A Highly Efficient, Stable, and Rapid Approach for
Ex Vivo Human Liver Gene Therapy Via a FLAP
Carlo Giannini,1,3Serban Morosan,1,4J. Guilherme Tralhao,1,5Jacques Emmanuel Guidotti,1,7Serena Battaglia,1,3
Karine Mollier,2Laurent Hannoun,6Dina Kremsdorf,1Helene Gilgenkrantz,7and Pierre Charneau2
Allogenic hepatocyte transplantation or autologous transplantation of genetically modified
hepatocytes has been used successfully to correct congenital or acquired liver diseases and
are neither easily maintained in culture nor efficiently genetically modified and are very
sensitive to dissociation before their reimplantation into the recipient. These difficulties
have greatly limited the use of an ex vivo approach in clinical trials. In the present study, we
have shown that primary human and rat hepatocytes can be efficiently transduced with a
FLAP lentiviral vector without the need for plating and culture. Efficient transduction of
nonadherent primary hepatocytes was achieved with a short period of contact with vector
particles, without modifying hepatocyte viability, and using reduced amounts of vector. We
high levels of transduction. Moreover, transplanted into uPA/SCID mouse liver, lentivirally
differentiated and functional phenotype as assessed by the stable detection of human albu-
min and antitrypsin in the serum of the animals for months. In conclusion, the use of FLAP
lentiviral vectors allows, in a short period of time, a high transduction efficiency of human
functional and reimplantable hepatocytes. This work therefore opens new perspectives for
the development of human clinical trials based on liver-directed ex vivo gene therapy.
approach has been limited mainly by the shortage of do-
rthotopic liver transplantation is the only effec-
tive therapy for several congenital and acquired
nor livers, the prohibitive cost, and the requirement of
lifelong immunosuppressive treatment. Segmental trans-
plants or living-donor liver transplantation has only mar-
ginally impacted on organ scarcity, and 10% of patients
still die while on waiting lists for liver transplantation.
Hepatocyte transplantation can be considered in some
cases as “bridging therapy” to enable a short-term im-
provement of liver functionality until a donor liver be-
comes available.1Hepatocyte transplantation has been
used successfully in humans as an alternative therapy to
based on transplantation of genetically engineered hepa-
tocytes should be considered as an alternative strategy
applicable to liver disorders as well as to metabolic disor-
ders nondirectly involving the liver. Ex vivo cell transduc-
tion before transplantation into the host avoids systemic
dissemination of viral particles, ensures a tissue-specific
expression of the transgene, and markedly reduces the
amount of vector particles required for direct in vivo
Abbreviations: HIV, human immunodeficiency virus; uPA, urokinase plasmin-
ogen activator; H?1-AT, human ?1-antitrypsin; GFP, green fluorescent protein;
BrdU, bromodeoxyuridine; VSV-G, vesicular stomatitis virus-G.
From the1PASTEUR INSERM Unite ´ 370, Necker Institute, Paris, France;
2Groupe de Virologie Mole ´culaire et Vectorologie, Pasteur Institute, Paris, France;
3Department of Internal Medicine, University of Florence, Florence, Italy;4Faculty
of Veterinary Medicine, University of Agronomical Science and Veterinary Medi-
cine, Iasi, Romania;5Department of Surgery 3, H. U. Coimbra, Medical Faculty,
Coimbra, Portugal;6Service de Chirurgie Digestive et He ´patobiliaire, Ho ˆpital
Pitie ´-Salpe ˆtrie `re, Paris, France; and7INSERM Unite ´ 567, UMR 8104 CNRS
Cochin Institute, Paris, France.
Received January 10, 2003; accepted April 9, 2003.
Supported by grants from INSERM, ARC, and Pasteur Institute.
C.G., S.M., and J.G.T. contributed equally to this work.
Address reprint requests to: Carlo Giannini, Ph.D., PASTEUR INSERM Unite ´
370, Necker Institute, 156, rue de Vaugirard, 75015 Paris, France. E-mail:
firstname.lastname@example.org; fax: (33) 1-40615581.
Copyright © 2003 by the American Association for the Study of Liver Diseases.
transduction. Moreover, a highly expressed therapeutic
gene in a limited number of cells may be sufficient to
achieve normal secretion of the corresponding protein in
the systemic circulation and to allow the correction of a
metabolic disorder.4-6In this line, the hepatocyte consti-
tutes an ideal target for gene therapy thanks to its endog-
enous secretory role.
For long-term correction of metabolic deficiencies,
stable and continuous expression of the transgene is re-
quired. Oncoretroviral-based vectors have been preferen-
tially used until now for this purpose.4,7A therapeutic
cholesterolemia after autotransplantation of hepatocytes
retrovirally transduced with the low-density lipoprotein
receptor gene.8However, the requirement of hepatocyte
division to achieve efficient gene transfer with oncoretro-
viral vectors has restricted their use in ex vivo liver gene
Considerable improvements have recently been
achieved in vector technology with the introduction of
human immunodeficiency virus (HIV)-1–derived lenti-
viral vectors.9,10These vectors maintain the capacity of
HIV to efficiently transduce nondividing cells and can be
obtained at high titers. After systemic administration,
VSV-G pseudotyped lentiviral vector can transduce a
broad spectrum of target cell types with extensive trans-
duction of several other organs (namely, spleen, bone
few reports have been published on the use of lentiviral
vectors for in vivo or ex vivo liver targeting. Moreover,
discordant results have been obtained concerning the re-
quirement of cell division to reach an efficient in vivo
lentiviral transduction of the hepatocytes.12-15
rat hepatocytes can be transduced very efficiently with a
FLAP lentiviral vector without plating and culture. The
presence of DNA FLAP sequences within the vector con-
struct (central poly-purine tract sequence and central
tion of nonadherent hepatocytes. In addition, after ge-
netic modification, adult primary human hepatocytes
were able to repopulate large areas of uPA/SCID mouse
liver, thus emphasizing their ability to proliferate and to
maintain a differentiated phenotype. These results open
new perspectives to liver-directed human ex vivo gene
Materials and Methods
Primary Hepatocyte Isolation. Primary rat hepato-
cytes were isolated by liver perfusion with a collagenase
blend (Liberase; Roche Molecular Biochemicals, Mann-
heim, Germany) as previously described.16
Primary human hepatocytes were isolated as described
elsewhere17from the healthy liver tissue of surgical liver
biopsy specimens (approximately 20-25 cm3) collected
after informed consent had been obtained from patients
undergoing therapeutic partial hepatectomy for liver me-
tastasis and benign hepatic tumor. Subjects with viral in-
fections (hepatitis C virus, hepatitis B virus, HIV),
cirrhosis, and primary hepatic carcinomas were not con-
least 3 cm from the metastasis. Collagenase H (Roche
Molecular Biochemicals) perfusion (500 ?g/mL, 2.4
extensive washing of the liver tissue with HEPES/ethyl-
enediaminetetraacetic acid buffer (pH 7.4) using a cathe-
ter inserted into the vessels on the cut surface of the
18 mL/min, and all solutions were prewarmed at 37°C.
Cells were then washed twice at 100g for 2 minutes, and
hepatocytes were separated from nonparenchymatous
cells by Percoll fractionation (30% isotonic Percoll solu-
tion, centrifuged at 450g for 4 minutes). Viable cells were
determined by trypan blue exclusion.
Lentiviral Vectors. TRIP-?U3-CMV-GFP (referred
described elsewhere.10,18Vector particles were produced
by the transient calcium phosphate cotransfection of
293T cells as previously described.19The concentrations
of vectors were normalized according to the p24 (HIV-1
capsid protein) content of supernatants. The titer of the
TRIP-GFP vector, measured on MT4 cell line (a very
permissive lymphocyte cell line), was 1010transducing
units/mL, and the p24 concentration was 97 ng p24 an-
In all experiments, the vectors were concentrated by
ultracentrifugation except when otherwise mentioned
(“nonconcentrated medium”). In the case of a noncon-
centrated vector supernatant, Dulbecco’s modified Eagle
medium was replaced by serum-free Williams medium in
293T cells 24 hours after transfection. The vector super-
natant was collected after a further 24 hours, cleared by
centrifugation (800g for 10 minutes), aliquoted, and
stored at ?80°C until use. The p24 antigen concentra-
tion of nonconcentrated vector stock was 740 ng/mL.
Nonadherent Transduction and Culture of Pri-
mary Hepatocytes. Immediately after isolation, hepato-
cytes were incubated at 37°C in Williams medium (107
cells/mL) to which lentiviral vector stock solutions were
added and maintained under rolling agitation. In case of
nonconcentrated vectors, hepatocytes were directly resus-
pended in vector supernatant. After the time intervals
HEPATOLOGY, Vol. 38, No. 1, 2003GIANNINI ET AL.115
indicated, the cells were washed twice and resuspended in
complete Williams medium (1 mg/mL bovine serum al-
bumin, 100 ?g/mL streptomycin, 100 U/mL penicillin,
25 nmol/L dexamethasone, and 5 ?g/mL bovine insulin)
2 to 3 ? 104cells/cm2. After 5 hours, serum-containing
dermal growth factor, 25 ng/mL; sodium pyruvate, 20
mmol/L). At the indicated time points, the cells were
washed twice with phosphate-buffered saline and fixed
with a 4% paraformaldehyde solution at 4°C for 15 min-
utes. Hoechst staining was performed by incubatingfixed
cells for 10 minutes in the presence of a 0.5 ?g/mL solu-
tion of Hoechst 33258 in phosphate-buffered saline. The
results presented herewith correspond to the average of 3
independent experiments; at least 500 cells were analyzed
per condition during each experiment.
Generation of uPA/SCID Mice. To generate the
uPA/SCID mice, we crossed the Alb-uPA (Jackson Lab-
oratories, Bar Harbor, ME) and SCID mouse lines
(IFFA-CREDO, Lyon, France) and, through selective
backcrossing, bred SCID trait to homozygosity. uPA
p1, 5?-ATTCTGGAGGACCGCTTATCTGT-3?; p2,
gous mice were used for transplantation experiments. Ho-
mozygosity of the SCID trait was confirmed by quantifica-
tion of total serum immunoglobulin G using a sandwich
enzyme-linked immunosorbent assay technique.
Liver Cell Transplantation. uPA/SCID mice were
anesthetized, the spleen was exteriorized through a small,
left flank incision, and a syringe with a 30-gauge needle
was used to inject 100 ?L of the cell suspension (5-6 ?
105viable human hepatocytes) in Williams medium (In-
vitrogen, Carlsbad, CA). The spleen was replaced in the
abdominal cavity, and the incision site was closed with
uPA/SCID mice were killed at 4 (n ? 3), 8 (n ? 3),
and 12 weeks (n ? 5) after the transplantation of hepato-
cytes transduced with green fluorescent protein (GFP)-
expressing lentiviral vectors.
Enzyme-Linked Immunosorbent Assay for Human
sera were quantified by a standard sandwich enzyme-
linked immunosorbent assay. Briefly, mouse anti-human
land, Gilbertsville, PA) were used as antigen-specific cap-
ture antibodies. After overnight coating, nonspecific
binding was blocked by 1 hour of incubation with 1%
bovine serum albumin at 37°C. After washing, 75 ?L/
well of 1:10 diluted mouse sera was added and incubated
overnight at 4°C. Rabbit anti-human albumin (0.16 ?g/
mL; Sigma Chemical Co.) and rabbit anti–h?1-AT (2
?g/mL; Sigma Chemical Co.) were used as antigen-spe-
cific indicator antibody. Horseradish peroxidase–conju-
gated anti-rabbit antibodies were purchased from Sigma
Chemical Co. The chromogen and the substrate were
used according to the manufacturer’s indications (Sigma
Chemical Co.). Absorbance values (405 nm) were con-
verted to micrograms per milliliter by the comparison
with a standard curve performed by using serial dilution
of purified human albumin and h?1-AT (Sigma Chemi-
Immunohistochemical Staining for h?1-AT. Fro-
zen liver sections (7 ?m) were immediately fixed for 10
minutes at 4°C in 4% paraformaldehyde solution and
then permeabilized by incubation in 0.2% Triton X-100
and 10% goat serum solution for 30 minutes. Endoge-
nous peroxidase activity was blocked with 0.3% H2O2in
phosphate-buffered saline for 5 minutes at room temper-
ature and then incubated with rabbit anti–h?1-AT anti-
bodies (Dako, Glostrup, Denmark) for 1 hour at room
temperature. Slides were developed using a secondary
anti-rabbit antibody conjugated with horseradish peroxi-
dase (EN-VISION-kit; Dako) following the manufactur-
er’s instructions. After immunohistochemistry, sections
were counterstained with Meyer’s hematoxylin (Dako).
Bromodeoxyuridine Staining. Primary human hep-
atocytes were transduced in a nonadherent manner with
cultured for 96 hours with 10 ?mol/L bromodeoxyuri-
dine (BrdU) in the absence or presence of 25 ng/mL epi-
washed twice and then fixed and stained following the
kit I; Roche Molecular Biochemicals). BrdU positivity of
cultured hepatocytes was assessed by light microscopy. At
least 500 cells were analyzed per data point. Presented
data represent the average of 3 independent experiments.
Statistical Analysis. P values were determined using
Fisher’s exact test.
Efficient Transduction of Nonadherent Human
and Rat Primary Hepatocytes Using FLAP Lentiviral
Vectors. To analyze the transduction rate of human and
rat primary hepatocytes with the HIV-1–derived FLAP
116GIANNINI ET AL. HEPATOLOGY, July 2003
in the presence of increasing concentrations of lentiviral
vector particles expressing GFP under the control of the
cytomegalovirus promoter (Fig. 1). Human hepatocytes
were transduced while still in suspension during a 2-hour
vector particle concentration corresponding to 1.5 ?g
p24 antigen/mL (Fig. 1A and B). Primary rat hepatocytes
were less permissive to transduction. At 1.5 ?g p24 anti-
gen/mL, approximately 60% of rat hepatocytes were
transduced. A plateau was reached at 70% efficiency for a
vector particle concentration of 3?g p24 antigen/mL.
These results indicate that primary rat and human hepa-
tocytes are efficiently transduced by FLAP lentiviral vec-
rat cells, we did not observe any toxicity induced by the
vector, even at the highest doses tested (5 ?g/mL of p24
antigen, data not shown).
A 100% hepatocyte transduction was achieved using
nonultracentrifugated vector supernatant conditioned in
Interestingly, transduction to homogeneity was achieved
compared with experiments performed using concen-
trated vector stocks (1,500 ng p24 antigen/mL).
To confirm that lentiviral vectors are able to transduce
nondividing hepatocytes, freshly isolated, nonadherent
human and rat hepatocytes were transduced (1.5 ?g p24
antigen/mL of TRIP-GFP) and further cultured either
with or without epidermal growth factor. BrdU was
added in the medium and maintained throughout the
Fig. 1. Efficient transduction of nonadherent, primary human and rat hepatocytes using FLAP lentiviral vectors. (A) Direct fluorescence images of
human (upper panels) and rat (lower panels) primary hepatocytes transduced using a nonadherent protocol with concentrated TRIP-GFP vectors.
Green fluorescence corresponds to GFP-expressing cells, whereas nuclei are counterstained in blue with Hoechst 33258 (original magnification
?250). (B) Summary table showing the percentage of transduced human and rat hepatocytes with different concentrations of TRIP-GFP vector. Cells
were incubated for 2 hours with vector solutions before washing and plating and were analyzed 4 days after transduction. (C) Histogram representing
transduction efficiency of primary human hepatocytes with indicated concentrations of nonconcentrated vector supernatant. Hepatocytes were
transduced for 2 hours in a nonadherent manner (see Materials and Methods).
HEPATOLOGY, Vol. 38, No. 1, 2003GIANNINI ET AL. 117
experiment (96 hours) as a marker for dividing cells. In-
corporation of BrdU was found in 63% ? 7.5% of hepa-
tocytes when cultured in the presence of epidermal
growth factor and in 7.6% ? 4.4% when cultured in
complete Williams medium without mitogenic stimula-
tion. An equivalent transduction efficiency of human
hepatocytes was obtained in both culture conditions
whether cells had incorporated BrdU or not (96.4% ?
3.6% vs. 97.1% ? 2.9% of GFP-positive cells) (Fig. 2A).
A direct delivery of GFP, referred to as a pseudotrans-
duction mechanism, could have explained these results.
To test this hypothesis, cells were treated with 1 ?mol/L
nevirapine (Boehringer Ingelheim, Ridgefield, CT), a
nonnucleoside HIV-1 reverse-transcriptase inhibitor,
leading to a dramatic decrease in transduction efficiency
vectors to transduce primary hepatocytes with equal effi-
ciency, whether or not they are induced to proliferate.
Efficient Transduction of Human and Rat Primary
Hepatocytes Can Be Achieved Using a Rapid Trans-
duction Protocol. The time required for ex vivo manip-
ulation of primary hepatocytes (isolation, transduction,
washing, and transplantation) has to be optimally mini-
mized in the view of a clinical liver gene therapy trial.
Therefore, we studied the kinetics of transduction of hu-
man and rat liver cells while in suspension with a subop-
timal dose of vector particles as determined previously
vector particles was sufficient to obtain 70% of the max-
imum value of transduction (34.6% ? 4.1% at 30 min-
utes vs. 47.7% ? 2% at 240 minutes and 43.7% ? 4.9%
human hepatocytes, respectively). This proportion in-
creased with time of contact and reached a plateau after 2
hours of incubation (43.7% ? 1.2% and 55% ? 7.3%
for rat and human hepatocytes, respectively) (Fig. 3B).
The permissiveness of rat hepatocytes to transduction
differed from that seen in human cells but exhibited sim-
ilar transduction kinetics (Fig. 3B). Interestingly, the
viability of nonadherent human and rat primary hepato-
cytes, as determined by trypan blue exclusion, was not
substantially modified by incubation with lentiviral vec-
tors, even after the longest period tested (90% ? 1.4%
and 92% ? 1.2% of viable hepatocytes for rat and hu-
man, respectively) (Fig. 3C).
Internal FLAP Is Essential for the Efficient Trans-
duction of Nonadherent Primary Human Hepato-
cytes. To determine the impact of DNA FLAP on the
efficiency of primary human hepatocyte transduction,
these cells were transduced using vectors containing the
DNA FLAP (TRIP-GFP) or not (HR-GFP). Transduc-
tions were conducted with vector concentrations, which
increased from 75 to 3,000 ng p24 antigen/mL (Fig. 4).
For each concentration, the TRIP vector was found to be
more efficient than the HR vector for nonadherent hepa-
8-fold (at 3,000 ng p24 antigen/mL; 98.8% ? 1.1% vs.
12.7% ? 1.9%) to 20-fold (at 250 ng p24 antigen/mL;
statistically significant for each vector concentration (P ?
.01). Moreover, TRIP-transduced hepatocytes exhibited
much brighter GFP fluorescence, suggesting multiple
vector genome integrations per cell using high vector
doses (data not shown).
Fig. 2. Efficient transduction of nonreplicating primary human hepa-
tocytes by FLAP lentiviral vectors. (A) BrdU incorporation (left panels)
and GFP expression (right panels) in primary human hepatocytes trans-
duced with 1.5 ?g p24 antigen/mL of TRIP-GFP vectors after 96 hours
of culture in the absence (upper panels) or presence (lower panels) of
mitogenic stimulation (epidermal growth factor). In the right panels,
nuclei are counterstained in blue with Hoechst 33258 (original magni-
fication ?250). (B) Effect of treatment with nevirapine on primary human
hepatocyte transduced with 1.5 ?g p24 antigen/mL. Cells were treated
(white bar) or not (gray bar) with nevirapine (NVP). At least 500 cells
were analyzed for GFP fluorescence per data point; presented data
represent the average of 3 independent experiments.
118GIANNINI ET AL.HEPATOLOGY, July 2003
These data show that the DNA FLAP structure is a
primary hepatocytes ex vivo.
Ex Vivo Transduced Human Hepatocytes, Trans-
planted in Mice, Maintain Transgene Expression,
Proliferation Potential, and Differentiation Status.
uPA/SCID mice have already been shown to provide
an excellent model for xenogenic repopulation of their
liver with human hepatocytes.20To show the ability of
transduced adult human hepatocytes to engraft, main-
tain, and proliferate, we therefore transplanted primary
adult human hepatocytes, transduced in a nonadherent
manner with TRIP-GFP, into the spleen of uPA/SCID
mice. Due to the continuous selective pressure induced
by the cytotoxicity of uPA transgene expression in the
liver, transduced human hepatocytes underwent strong
proliferative stimulation and repopulated large por-
tions of mouse liver, as shown by the presence of large
clusters of hepatocytes expressing high levels of GFP
(Fig. 5A). Successful transplantation was assessed by
the presence of human hepatocyte-specific proteins in
the serum of transplanted animals as well as the immu-
Fig. 3. Efficient transduction of primary human and rat hepatocytes can be achieved using a rapid transduction protocol. (A) Direct fluorescence
images of human primary hepatocytes incubated with TRIP-GFP vectors at 600 ng p24 antigen/mL for indicated times and then plated and cultured
for 96 hours. Green fluorescence corresponds to GFP-positive cells; nuclei are counterstained in blue with Hoechst 33258 (original magnification:
left and center panels, ?200; right panel, ?100). (B) Transduction efficiency of human (black lines) and rat (gray lines) primary hepatocytes
incubated with 600 ng p24 antigen/mL of TRIP-GFP vectors for indicated times. At least 500 cells were analyzed for GFP fluorescence per data point;
presented data represent the average of 3 independent transduction experiments. (C) Viability of human (black lines) and rat (gray lines) primary
hepatocytes incubated with 600 ng p24 antigen/mL of TRIP-GFP vectors for indicated times. Tests were performed in triplicate for each data point.
Presented results correspond to the average of 3 independent transductions. Standard error was always less than 1.5%.
Fig. 4. Internal FLAP is essential for the efficient nonadherent trans-
duction of primary human hepatocytes. Percentage of GFP-positive hu-
man primary hepatocytes transduced in a nonadherent manner for 2
hours with HR-GFP or TRIP-GFP lentiviral vectors at indicated concentra-
tions. At least 500 cells were analyzed for GFP fluorescence per data
point; presented data represent the average of 3 independent experi-
HEPATOLOGY, Vol. 38, No. 1, 2003 GIANNINI ET AL.119
nohistochemical detection of human hepatocyte-spe-
cific markers. An average rating of 70% successful
engraftment was obtained (11 positive of 16 trans-
planted mice). We evaluated the level of liver repopu-
lation by quantitative imaging on liver sections of
animals at week 12 after transplantation by using both
GFP and histochemical staining of h?1-AT. Human
cell clusters ranged from 8.9% to 16.1% (mean, 13.5%
? 9.1%) of total area of liver sections.
The human origin of GFP-expressing hepatocytes was
confirmed by immunohistochemical staining with anti-
h?1-AT (Fig. 5B) and anti-human hepatocyte antibodies
(data not shown). The absence of GFP-negative clusters
of human hepatocytes indicates the stability of the trans-
gene expression in mouse liver up to 12 weeks after trans-
To confirm the function and viability of human trans-
concentrations of human albumin and human ?1-AT
increased with time, thus confirming the expansion of
human hepatocytes, which reached a plateau level 4 to 6
weeks after transplantation and persisted for several
months thereafter (Fig. 5C and D).
tion of human albumin and ?1-AT were found in a con-
trol group of 10 uPA/SCID mice transplanted with
nontransduced human hepatocytes, indicating that the
transduction process does not modify cell physiology
(data not shown).
Hepatocyte transplantation can be considered as an
alternative approach to orthotopic liver transplanta-
tion. Moreover, the autotransplantation of genetically
engineered hepatocytes avoids immune rejection and
side effects of lifelong immunosuppressive treatment.
As opposed to direct in vivo administration, such an ex
vivo gene transfer (1) allows a specific and exclusive ex-
pression of the transgene in selected target cells, avoiding
the accidental transduction of undesired tissues by sys-
temic dissemination, and (2) requires vector amounts
compatible with standard laboratory scale vector prepara-
tions. However, only a few examples of liver-directed ex
vivo gene transfer in humans or large animal models have
ficulties, mainly related to the production and transduc-
enabled efficient gene transfer in plated simian hepato-
cytes using multiple rounds of transduction before rein-
troduction into the animal.22
The development of lentiviral vectors seems to have
Fig. 5. Lentivirally ex vivo transduced human hepatocytes transplanted into uPA/SCID mice maintain transgene expression, proliferation potential,
and differentiation status. (A) Direct fluorescence images of 2 representative clusters of GFP-positive human hepatocytes, ex vivo transduced with
TRIP-GFP vectors (green cells), integrated in mouse liver parenchyma (counterstained in red by 7-amino-actinomycin D), and still present 3 months
after human hepatocyte transplantation (frozen liver section, 7 ?m). (B) Immunohistochemical staining of a representative uPA/SCID mouse frozen
liver section with antibodies directed against h?-1 AT (left panel) and a serial section (right panel) analyzed by direct fluorescence after Hoechst
33258 counterstaining (animal killed 3 months after human hepatocyte transplantation). (C and D) Serum concentrations of (C) human albumin and
(D) h?1-AT detected in the uPA/SCID mouse at indicated times. Results correspond to the average of the values found in 5 transplanted uPA/SCID
120 GIANNINI ET AL.HEPATOLOGY, July 2003
ex vivo liver gene therapy.23,24We show herewith that,
using a lentiviral vector, we can bypass the critical steps of
the hepatocyte ex vivo transduction procedure, namely,
the plating and further deleterious trypsin dissociation of
cultured hepatocytes. We obtained a highly efficient gene
transfer using limited amounts of a lentiviral vector in a
very limited period of time in both rat and human hepa-
tocytes that are still in suspension. Moreover, this trans-
duction protocol does not alter hepatocyte viability.
The presence of a DNA FLAP structure in TRIP vec-
tors was crucial to this efficient gene delivery.25,26Dele-
tion of the DNA FLAP caused 8- to 20-fold reduction in
transduction efficiency that was not compensated for by
increasing the vector concentration. Finally, using the
medium recovered from vector-producing cells, we pro-
pose a simple protocol of hepatocyte transduction, avoid-
ing ultracentrifugation steps. Indeed, we noted a higher
transduction efficiency using a nonconcentrated vector
preparation. This observation can be explained by partial
damage to vector particles during the ultracentrifugation
and resuspension steps.
During the preparation of this report, Nguyen et al.
showed a 50% to 60% transduction efficiency using a
gene transfer protocol of nonadherent human hepato-
has not been addressed in the present study because, in
our hands, transduction efficiency and hepatocyte viabil-
ity close to 100% were achieved without antioxidants.
The significant improvement in transduction efficiency
between these recent results and those presented here
could be explained by differences in the gene transfer
protocol such as transduction volume or continuous roll-
ing agitation, allowing a more rapid and more efficient
transduction with a reduced amount of vector. Further-
more, we showed that primary human hepatocytes, iso-
lated from adult donors, transduced without plating were
able to colonize and repopulate large areas of uPA/SCID
mouse livers (Fig. 5A and B). Transduced human cells
a differentiated phenotype, as confirmed by the stable
detection of GFP in the liver and albumin and ?1-AT in
mouse sera and liver for several months (Fig. 5C and D).
These data strongly support the hypothesis that differen-
tiated hepatocytes isolated from adult donors can indeed
be used for liver-directed ex vivo gene therapy and, under
appropriate conditions, are able to replicate and repopu-
late a recipient liver, expanding the number of cells carry-
ing the transgene and maintaining a functional
phenotype. The presented results suggest that the global
expansion of human transplanted hepatocytes is not due
to some side effects of the transduction process, as con-
firmed by the analogous engrafting and repopulation po-
in uPA/SCID mice. Moreover, previous studies per-
formed in uPA immunodeficient mice described an
equivalent repopulation efficiency (up to 15%) by trans-
planting adult human hepatocytes.28Nevertheless, the
promoter and enhancer sequences in lentiviral vectors
used in this study have been deleted (?U3 versions). This
modification abolishes transactivation of cellular proto-
oncogenes by the vector LTR as recently observed in the
SCID X1 French trial (M. Cavazzana, personal commu-
nication, March 2003). Retroviral vector safety can be
herpes thymidine kinase, enabling elimination of trans-
tagenesis and frequency of potentially tumorigenic events
after lentiviral-mediated gene transfer remains to be di-
Previous studies have shown that allogenic hepatocyte
transplantation could improve the clinical status of pa-
tients harboring an inherited liver disorder such as orni-
thine carbamoyltransferase deficiency, ?1-AT deficiency,
glycogen storage disease type IA, and Crigler-Najjar syn-
drome.1-3,29In the patient with Crigler-Najjar syndrome,
transplanted hepatocytes, although representing only 2%
to 5% of total liver mass, were able to express sufficient
levels of uridine diphosphoglucoronate glucoronosyl-
transferase activity to reduce the need for photothera-
py.3We can therefore hypothesize that, using our ex
vivo approach with a lentiviral vector expressing the
therapeutic gene at high levels, a limited number of a
patient’s hepatocytes overexpressing the therapeutic gene
may be sufficient to restore a functional phenotype in an
autotransplantation protocol. In addition, some acquired
liver diseases, such as viral hepatitis C, can be targeted by
by transplanting hepatocytes transduced with vectors en-
coding virus-specific interfering RNAs or ribozymes.
These cells, protected by the viral infection, could prolif-
erate and colonize the host liver, taking advantage of the
clearing of infected hepatocytes by the host immune sys-
In conclusion, we provide here an efficient, simple,
and very rapid protocol for ex vivo gene transfer in
human hepatocytes with a FLAP lentiviral vector. This
protocol allows us to circumvent some of the major
obstacles previously encountered in ex vivo liver gene
therapy clinical trials, among which are the require-
ment to plate and culture hepatocytes during several
days, the huge quantities of viral particles needed, and
the high proportion of hepatocyte mortality after dis-
HEPATOLOGY, Vol. 38, No. 1, 2003 GIANNINI ET AL.121
sociation from culture plates. This approach could be
applied not only to pathologies involving liver function
deficiencies but also, because of the unique secretory
characteristics and strategic position of the liver, to
deficiencies in secreted factors such as hemophilia23or
to lysosomal enzyme deficiencies such as glycogen stor-
age disease type II (Pompe’s disease),30,31Fabry’s dis-
ease,32or Tay-Sachs disease.33
chot for critical reading of the manuscript, Chantal Des-
douets for helpful discussion, Olivier Bregerie for
technical assistance, and Martine Netter for her indis-
pensable aid in the preparation of figures.
The authors thank Christian Bre ´-
1. Strom SC, Fisher RA, Thompson MT, Sanyal AJ, Cole PE, Ham JM,
Posner MP. Hepatocyte transplantation as a bridge to orthotopic liver
transplantation in terminal liver failure. Transplantation 1997;63:559-
2. Muraca M, Gerunda G, Neri D, Vilei MT, Granato A, Feltracco P,
Meroni M, et al. Hepatocyte transplantation as a treatment for glycogen
storage disease type 1a. Lancet 2002;359:317-318.
3. Fox IJ, Chowdhury JR, Kaufman SS, Goertzen TC, Chowdhury NR,
Warkentin PI, Dorko K, et al. Treatment of the Crigler-Najjar syndrome
type I with hepatocyte transplantation. N Engl J Med 1998;338:1422-
4. Chowdhury JR, Grossman M, Gupta S, Chowdhury NR, Baker JR Jr,
gene therapy in LDLR-deficient rabbits. Science 1991;254:1802-1805.
based gene therapy. J Hepatobiliary Pancreat Surg 2001;8:51-57.
6. Wilson JM. Adenovirus-mediated gene transfer to liver. Adv Drug Deliv
7. Gupta S, Wilson JM, Chowdhury JR. Hepatocyte transplantation: devel-
opment of new systems for liver repopulation and gene therapy. Semin
Liver Dis 1992;12:321-331.
8. Grossman M, Rader DJ, Muller DW, Kolansky DM, Kozarsky K, Clark
BJ 3rd, Stein EA, et al. A pilot study of ex vivo gene therapy for homozy-
gous familial hypercholesterolaemia. Nat Med 1995;1:1148-1154.
into human lymphocytes by a defective human immunodeficiency virus
type 1 vector. J Virol 1991;65:532-536.
10. Naldini L, Blomer U, Gallay P, Ory D, Mulligan R, Gage FH, Verma IM,
et al. In vivo gene delivery and stable transduction of nondividing cells by
a lentiviral vector. Science 1996;272:263-267.
11. Pan D, Gunther R, Duan W, Wendell S, Kaemmerer W, Kafri T, Verma
IM, et al. Biodistribution and toxicity studies of VSVG-pseudotyped len-
of in vivo transduction of bone marrow. Mol Ther 2002;6:19-29.
12. Kafri T, Blomer U, Peterson DA, Gage FH, Verma IM. Sustained expres-
sion of genes delivered directly into liver and muscle by lentiviral vectors.
Nat Genet 1997;17:314-317.
13. Park F, Ohashi K, Chiu W, Naldini L, Kay MA. Efficient lentiviral trans-
duction of liver requires cell cycling in vivo. Nat Genet 2000;24:49-52.
14. Pfeifer A, Kessler T, Yang M, Baranov E, Kootstra N, Cheresh DA, Hoff-
man RM, et al. Transduction of liver cells by lentiviral vectors: analysis in
living animals by fluorescence imaging. Mol Ther 2001;3:319-322.
15. VandenDriessche T, Thorrez L, Naldini L, Follenzi A, Moons L, Berne-
man Z, Collen D, et al. Lentiviral vectors containing the human immuno-
deficiency virus type-1 central polypurine tract can efficiently transduce
16. McIntyre M, Desdouets C, Senamaud-Beaufort C, Laurent-Winter C,
Lamas E, Brechot C. Differential expression of the cyclin-dependent ki-
nase inhibitor P27 in primary hepatocytes in early-mid G1 and G1/S
transitions. Oncogene 1999;18:4577-4585.
17. Mazier D, Beaudoin RL, Mellouk S, Druilhe P, Texier B, Trosper J,
Miltgen F, et al. Complete development of hepatic stages of Plasmodium
falciparum in vitro. Science 1985;227:440-442.
18. Sirven A, Ravet E, Charneau P, Zennou V, Coulombel L, Guetard D,
Pflumio F, et al. Enhanced transgene expression in cord blood CD34(?)-
mouse repopulating cells, following transduction with modified trip lenti-
viral vectors. Mol Ther 2001;3:438-448.
Dubart-Kupperschmitt A, et al. The human immunodeficiency virus
type-1 central DNA flap is a crucial determinant for lentiviral vector nu-
clear import and gene transduction of human hematopoietic stem cells.
20. Mercer DF, Schiller DE, Elliott JF, Douglas DN, Hao C, Rinfret A,
Addison WR, et al. Hepatitis C virus replication in mice with chimeric
human livers. Nat Med 2001;7:927-933.
21. Kay MA, Baley P, Rothenberg S, Leland F, Fleming L, Ponder KP, Liu T,
et al. Expression of human alpha 1-antitrypsin in dogs after autologous
transplantation of retroviral transduced hepatocytes. Proc Natl Acad Sci
U S A 1992;89:89-93.
22. Andreoletti M, Loux N, Vons C, Nguyen TH, Lorand I, Mahieu D,
Simon L, et al. Engraftment of autologous retrovirally transduced
hepatocytes after intraportal transplantation into nonhuman primates:
implication for ex vivo gene therapy. Hum Gene Ther 2001;12:169-
23. Park F, Ohashi K, Kay MA. Therapeutic levels of human factor VIII and
IX using HIV-1-based lentiviral vectors in mouse liver. Blood 2000;96:
24. Galimi F, Verma IM. Opportunities for the use of lentiviral vectors in
human gene therapy. Curr Top Microbiol Immunol 2002;261:245-254.
25. Zennou V, Petit C, Guetard D, Nerhbass U, Montagnier L, Charneau P.
HIV-1 genome nuclear import is mediated by a central DNA flap. Cell
26. Zennou V, Serguera C, Sarkis C, Colin P, Perret E, Mallet J, Charneau P.
the brain. Nat Biotechnol 2001;19:446-450.
27. Nguyen T, Oberholzer J, Birraux J, Majno P, Morel P, Trono D. Highly
efficient lentiviral vector-mediated transduction of nondividing, fully re-
implantable primary hepatocytes. Mol Ther 2002;6:199.
28. Dandri M, Burda MR, Torok E, Pollok JM, Iwanska A, Sommer G,
Rogiers X, et al. Repopulation of mouse liver with human hepatocytes
and in vivo infection with hepatitis B virus. HEPATOLOGY 2001;33:
disease type II (Pompe disease). Mol Med Today 2000;6:245-251.
31. Amalfitano A, McVie-Wylie AJ, Hu H, Dawson TL, Raben N, Plotz P,
Chen YT. Systemic correction of the muscle disorder glycogen storage
disease type II after hepatic targeting of a modified adenovirus vector
encoding human acid-alpha-glucosidase. Proc Natl Acad Sci U S A 1999;
32. Ziegler RJ, Yew NS, Li C, Cherry M, Berthelette P, Romanczuk H, Ioan-
nou YA, et al. Correction of enzymatic and lysosomal storage defects in
Fabry mice by adenovirus-mediated gene transfer. Hum Gene Ther 1999;
33. Guidotti JE, Mignon A, Haase G, Caillaud C, McDonell N, Kahn A,
Poenaru L. Adenoviral gene therapy of the Tay-Sachs disease in hex-
osaminidase A-deficient knock-out mice. Hum Mol Genet 1999;8:831-
122GIANNINI ET AL.HEPATOLOGY, July 2003