Human Gene Therapy - April 2023 - 290

290
RANA ET AL.
INTRODUCTION
THE LIVER IS an important target for in vivo gene therapy
due to accessibility via peripheral vein injection, slow
hepatocyte turnover, and an inherently tolerogenic microenvironment
that favors long-term therapeutic gene
expression.1-3 Although many viral and nonviral gene
delivery methods into the liver have been described,4 arguably
the most effective vehicle to date has been the
adeno-associated virus (AAV) vector, with one FDAapproved
product for hemophilia B (Hemgenix) and
multiple treatments in late phase III clinical trials
(NCT04370054, NCT05345171, NCT03370913,
NCT03392974). One of the factors driving the mainstream
application of AAV as the gene delivery vehicle of choice
is its relative simplicity. The virion comprises a small
*20-25nm protein capsid that can package *4.7 kb of
the therapeutic transgene, flanked by cis acting 145 base
pair inverted terminal repeats (ITRs).5
Hepatic gene therapy with AAV vectors offers the potential
for long-lived correction of monogenic disorders
such as hemophilia A and B, ornithine transcarbamylase
deficiency, phenylketonuria, glycogen storage disease
type Ia, Fabry disease, and Crigler Najjar syndrome,6-16
among others. A large predictor of therapeutic efficacy is
hepatotropism of the vector capsid following intravenous
infusion. Serotypes considered for liver targeting in clinical
trials so far include AAV2,17 AAV8,10,11 and
AAV9,4,18 although their capacity for effectively transducing
human hepatocytes might not be optimal. The
capsid is also the major determinant of recognition by the
host immune system, resulting in the production of neutralizing
antibodies (NAb). Up to 80% of the human
population have pre-existing NAbs to the commonly used
AAV serotypes, serving as a major patient exclusion
criterion.19-22
Selecting a suitable vector serotype for hepatic gene
therapy therefore involves a calculated balance between
tissue tropism and evading immune responses. Approaches
to circumvent pre-existing NAbs include testing
other vector serotypes with lower pre-existing immunity
such as AAV56,23 and AAV624 (NCT04046224,
NCT03587116, NCT04370054), although these serotypes
might not transduce hepatocytes as efficiently.25,26 This
loss in efficiency is often offset by administering up to
6 · 1013 vg/kg titers in patients,6,27 which increases the
risk of eliciting potent immune responses against both the
vector and the transgene product.3,28-30 Thus, it is challenging
to find an existing natural capsid that checks all the
boxes for safe and efficacious hepatic gene delivery. More
effective vector capsid variants are needed to propel
clinical advancements in the gene therapy field.
Efforts to address this problem include the selection of
novel capsid serotypes isolated from ancestral or other
mammalian species.31-33 Alternatively, vector engineering
by capsid shuffling, rational design by site-directed
mutagenesis, directed evolution in the cell of interest, or a
combination of these strategies have been used for improving
transduction efficiency into the tissue of choice
and/or selecting for NAb evasion.34-38 This can result in
an entirely new recombinant capsid carrying components
of the parental capsids34,39,40 or can result in the incorporation
of modified amino acid sequences into the parent
serotype. Recently, AAV3B was identified as an emerging
clinical trial candidate for liver gene therapy because of its
improved tropism to hepatocytes.26,41,42 Its properties can
be further enhanced by incorporating targeted mutations.43-45
However, studies suggest a high prevalence of
pre-existing antibodies to AAV3 among individuals,46,47
which may exclude a large percentage of the eligible patient
population.
To address this, we generated an AAV3B combinatorial
targeted variant library and used directed evolution to
isolate enhanced liver targeting variants.35 Variants with
targeted mutations in the capsid variable region (VR)
loops36,48 gained a selective advantage after five rounds of
selection in human hepatocarcinoma spheroid cultures.35
In vivo, the variant AAV3B-DE5 exhibited enhanced human
hepatocyte tropism in a human liver chimeric mouse
model. Importantly, AAV3B-DE5 showed reduced seroreactivity
to human intravenous immunoglobulin (IVIg)
as well as to individual serum samples from 100 healthy
North American donors.35
The current study originated as a reanalysis of the data
set generated in our previous study. It was inspired by the
intriguing claims made by de Alencastro et al. in a recent
publication49 that AAV replication in the presence of adenovirus
helper factors can select for variants that replicate
more efficiently but are not necessarily better transducers.
Further, the authors claimed that using a low multiplicity
of infection (MOI) and multiple rounds of selection can be
detrimental to selection of the most desired variants.
Considering that we had used adenovirus help, low starting
MOIs and multiple (five) rounds of selection in our
study, this raised valid concerns about whether superior
transducing variants from early selection rounds were
possibly lost due to the faster replication kinetics of
AAV3B-DE5.
We therefore set out to screen earlier iterations from our
original dataset. In addition, we also simplified our dataset
by ignoring mutations in VR-I and treating everything
outside VR-IV to VR-VII as wild type (wt), after observing
that several sequences in the original dataset differed
only in VR-I without a concomitant difference in
enrichment.
This narrowed search identified two new variants that
were further tested for improved in vivo transduction of
human hepatocytes. One of these variants, AAV3B-V04,
was shown to have significantly improved tropism for
human hepatocytes compared to our previously identified
AAV3B-DE5 capsid. Interestingly, AAV3B-V04 was less
Human Gene Therapy - April 2023

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