion of cells 
(Southern blot measured 0.1 to 1.0 proviral copies/cell; Ref. 19). The retroviral vectors 
express levels of recombinant LDL receptor protein that exceeds normal endogenous levels by at 
least 5 fold. The net level of LDL receptor function in the transduced population of FH 
hepatocytes should be between 50% to 500% of the level of endogenous receptor assuming 
efficiencies of gene transfer described above. We propose to transplant approximately 5 x 10 9 
hepatocytes or 2% of the total number of hepatocytes that are in an intact liver (approximately 
2.5 x 10 11 hepatocytes). We, therefore, estimate that the level of genetic reconstitution 
should be between 1 and 10% of endogenous LDL receptor function in FH recipients. 
Several important questions emerge regarding the feasibility and efficacy of this form of 
therapy. The first question relates to the feasibility of obtaining this level of genetic 
reconstitution for prolonged periods of time using ex vivo gene therapy. This assumes that 1) 
the infused hepatocytes efficiently engraft and persist, and 2) the hepatocytes continue to 
express the recombinant gene and retain the differentiated functions necessary to affect 
improved hypercholesterolemia. Our experiments in the WHHL rabbit, described in section 
II. 6.1 and in Appendix G, indicate that the hypothesis and assumptions noted above are valid and 
that the human experiments are feasible (18). Using a similar protocol to that proposed for the 
human experiments, we have transplanted 2 x 10 8 autologous, genetically-corrected 
hepatocytes (approximately 2% of the total hepatocytes in the liver of a rabbit) into 3 kg WHHL 
rabbits and have achieved genetic reconstitution in liver, as measured by quantitative RNase 
protection analysis, equal to 2-4% of normal endogenous RNA levels. The level of recombinant 
derived RNA in liver did not diminish for a period of 6 months following the transplant, the end 
point of the experiment. The preclinical studies in the WHHL rabbit indicate that the 
genetically corrected hepatocytes engraft and continue to express the recombinant gene for long 
periods of time. 
The other critical question relates to the expected therapeutic efficacy of 2-4% hepatic 
LDL receptor reconstitution. Several lines of evidence suggest that this level of genetic 
reconstitution may lead to substantial improvements in hypercholesterolemia. Indirect 
evidence to support this hypothesis in man was provided by studies of Sprecher et al., who 
characterized a population of FH homozygotes with respect to serum cholesterol and residual 
LDL receptor activity (41). They found an inverse correlation between serum cholesterol and 
residual LDL receptor activity in cultured fibroblasts. More direct support for this hypothesis 
has been provided in a series of hepatocyte transplantation experiments in the WHHL rabbit. 
Two groups have demonstrated substantial declines in serum cholesterol (25-40% decrease) in 
WHHL rabbits transplanted with limited numbers of hepatocytes (2-5% of the total number of 
hepatocytes in a liver) derived from an LDL receptor expressing allogeneic donor (17, 89). 
We have shown similar improvements in serum cholesterol (decreases of 25-40%) in WHHL 
rabbits transplanted with allogeneic or autologous hepatocytes (representing 2% of the total 
number of hepatocytes in a liver) that were genetically corrected in vivo with recombinant 
retroviruses (16, 18). The level of genetic reconstitution obtained in the WHHL gene therapy 
experiments was confirmed to be approximately 4% of endogenous hepatic levels using a 
quantitative RNase protection assay. 
[192] 
Recombinant DNA Research, Volume 15 
