RNAi as Antiviral Therapy: The HIV-1 Case - RNA Interference from Biology to Therapeutics - page 227

strand [ 80, 158, 159 ]. Such off-target effects are notoriously difficult to predict, and
one should carefully screen for adverse effects in appropriate experimental systems.
Another problem relates to the induction of an immune response by transfected
siRNA and intracellularly expressed shRNA duplexes [ 160, 161 ], but this effect can
be minimized by optimal design of the si/shRNA molecule [ 162 ] . We recently
developed an ultrasensitive assay to score minor cell growth retardation effects in
stably transduced shRNA-expressing cell lines [ 163 ] .
When potent antiviral shRNAs have been identified in cell culture experiments,
one can move to relevant preclinical models to critically assess the safety and
efficacy of the proposed therapy. A simple and efficient in vitro test system to mea-
sure the impact of shRNA expression on cell viability is to perform a coculture of
the transduced GFP + cells and nontransduced GFP − cells (Eekels and Berkhout,
submitted for publication). A reduction in the percentage of GFP + cells over time
forms an indication of delayed cell growth and RNAi toxicity. Outgrowth of the
transduced and thereby protected cells should occur in the presence of HIV-1, which
can also be screened for by using simple FACS analysis of a mixed cell culture.
The SIV/macaque model [ 164 ] is used extensively for vaccination studies but
could also be considered for testing of an RNAi gene therapy. However, this model
has several limitations. First, anti-HIV shRNA cannot easily be tested against SIV
because of sequence dissimilarity. Thus, one should either convert the anti-HIV into
anti-SIV shRNA, which may affect their inhibitory power, or HIV-1 target sequences
should be incorporated into the SIV test genome. Second, transduction of the HIV-
based lentiviral vector is restricted by TRIM5a in macaque cells [ 165 ] . Third,
macaque experiments are rather expensive, and the number of animals may be
restricted because of budgetary or ethical reasons.
Most of these limitations do not apply to humanized immune system (HIS)
mouse models [ 166, 167 ]. All major human myeloid and lymphoid cellular com-
partments develop and mature from input human stem cells in the most recent HIS
mouse [ 168- 170 ]. This model provides access to in vivo and ex vivo experimenta-
tion on human T cells [ 171 ]. HIS mice can be infected by intravenous injection of
the virus but also via rectal and vaginal transmission routes. Infection results in
viremia and depletion of human CD4 + cells [ 172- 177 ]. We used this model to test
the safety and efficacy of a lentiviral-based gene therapy in hematopoietic stem
cells [ 4 ]. These animal models and their advantages and disadvantages have been
reviewed [ 178 ] .

11.5 Safety Concerns of Gene Therapy

By now, more than 1,700 clinical trials involving a gene therapy have been
performed ( http://www.wiley.co.uk/genmed/clinical ), and trials up to 2007 are
reviewed in [ 179 ]. The first gene therapy patient was treated in 1989 for adenosine
deaminase deficiency, a severe form of combined immunodeficiency (SCID) [ 180 ] .
Another gene therapy patient was treated in 1999 for a genetic liver disease and