Human Gene Therapy - April 2023 - 278

278
WESTHAUS ET AL.
encoding barcoded expression cassettes compatible with
NGS-based readout at the cell entry (DNA) and transgene
expression (RNA/cDNA) levels.31,36,43 AAV2, AAV5,
AAV8, and the bioengineered AAV-LK03 chosen as
clinical data from liver-targeted human studies was publicly
available.11 AAV-NP59 and AAV3b were included
because they were the bioengineered variant ofAAV2 and
the closest natural relative of AAV-LK03, respectively
(Supplementary Fig. S1).
The mix of the six AAV candidates was used to transduce
all different preclinical models evaluated in this
study, whereas the individual AAV candidates underwent
seroprevalence studies using individual human sera. The
functional data obtained were compared with publicly
available data on AAV performance in clinical trials as
well as between the models to identify differences (Supplementary
Fig. S1).
In vitro and ex vivo models of the human liver
The simplest models of human hepatocytes are based
on cell lines derived from hepatocellular carcinoma, such
as HuH-7, or hepatocellular blastoma, such as HepG2
cells.48,49 These lines are immortalized and self-renewing,
allowing for low-cost high-throughput experimentation
using standard laboratory equipment and reagents. To
evaluate both cell types for their potential to serve as biologically
predictive models ofhuman primary hepatocyte
transduction, cells were transduced at a dose of 1,000 and
200 vg/cell and were harvested 72 h after AAV exposure.
Cell pellets were processed for DNA and RNA/cDNA,
which allowed us to evaluate relative vector performance
at the cell entry and transgene expression levels, respectively
(Supplementary Fig. S2). The NGS results showed
that AAV2 outperformed the other variants at the DNA
and RNA levels in both HuH-7 and HepG2 cells. Of
interest, whereas AAV-LK03 and AAV3b performed
similarly at the cell entry level (DNA) in HuH-7 cells,
AAV-LK03 outperformed AAV3b at the transgene expression
level (RNA). In HepG2 cells, AAV-LK03 and
AAV3b performed similarly but were less efficient at both
DNA and RNA/cDNA levels than AAV5. The strong
performance for AAV2 in those cell lines was expected
based on data from HuH-7 cells we reported previously.36
Next, we studied the six vectors in iHeps and adult stem
cell-derived ductal organoids. It quickly became apparent
that the transcriptional dominance of AAV2, as shown in
the immortalized cell lines, did not fully translate to iHeps.
AAV2 was still able to enter the cells at the highest efficiency
at the DNA level, followed by AAV-LK03,
AAV3b, AAV5, AAV-NP59, and AAV8 (Fig. 1b), but at
the transgene expression level, AAV-LK03 outperformed
all other vectors (Fig. 1c).
Using adult stem cell-derived ductal organoids consisting
of liver ductal progenitor cells,50 we found AAV5
Figure 1. Ex vivo results in human and NHP liver models. (a) Schematic of transduction of indicated in vitro and ex vivo models. (b) NGS read contribution (%)
for each AAV from extracted DNA. (c) NGS read contribution (%) for each AAV from mRNA-derived complementary DNA. AAV, adeno-associated virus; NGS,
next-generation sequencing; NHP, nonhuman primate.

Human Gene Therapy - April 2023

Table of Contents for the Digital Edition of Human Gene Therapy - April 2023