A new surgical approach to improve gene transfer in liver using lentiviral vectors
Anne Dariel
a, Tuan Huy Nguyen
b, Virginie Pichard
b, Françoise Schmitt
a, Dominique Aubert
b, Nicolas Ferry
b, Guillaume Podevin
a,⁎
aPediatric Surgery Department, University Hospital, 44093 Nantes, France
bBiothérapies Hépatiques, INSERM EA4274, University Hospital, 44093 Nantes, France
Received 11 April 2008; revised 10 July 2008; accepted 14 August 2008
Key words:
Liver;
Gene therapy;
Lentivirus;
Crigler-Najjar
Abstract
Purpose: Metabolic inherited liver diseases are attractive targets for gene therapy. Recombinant lentiviruses are very powerful viral vectors able to infect nonmitotic cells. We wanted to develop a new surgical approach to improve gene transfer in adult liver using low viral doses.
Materials and methods:Adult rats were injected with 2.108 infectious particles of lentiviral vectors encoding the green fluorescent protein marker gene under control of a liver-specific promoter transthyretin. In the control group (n = 5), gene delivery was performed by inflow intraportal injection.
In the surgical group (n = 5), liver was completely excluded from systemic circulation before viral injection in infrahepatic vena cava with high pressure.
Results:At day 9, transduction efficiency was 14.35% in the surgical group 3 and 0.39% in the control group (P= .016). At month 2, the number of transduced hepatocytes decreased in the most part of rats, except in half of rats in the surgical group. Antibodies against green fluorescent protein were detected in all rats at month 2, except one in the surgical group.
Conclusions:We developed a new surgical approach allowing an efficient transduction of hepatocytes in adult rats using lentivirus at low viral doses. We have now to control the immune response to permit long-term expression of transgene.
© 2009 Elsevier Inc. All rights reserved.
Liver is a key organ for most metabolic pathways[1]. The molecular knowledge of inborn errors of liver metabolism has significantly improved, and metabolic-inherited liver diseases are now attractive targets for gene therapy with the aim to correct protein deficiencies. Recombinant lentiviruses are very powerful viral vectors for gene therapy. Such as
other retroviral vectors, they have the ability to integrate their transgene in genome of host cells, a prerequisite for long- term expression. Moreover, in contrast to other retroviruses such as murine oncoretrovirus, lentiviral vectors are able to infect nonmitotic cells provided they are in the G1 stage of the cell cycle [2]. Therefore, they could be used to correct metabolism deficiencies in the adult liver, which is composed of nondividing hepatocytes.
Our favorite model of metabolic inherited disorder is Crigler-Najjar type I disease. It is a rare (1/1 million birth) recessive inherited disease because of a complete defect in
⁎ Corresponding author. Sce chirurgie infantile, HME, 7 quai Moncousu, 44093 Nantes, France. Tel.: +33 240 083 585; fax: +33 240 083 546.
E-mail address:guillaume.podevin@chu-nantes.fr(G. Podevin).
www.elsevier.com/locate/jpedsurg
0022-3468/$–see front matter © 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.jpedsurg.2008.08.020
bilirubin conjugation that is characterized by high serum level of nonconjugated bilirubin. The existence of a rat model of the disease, the Gunn rat, has made possible to test various gene transfer strategies to correct the phenotype. We previously demonstrated that using lentiviral vectors more than 10% of transduction efficiency was required to correct completely the hepatic disorder [3,4]. In this study, lentiviral vectors were injected in newborn animals in which hepatocytes are cycling.
Such a direct delivery is not possible in adult animals because adult hepatocytes are in the G0 stage of the cell cycle and hence not susceptible to lentivirus injection.
We now want to develop a new surgical approach to improve gene transfer in adult liver using low viral doses of lentiviral vectors—the asanguineous liver in ischemia- hyperpressure procedure. This approach has 2 main advantages. Firstly, injection of lentiviral vectors after vascular exclusion of liver from systemic circulation could create a viral retention and concentrate virus in the liver decreasing extrahepatic dissemination. Secondarily, the liver sinusoidal endothelium contains fenestrations of 100 to 150 nm in diameter according to different species. The size of lentiviral virus is close to that value. Nevertheless, Snoeys et al[5]demonstrated that with a peripheral injection of other viruses, such as recombinant adenoviruses (93 nm size), the sinusoidal endothelium could restrain hepatocyte transduc- tion. Hyperpressure could force recombinant lentiviruses to cross the endothelial barrier as for adenoviruses[6]. Finally, it also has been shown that a transient liver ischemia contributed to an increase in the fenestration diameter.
Here, we used heterozygous Gunn rats to test our hypothesis using lentiviral vectors carrying a green fluor- escent protein (GFP) reporter gene.
1. Materials and methods 1.1. Animals
Heterozygous Gunn rats of both sexes and weighing 145- 175 g were used in this study. They were obtained from our breeding colony by mating homozygous Gunn males with heterozygous females. Animals were maintained under a 12- hour light cycle and fed ad libitum [3]. All animals were housed at the animal facilities of Nantes University Medical School (Nantes, France) and received human cares according to the guidelines of the French Ministère de l'Agriculture (Paris, France).
1.2. Surgical procedures
All surgical procedures were conducted on deeply anesthetized animals using isoflurane inhalation (3% vol/
vol in air).
Asanguineous liver in ischemia-hyperpressure procedure.
Rats underwent laparotomy, and liver was totally excluded
from systemic blood stream by successively clamping the hepatic artery, the portal vein, the suprahepatic vena cava, and the infrahepatic vena cava. A 22-gauge catheter was introduced in the infrahepatic vena cava, and 8.9 mL/kg body weight of viral supernatant was infused, corresponding to 10 cm of water of intrahepatic pressure. This intrahepatic pressure was measured with a graduated water column linked to the catheter. The supernatant was removed from the infrahepatic vena cava by gentle aspiration with a syringe after a 10-minute viral incubation, and blood flow was reestablished. Total ischemia did not exceed 20 minutes. A blood sample was taken in the infrahepatic vena cava 5 minutes after having removed clamps to assay circulating viral particles on Hepatoma HuH7 cells. Nine days after vector injection, the median liver lobe was removed for quantitative real time polymerase chain reaction (q-PCR) and immunohistochemistry analysis. Two months after surgery, rats were killed and liver was also analyzed by q-PCR and immunohistochemistry.
Portal vein injection procedure. Viral supernatant was injected in the portal vein in 1 minute without vascular exclusion of liver using a 30-gauge syringe. The infused volume was adjusted to 8.9 mL/kg body weight.
Preischemia procedure. Asanguineous liver in ischemia- hyperpressure procedure was performed as described above but infusing phosphate buffered saline (PBS). After 20 minutes of vascular exclusion of liver, clamps were removed and viral supernatant was injected through the portal vein in 1 minute.
1.3. Description, production, and titration of lentiviral vectors
HeLa, HuH7, and 293T cells were maintained in Dulbecco modified Eagle medium containing 10% fetal bovine serum, 10 mmol/L of 4-(2-hydroxyethyl)-1-piperazine- ethanesulfonic acid (HEPES), 2 mmol/L of glutamine, and antibiotics.
High-titer lentiviral vector stocks were generated as previously described by calcium phosphate–mediated tran- sient transfection of 293T cells by the 3 following plasmids [7]: the vector transfer plasmid, the packaging plasmid psPAX2, and the vesicular stomatitis virus G protein (VSVG) envelope protein-coding plasmid pMD2G. These self- inactivated transfer vectors harbored the GFP complementary DNA under the control of a promoter that specially directs transcription to the liver, the murine transthyretin promoter (mTTR) fused to a synthetic hepatocyte-specific enhancer [3]. They also harbored thecis-acting central polypurine tract (cPPT) and the central termination sequence (CTS) from the human immunodeficiency virus 1, which facilitates the nuclear translocation of preintegrative vector complexes and the posttranscriptional regulatory element from the woodchuck hepatitis virus, which increased transgene expression. Vector titers were determined on HeLa cells by q-PCR using Mesa Green q-PCR MasterMix Plus
(Eurogentec, France) and with a ABI Prism 7700 sequence detection system. For this titration, the following primers (Eurogentec) specific for 5′-untranslated lentiviral vectors were used: GAG-F, GGAGCTAGAACGATTCGCAGTTA;
GAG-R, GGTTGTAGCTGTCCCAGTATTTGTC. For nor- malization of the amount of genomic DNA, the following primers specific for §-actin gene were used: HB2-F, TCCGTGTGGATCGGCGGCTCCA; HB2-R, CTGCTTG- CTGATCCACATCTG. A standard curve was generated by using dilutions of lentiviral vector plasmid in genomic DNA extracted from HeLa cells as previously described[3].
1.4. Liver function assays
Blood samples were drawn from retroorbital sinus.
Alanine aminotransferase (ALAT) activity was measured at the routine biochemistry department of Nantes University Hospital, using ALAT/glutamate pyruvate transaminase kit according to the manufacturer's instruction (Roche, Meylan, France) on Hitachi 717 analyzer.
1.5. Quantitative real time PCR analysis
High-molecular-weight DNA (100 ng) was subjected to amplification by q-PCR on an ABI Prism 7000 using SYBR green (Mesa Green q-PCR mix, Eurogentec) in triplicates. For this q-PCR, the following primers specific for 5′-untranslated lentiviral vector were used: GAG-F, GGAGCTAGAAC- G AT T C G C A G T TA a n d G A G - R , G G T T G TA G - CTGTCCCAGTATTTGTC. For normalization of the amount of genomic DNA, the following primers specific for the X-chromosomal rat HPRT sequence were used: HPRT-F, GCGAAAGTGGAAAAGCCAAGT and HPRT-R, GCCA- CATCAACAGGACTCTTGTAG. A standard curve was constructed by using dilutions of lentiviral vector DNA in rat genomic DNA simulating the presence of 100 to 0.01 vector copy per haploid genome. Vector copy numbers were calculated by interpolating Ct(GAG)-Ct(HPRT) sample values to standard curve values.
1.6. Antibody detection
We detected the presence of antibodies against GFP in rat serum by direct enzyme-linked immunosorbent assay.
Ninety-six-well dishes were coated overnight at 4°C with purified GFP (Clontech) at 0.5 (g/well). After rinsing with PBS/Tween 20 (0.5% vol/vol), serial dilutions of the serum were incubated for 1.5 hours at 37°C. After washing with PBS/Tween 20, the presence of antibodies was revealed using biotinylated antirat IgG immunoglobulin and strepta- vidin/peroxidase followed incubation with 2,2′-Azinobis 3- ethylbenzothiazoline-6-sulfonic acid diammonium salt (ABTS) substrate (Roche). Optical density was read at 405 nm.
1.7. Immunohistochemistry
The presence of GFP-positive hepatocytes was assessed by immunohistochemistry on formalin-fixed/paraffin- embedded sections (5 μm). Paraffin was extracted from sections and endogenous peroxidase activity was inhibited by incubation for 30 minutes in a 3% H2O2 solution in PBS. Monoclonal primary mouse anti-GFP antibody (Clontech) diluted 1:100 in PBS containing bovine serum albumin (2% wt/vol) and Tween 20 (0.1% vol/
vol) was applied for 2 hours at room temperature. The GFP-positive cells were revealed with biotinylated goat antimouse immunoglobulin and streptavidin-peroxidase (Vectastain Universalis ABC kit) using diaminobenzidine (Vector) as a chromogenic substrate. Slides were counter- stained with hematoxylin, and the GFP-positive cell index was calculated as the percentage of positively stained cells in 10 fields at ×40 magnification (one field contained 300 hepatocytes).
1.8. Statistical analysis
Mann-Whitney test was used to compare quantitative data between groups (Pb.05).
Fig. 1 Green fluorescent protein immunochemistry at day 9. Representative histochemical analysis of liver of rats injected with lentiviral vectors expressing GFP in groups 1 (portal vein) and 3 (ischemia-hyperpressure). Transduced hepatocytes were identified by immunostaining of liver biopsy specimens obtained in groups 1 (left panel) and 3 (right panel) on day 9 after gene transfer using antibodies directed against GFP. The GFP-positive cells appear brown (original magnification, ×100). (For interpretation of references to color in this figure legend, the reader is referred to the web version of this article.)
2. Results
2.1. Hepatocyte transduction
Rats were injected with 2.108 viral infectious particles of lentivirus mTTR-GFP. In control groups, injection was realized according to the portal vein injection procedure for group 1 (n = 5), the preischemia procedure for group 2 (n = 3), and the asanguineous liver in ischemia-hyperpres- sure procedure for group 3, the surgical group (n = 5).
To determine the transduction efficiency, GFP immuno- histochemistry was performed 9 days after viral injection on liver biopsy specimens. Transduction efficiency was 14.35%
(2.02-21.82) in group 3 (ischemia-hyperpressure), whereas the transduction efficiency was 0.39% (0.03-2.37) in group 1 (portal vein) and 1.19% (0.29-2.97) in group 2 (pre- ischemia). Hepatocyte transduction was significantly higher in group 3 than that in group 1 (P= .016). There was no statistically significant difference between the control groups (P= .47) (Fig. 1andTable 1).
To confirm these results, we determined the number of vector copies existing in liver biopsy specimens by q-PCR. In group 3 (ischemia-hyperpressure), the median number of copy per haploid genome was 4.75 × 10−2 (2.14 ×10−2-12.37 × 10−2). In group 1 (portal vein), the q-PCR values were significantly lower than that in group 3 as follows: 0.81 × 10−2 (0.49 × 10−2-1.23 × 10−2) vector copies per haploid genome (P = .009) (Table 2).
2.2. Long-term expression
To assess the persistence of transduced hepatocytes at long-term, rats were killed 2 months after vector injection, and GFP immunohistochemistry was performed on liver
biopsy specimens. In group 1 (portal vein), the number of transduced hepatocytes decreased in 2 rats and remained the same in the others, whereas in group 2 (preischemia), the decrease of transduction efficiency affected all rats. In group 3 (ischemia-hyperpressure), the number of transduced hepatocytes also decreased in 2 rats to reach the values of control groups (b1%), whereas in the other two, the transduction efficiency decreased but remained high (9.5%
and 17%) (Table 1). These results were confirmed by q-PCR.
2.3. Liver toxicity
Hepatocellular toxicity of lentiviral vectors and surgical approach of asanguineous liver in ischemia-hyperpressure was analyzed by measuring serum ALAT levels before viral injection, at days 1 and 2 after viral injection. Before viral injection, median ALAT levels were 0.82 (kat/L, 1.03 (kat/L, and 1.19 (kat/L, respectively, in groups 1 (portal vein), 2 (preischemia), and 3 (ischemia-hyperpressure). A transient increase to moderate levels was detected in serum at day 1 in group 2 (1.12 (kat/L) (not in groups 1 and 3, respectively, 0.77 and 1.15 (kat/L), but this increase was not statistically significant (PN.05). The ALAT levels decreased to normal at day 2. There was no significant difference in ALAT levels between groups (PN.05).
2.4. Antibodies against GFP
Antibodies against GFP were detected in sera of all injected rats at month 2, except in one rat of group 3.There was no statistically significant difference for antibodies concentrations in serum between the 3 groups (PN.05).
Table 1 Transduction efficiency at day 9 and month 2 Rat % transduction
at day 9
% transduction at month 2 Group 1 (portal vein) 53 0.15 0.18
54 0.39 0.1
55 2.37 -
92 1.97 0.37
93 0.03 0.6
Group 2 (preischemia) 56 2.97 0
57 0.29 0.03
58 1.19 0.1
Group 3 (ischemia- hyperpressure)
50 14.35 -
51 11.54 0.03
52 2.02 0.13
94 21.82 17
95 16.1 9.5
This presents the comparison between the liver transduction efficiency in the different experimental groups at day 9 and month 2 after gene transfer to evaluate the persistence of GFP-positive hepatocytes at long-term.
Table 2 Quantitative real time PCR analysis of lentiviral gene transfer in liver
Rat No. of copies (×10−2)
Median (×10−2)
Group 1 (portal vein) 53 0.81 0.812⁎
54 0.53 55 1.23 92 0.49 93 0.91
Group 2 (preischemia) 56 2.45 1.68
57 1.46 58 1.68 Group 3 (ischemia-
hyperpressure)
50 2.14 4.75⁎
51 12.37 52 4.75 94 12.16 95 4.49
High-molecular-weight DNA was extracted from liver biopsy speci- mens 9 days after gene transfer, and the number of lentiviral vectors copies per haploid genome was determined in each group by quantitative real time PCR.
⁎ P= .009 between groups 1 and 3.
2.5. Postdelivery circulating viral particles
Viral load was measured in blood samples 5 minutes after viral injection or removal of vascular clamps. Viral input was 2.108 infectious particles in each group. The median number of circulating infectious particles 5 minutes after the end of surgery was significantly lower (P= .016) in group 3 (3.57 × 104) than that in group 1 (2.36 × 105) but was not statistically different between groups 1 and 2 (P = .07). Ischemia- hyperpressure resulted in a higher decrease (5602-fold decrease) in circulating viral particles than portal vein injection (847-fold decrease) (Table 3). At day 1, viral load was not detectable any more.
3. Discussion
When lentiviral vectors are administered to adult rats, hepatocytes are hardly transduced without high viral doses, and nonparenchymal cells account for most of transduced liver cells [8]. We developed a new surgical approach, asanguineous liver in ischemia-hyperpressure, which improved gene transfer in liver using lentiviral vectors in adult rats. In previous experiments, we injected high viral titers, and transduction efficiency was very efficient (45%) but was the same in the surgical group than in the control groups (data not shown). By injecting low viral doses, the transduction efficiency was superior to 10% in the surgical group, required transduction efficiency to correct completely the hepatic disorder and was significantly higher than in the intraportal injection control group (P = .016). A poor transduction efficiency of peripheral delivery was expected.
Nevertheless, we showed with the preischemia procedure that transient liver ischemia and hyperpressure realized
without simultaneous retention of viral vectors were not sufficient to significantly increase hepatocyte transduction.
The high efficiency of the surgical procedure was probably because of the combination of 2 mechanisms. The first and probably the most important mechanism was the retention of lentiviral vectors in liver facilitating contact between vectors and hepatocytes. The second was the hyperpressure brought on sinusoidal endothelium that could help viral vectors to cross fenestrations to reach hepatocytes. Interestingly, the asanguineous liver in ischemia-hyperpressure procedure decreased the level of circulating infectious particles after viral delivery, but there was no correlation between the number of GFP-positive hepatocytes at long-term and antibody levels in serum. Indeed, we were able to take off part of the virus before removal of vascular clamps. Vector dissemination could probably be further decreased if we would have washed the liver before removal of clamps. This wash step could be planned in larger animal but is difficult to carry out in rat because of the 20-minute maximum ischemia
—the liver injury beyond this period making the results not interpretable. However, our present results demonstrate that our new approach of gene therapy increased the biosafety of using viral vectors because of the more efficient reduction of the systemic postdelivery viral load in the surgical group than in the control groups. Finally, neither lentiviral vectors' administration nor the surgical procedure was responsible for liver toxicity in this study indicating that it should be clinically applicable.
We showed a decrease in GFP-positive hepatocytes in more than half of rats 2 months after lentivirus mTTR- GFP injection. The disappearance of transduced cells could be related either to the shutoff of the mTTR promoter or to the elimination of transduced hepatocytes. Detection of antibodies against GFP in serum demonstrated the induction of a humoral immune response but could not
Table 3 Postdelivery circulating infectious particles
Rat Viral input No. of circulating infectious particles
Median Fold decrease
Group 1 (portal vein) 53 2.108 1.56 × 105 2.36 × 105⁎ 847
54 1.37 × 105
55 4.19 × 105
92 2.36 × 105
93 3.67 × 105
Group 2 (preischemia) 56 3.96 × 105 1.04 × 106 192
57 1.04 × 106
58 1.18 × 106
Group 3 (ischemia-hyperpressure) 50 3.57 × 104 3.57 × 104⁎ 5602
51 2.74 × 104
52 3.19 × 104
94 1.37 × 105
95 7.60 × 104
Circulating infectious viral particles were evaluated in blood samples 5 minutes after viral injection in groups 1 and 2 and after removal of vascular clamps in group 3. The number of circulating particles was calculated by multiplying the viral titer measured in blood samples by the animal blood volume (15 mL/
250 g of body weight).
⁎ P= .017 between groups 1 and 3.
explain the clearance of GFP expressing cells. The mTTR hepatospecific promoter already allowed our team to achieve a complete and long-term correction of hyperbilir- ubinemia in newborn Gunn rats—animal models of the Crigler-Najjar type 1 disease[3]. Another team successfully used the mTTR promoter without shuttoff at long-term[9].
Furthermore, immunohistochemistry showed a decrease of transduced hepatocytes confirmed by q-PCR. Therefore, the more likely hypothesis is the involvement of a cytotoxic immune response directed against the transgene product in elimination of transduced hepatocytes. Accordingly, Brown et al [11] recently showed that a lentiviral vector carrying the mTTR promoter did not prevent a cytotoxic response against the transgene product, leading to the clearance of transduced hepatocytes.
To prove that elimination of transduced hepatocytes was dependent on cytotoxic T lymphocytes in rat, we planned to induce a CD8-depletion using anti-CD8 OX-8 antibodies. As we described [10], immunotolerance to transgene product could be induced with CTLA4Ig by direct injection of the drug or by injection of recombinant adenoviruses encoding CTLA4Ig. Another strategy will be to prevent expression of the mTTR lentiviral vector in antigen-presenting cells by incorporating the target sequence for the micro RNA 142pT in the vector backbone, as recently described[11].
In conclusion, we developed a new surgical approach, asanguineous liver in ischemia-hyperpressure, which allowed an efficient transduction of more than 10% of hepatocytes in adult rats using lentivirus mTTR-GFP at low viral doses. Nevertheless, we have now to confirm these results with a higher number of animals and also control the immune response to allow long-term expression of transgene. Interestingly, this approach decreased the postdelivery viral load, which could be in favor of the decrease of the viral dissemination and the increased biosafety, in association with the use of low viral doses.
Such results give support to perfect this surgical strategy in
Gunn rats, animal model of Crigler-Najjar type 1 disease, before developing preclinical studies in nonhuman pri- mates followed by clinical trials in humans to correct inherited disorder of liver.
References
[1] Ferry N. Gene therapy and liver diseases. Gastroenterol Clin Biol 2003;27:288-90.
[2] Kafri T, Blomer U, Peterson DA, et al. Sustained expression of genes delivered directly into liver and muscle by lentiviral vectors. Nat Genet 1997;17(3):314-7.
[3] Nguyen TH, Belloti-Privato M, Aubert D, et al. Therapeutic lentivirus- mediated neonatal in vivo gene therapy in hyperbilirubinemic Gunn rats. Mol Ther 2005;12(5):852-9.
[4] Nguyen TH, Aubert D, Bellodi-Privato M, et al. Critical assessment of lifelong phenotype correction in hyperbilirubinemic Gunn rats after retroviral mediated gene transfer. Gene Ther 2007;14(17):1270-7.
[5] Snoeys J, Lievens J, Wisse E, et al. Species differences in transgene DNA uptake in hepatocytes after adenoviral transfer correlate with the size of endothelial fenestrae. Gene Ther 2007;14(7):604-12.
[6] Lievens J, Snoeys J, Vekemans K, et al. The size of sinusoidal fenestrae is a critical determinant of hepatocyte transduction after adenoviral gene transfer. Gene Ther 2004;11:1523-31.
[7] Nguyen TH, Oberholzer J, Birraux J, et al. Highly efficient lentiviral vector-mediated transduction of nondividing, fully reimplantable primary hepatocytes. Mol Ther 2002;6(2):199-209.
[8] Follenzi A, Sabatino G, Lombardo A, et al. Efficient gene delivery and targeted expression to hepatocytes in vivo by improved lentiviral vectors. Hum Gene Ther 2002;13(2):243-60.
[9] Vigna E, Amendola M, Benedicenti F, et al. Efficient Tet-dependent expression of human factor IX in vivo by a new self-regulating lentiviral vector. Mol Ther 2005;11(5):763-75.
[10] Aubert D, Menoret S, Chiari E, et al. Cytotoxic immune res ponse blunts long-term transgene expression after efficient retro- viral-mediated hepatic gene transfer in rat. Mol Ther 2002;5(4):
388-96.
[11] Brown BD, Venneri MA, Zingale A, et al. Endogenous microRNA regulation suppresses transgene expression in hematopoietic lineages and enables stable gene transfer. Nat Med 2006;12(5):585-91.
