Adeno-associated virus vector serotypes mediate sustained correction of bilirubin UDP glucuronosyltransferase deficiency in rats

ARTICLE

doi:10.1016/j.ymthe.2006.01.014

Adeno-associated Virus Vector Serotypes Mediate Sustained Correction of Bilirubin UDP Glucuronosyltransferase Deficiency in Rats Jurgen Seppen,1 Conny Bakker,1 Berry de Jong,1 Cindy Kunne,1 Karin van den Oever,1 Kristin Vandenberghe,1 Rudi de Waart,1 Jaap Twisk,2 and Piter Bosma1,* 1

Academic Medical Center Liver Center, 1105 BK Amsterdam, The Netherlands 2 AMT BV, 1105 BA Amsterdam, The Netherlands

*To whom correspondence and reprint requests should be addressed at Academic Medical Center Liver Center, S1-166, Meibergdreef 69, 1105 BK Amsterdam, The Netherlands. Fax: +31205669190. E-mail: [email protected].

Available online 3 April 2006

Crigler–Najjar (CN) patients have no bilirubin UDP glucuronosyltransferase (UGT1A1) activity and suffer brain damage because of bilirubin toxicity. Vectors based on adeno-associated virus (AAV) serotype 2 transduce liver cells with relatively low efficiency. Recently, AAV serotypes 1, 6, and 8 have been shown to be more efficient for liver cell transduction. We compared AAV serotypes 1, 2, 6, and 8 for correction of UGT1A1 deficiency in the Gunn rat model of CN disease. Adult Gunn rats were injected with CMV-UGT1A1 AAV vectors. Serum bilirubin was decreased over the first year by 64% for AAV1, 16% for AAV2, 25% for AAV6, and 35% for AAV8. Antibodies to UGT1A1 were detected after injection of all AAV serotypes. An AAV1 UGT1A1 vector with the liver-specific albumin promoter corrected serum bilirubin levels but did not induce UGT1A1 antibodies. Two years after injection of AAV vectors all animals had large lipid deposits in the liver. These lipid deposits were not seen in age-matched control animals. AAV1 vectors are promising candidates for CN gene therapy because they can mediate a reduction in serum bilirubin levels in Gunn rats that would be therapeutic in humans. Key Words: AAV, liver, gene therapy, Crigler–Najjar, Gunn rat

INTRODUCTION The majority of inherited liver diseases can be cured only by liver transplantation. Because these disorders are usually caused by a single gene defect, gene therapy is a much more attractive treatment. A good model system to study gene therapy for inherited liver diseases is the deficiency in bilirubin metabolism of Crigler–Najjar type 1 (CN) patients. Bilirubin is the breakdown product of the heme group of hemoglobin and other heme-utilizing enzymes such as cytochrome P450s. Because bilirubin is hydrophobic, it needs to be glucuronidated by the hepatic enzyme bilirubin UDP glucuronosyltransferase (UGT1A1) before it can be excreted into bile [1,2]. Patients with CN type 1 have no detectable UGT1A1 activity [3] and therefore have high serum levels of unconjugated bilirubin. Because unconjugated bilirubin is highly neurotoxic [4], UGT1A1 deficiency leads to brain damage and death if not treated. The only permanent treatment option for patients with CN is liver transplantation. The Gunn rat is

MOLECULAR THERAPY Vol. 13, No. 6, June 2006 Copyright C The American Society of Gene Therapy 1525-0016/$30.00

a natural mutant that has no UGT1A1 activity and is therefore a good model for CN and the development of gene therapy for this disease [5]. A number of gene therapy studies have achieved correction of UGT1A1 deficiency in Gunn rats. Naked DNA gene transfer [6], genetically modified fibroblasts [7], and viral vectors based on SV40 [8], adenovirus [9– 11], murine retrovirus [12,13], and lentivirus [14,15] have been used to correct the high serum bilirubin levels in Gunn rats. However, the only gene therapy vectors that have shown promise in a clinical trial of a liver deficiency are based on adeno-associated virus (AAV) [16,17]. The vectors used in these clinical trials are based on AAV serotype 2 and although these vectors have been shown to be safe and well tolerated [16], a drawback is their relative inefficiency for liver gene transfer. Studies in mice have shown that, even at high vector doses, AAV2 can transduce maximally 5–10% of hepatocytes [18]. More recently, the identification of novel AAV serotypes made it possible to construct AAV vectors with improved

1085

ARTICLE

transduction properties. Of these novel serotypes, AAV1 [19], AAV6 [20], and AAV8 [21] have been described as being more efficient for liver gene transfer [22–25] in mice. The aim of this study was to evaluate the efficiency of these different AAV serotypes in the Gunn rat model of the human metabolic liver deficiency CN.

RESULTS Construction of AAV Vectors We constructed an AAV vector in which the UGT1A1 cDNA was under control of the CMV promoter, CMVUGT1A1. This vector also includes the woodchuck hepatitis PRE, which has been shown to increase expression levels in AAV vectors [26]. We constructed a liver-specific AAV vector (ALB-UGT1A1) by using the albumin promoter enhancer region [27] to drive UGT1A1 expression. Recombinant AAV particles were made with Rep from AAV2 and Cap from serotypes 1, 2, 6, and 8. With ALBUGT1A1 we generated only the AAV serotype 1. In this

doi:10.1016/j.ymthe.2006.01.014

paper we designate the different AAV serotypes by the origin of their Cap genes only. To confirm the generation of infectious AAV particles, we used the different serotypes to transduce 293T cells and a hepatoma cell line, HepG2, which has no endogenous UGT1A1 activity. After infection with adenovirus, we harvested the cells and confirmed expression of UGT1A1 by immunohistochemistry and Western blotting (not shown). Correction of Serum Bilirubin Levels Injection of all UGT1A1 AAV pseudotypes lowered serum bilirubin levels (Fig. 1). To be able to determine accurately the relative efficiencies of the different CMVUGT1A1 AAV serotypes we compared the average serum bilirubin levels over periods of 1 to 26, 27 to 54, and 55 to 82 weeks after injection. We used analysis of variance using a mixed linear model to determine statistical significances. This analysis showed that AAV1 is the most efficient, followed by AAV8, AAV6, and AAV2 (Table 1). All AAV serotypes lowered serum bilirubin levels significantly for at least 26 weeks (Table 1).

FIG. 1. Correction of serum bilirubin by injection of AAV vectors. Male Gunn rats were injected with CMV-UGT1A1 vectors and serum bilirubin was measured. Dashed lines in each graph represent the same group of control or sham-operated rats, N = 14. The solid lines show the serum bilirubin concentrations of animals injected with the indicated AAV serotypes. Each AAV-treated group consisted of five animals up to 1 year, later time points are from fewer than five animals. All serotypes lowered serum bilirubin significantly over this period. Data shown are mean and standard deviation per time point.

1086

MOLECULAR THERAPY Vol. 13, No. 6, June 2006 Copyright C The American Society of Gene Therapy

ARTICLE

doi:10.1016/j.ymthe.2006.01.014

TABLE 1: Reduction in serum bilirubin in Gunn rats by AAV UGT1A1 gene transfer 1 to 26 weeks Control rats AAV1 AAV2 AAV6 AAV8

113 F 27 AM 33 F 12 AM, 89 F 18 AM, 82 F 23 AM, 69 F 23 AM,

P P P P

b = = b

Serum bilirubin, P value versus control rats 27 to 54 weeks

0.001 0.015 0.003 0.001

120 F 24 AM 52 F 15 AM, P 106 F 19 AM, P 88 F 17 AM, P 88 F 20 AM, P

b 0.001 = 0.094 = 0.001 = 0.001

55 to 82 weeks 131 F 17 AM 66 F 10 AM, P b 0.001 131 F 27 AM, P = 0.480 ND 107 F 18 AM, P = 0.004

Average serum bilirubin was calculated over periods of 1 to 26, 27 to 54, and 55 to 82 weeks after injection. Each AAV-treated group consisted of five animals until 1 year, later time points are from fewer than five animals. A mixed linear model analysis of variance was used to calculate P values. ND, not determined.

A significant reduction in serum bilirubin was sustained for up to 82 weeks with CMV-UGT1A1 AAV1 and AAV8. Rats injected with AAV6 were followed for only 54 weeks. Serum bilirubin levels of control rats increased in time and this trend was also observed in the AAV-treated rats. However, in AAV2- and AAV8-treated rats the magnitude of the therapeutic effect decreased at the later time points, with loss of significance for AAV2 at 27 weeks (Fig. 1, Table 1).

we observed the most intense band in liver for all serotypes. Because spleen also gave a strong signal with all serotypes except AAV6, we analyzed DNA from liver and spleen further by quantitative real-time PCR. AAV1 and AAV6 gave widespread transduction with a positive signal detected in most organs. Interestingly, AAV2 and AAV8 seem to be more specific, with the highest gene transfer observed in liver and spleen and, in the case of AAV2, duodenum.

Biodistribution of AAV Gene Transfer We determined AAV vector distribution by isolation of high-molecular-weight DNA and PCR to detect specifically the human UGT1A1 sequence (Fig. 2). As expected,

Quantitative Determination of AAV Gene Transfer by PCR We determined the transduction efficiency of liver and spleen by detection of the AAV vector by quantitative real-time PCR in genomic DNA from these tissues. The results are averages of tissues from four animals and summarized in Table 2. Values shown are AAV copies present per haploid genome. AAV1-transduced liver had the highest number of viral genomes per cell, which correlates well with this serotype also being the most efficient for the correction of UGT1A1 deficiency. AAV2 and AAV6 had the least amount of viral genomes per cell, which agrees with the inefficient correction of hyperbilirubinemia observed with these serotypes. Transduction of spleen by AAV vectors was similar for all serotypes but at least 1 order of magnitude lower than transduction of liver. Bile Analysis Confirms Functional UGT1A1 Gene Transfer We collected bile from Gunn rats injected with the AAV UGT1A1 serotypes 12 months after injection of AAV6 TABLE 2: Quantitative PCR determination of AAV gene transfer in liver and spleen

FIG. 2. Biodistribution of AAV-mediated UGT1A1 gene transfer. Highmolecular-weight DNA was isolated between 14 and 20 months after administration of AAV vectors for serotypes 1, 2, and 8 and at 12 months for serotype 6. DNA samples were analyzed for the presence of human UGT1A1 sequence by PCR as described. As loading control endogenous GAPDH was amplified. M, marker; Li, liver; Lu, lung; Ki, kidney; Sp, spleen; St, stomach; Du, duodenum; Il, ileum; Je, jejunum; Co, colon; Pa, pancreas; He, heart; Mu, skeletal muscle; Te, testis; Th, thymus; Br, brain.

MOLECULAR THERAPY Vol. 13, No. 6, June 2006 Copyright C The American Society of Gene Therapy

Serotype AAV-1 AAV-2 AAV-6 AAV-8

Liver 1.5 F 1 0.069 F 0.05 0.046 F 0.03 0.23 F 0.7

Spleen 3.0 F 3  10 3 14 F 30  10 3 0.028 F 0.01  10 12 F 20  10 3

3

The number of AAV vector copies per haploid genome as determined by quantitative realtime PCR is shown. Values are averages of four animals.

1087

ARTICLE

doi:10.1016/j.ymthe.2006.01.014

nides were detected in bile of all AAV-treated animals and were absent in control rat bile (Fig. 3). This shows that functional UGT1A1 is expressed after injection of all AAV serotype UGT1A1 vectors. The reduction of serum bilirubin by AAV2 vectors was modest but the presence of small amounts of bilirubin glucuronides in bile from these animals confirms functional UGT1A1 expression by AAV2.

FIG. 3. HPLC analysis of bilirubin glucuronides in bile. Bile was collected and analyzed for the presence of bilirubin glucuronides. HPLC traces of representative bile samples from AAV 1, 2, 6, and 8 are shown. For comparison, a trace from an untreated control Gunn rat is also shown. BDG, bilirubin diglucuronide; BMG, bilirubin monoglucuronide; UCB, unconjugated bilirubin. Bilirubin glucuronides are present only in bile from rats injected with CMV AAV vectors. Bile from control rats contains only unconjugated bilirubin.

and 14 to 22 months after injection of AAV1, AAV2, and AAV8 and analyzed it for the presence of bilirubin glucuronides by reverse-phase HPLC. Bilirubin glucuro-

Histological Demonstration of AAV-Mediated Gene Transfer to the Liver As described before [14], we were not successful in detecting UGT1A1 by immunohistochemistry in Gunn rats. We therefore injected Gunn rats with GFP AAV vectors to be able to demonstrate histologically AAV gene transfer (Fig. 4). Most of the GFP-positive cells were hepatocytes, as determined by their characteristic morphology. GFP-positive hepatocytes were detected in livers from rats injected with all AAV serotypes. As expected, the GFP-positive hepatocytes were most abundant in the AAV1-injected animals. However, the number of transduced hepatocytes was low. Counting of

FIG. 4. Histological demonstration of AAV gene transfer in liver. Gunn rats were injected with GFP AAV vectors and GFP was detected by immunohistochemistry. The different images show that GFPpositive hepatocytes (darkly stained cells) are present in animals injected with all AAV serotypes. Sections from AAV-injected rat livers are shown at original magnification of 10. To show accurately the specificity of the staining, a section of an untreated control rat liver is shown at original magnification of 5.

1088

MOLECULAR THERAPY Vol. 13, No. 6, June 2006 Copyright C The American Society of Gene Therapy

ARTICLE

doi:10.1016/j.ymthe.2006.01.014

three random fields of hepatocytes revealed that the percentage of transduced hepatocytes with all AAV serotypes was below 1%. Immune Response to UGT1A1 Because we have observed that UGT1A1 antibodies were generated after lentiviral UGT1A1 gene transfer [28], we investigated whether AAV gene transfer also induced an immune response to UGT1A1. Using an ELISA, we detected antibodies to UGT1A1 in animals injected with all serotypes of AAV CMV-UGT1A1 (Fig. 5). We confirmed the specificity of the UGT1A1 antibodies by Western blotting (not shown). To investigate the mechanism of the immune response to UGT1A1, we constructed an AAV1 UGT1A1 vector in which the liver-specific albumin promoter was driving UGT1A1 expression (ALB-UGT1A1). ALB-UGT1A1 also corrected serum bilirubin levels (Fig. 5). However, animals injected with this vector did not develop antibodies to UGT1A1 (Fig. 5). In this figure the immune response to AAV1 vectors is shown but the titers of antibodies to UGT1A1 that developed after injection of AAV serotypes 2, 6, and 8 were similar (not shown). Histology of Gunn Rat Livers Two years after the start of the experiment, we sacrificed the animals. Autopsies revealed the presence of large white nodules on the livers of all AAV-injected animals (Fig. 6A). Although age-matched sham-operated or untreated control rats have areas of fatty liver, the large nodules seen in the AAV-treated animals were absent. Histology on paraffin-embedded tissue showed that these nodules were large structures resembling fat deposits (Fig. 6C). Oil Red O staining on cryosections confirmed that these nodules are indeed lipid deposits (Figs. 6E and 6F). The lesions contained nuclei but these did not show an abnormal amount of metaphases. In contrast, a control

rat with a spontaneous liver tumor is shown in Fig. 6D. In this section numerous metaphases were seen, indicating actively proliferating tissue. We isolated high-molecularweight DNA from the lipid lesions to investigate whether integrated AAV sequences were responsible for the generation of these structures. PCR analysis of the DNA failed to detect AAV sequences in DNA obtained from the lipid deposits (not shown).

DISCUSSION We have shown that AAV1 vectors are more efficient for the correction of the UGT1A1 deficiency in the Gunn rat than AAV2, 6, and 8. Several reports document that AAV serotypes other than 2 are more efficient for the transduction of hepatocytes. The most efficient serotype for murine liver transduction seems to be AAV8, as a recent study documents that AAV8 transduction is not restricted in mouse liver. Vectors based on AAV8 were able to transduce complete murine livers, an efficiency not obtained with any other serotype [24]. However, we find that, in rats, AAV1 is the most efficient serotype. AAV2 transduction of rat liver is relatively inefficient; injection of 5  1012 AAV genomes per kilogram resulted in an average copy number of 0.069 vector copies per haploid genome. These results are in agreement with an earlier study in which a similar AAV2 transduction efficiency was obtained in rat liver [29]. The number of AAV genomes per cell exceeds the percentage of transduced cells that we detected by counting positive cells in immunohistochemistry of transduced livers. This discrepancy is explained by the property of AAV vectors to form large concatemers of viral copies after a productive transduction event [30]. Our study confirms that AAV serotypes other than 2 are more efficient for liver gene transfer. Tropism of AAV

FIG. 5. Use of a liver-specific promoter abrogates immune response to UGT1A1. Male Gunn rats were injected with CMV-UGT1A1 and ALB-UGT1A1 vectors. Serum bilirubin and antibody titers were measured as described. Both AAV1 ALB-UGT1A1 (N = 2) and CMV-UGT1A1 (N = 5) reduce serum bilirubin levels in Gunn rats as shown on the left. The dashed line represents control animals (N = 14). On the right, titrations of sera from ALB-UGT1A1- (N = 2) and CMVUGT1A1- (N = 2) injected animals are shown. A strong immunoreactivity is seen in sera from CMV-UGT1A1-injected animals only. AAV1 ALB UGT1A1 administration did not lead to the formation of UGT antibodies. Titrations of sera 10 and 17 weeks after administration of AAV vector are shown.

MOLECULAR THERAPY Vol. 13, No. 6, June 2006 Copyright C The American Society of Gene Therapy

1089

ARTICLE

doi:10.1016/j.ymthe.2006.01.014

FIG. 6. Liver histology 2 years after AAV vector administration. Gunn rats were sacrificed 2 years after the start of the experiment and tissues were processed for immunohistochemistry. (A) A macroscopic image of the left median lobes from two sham-operated control rats and AAV1-, 2-, and 8-injected animals. On the surface of the AAV-injected rat livers large white spots can be seen. These spots are absent from the age-matched control livers. (B and C) Hematoxylin/ phloxin staining of sham-operated control and AAV1injected rat livers, respectively. A large lesion can be seen in the liver from the AAV1-injected rat. (D) A spontaneous liver tumor in an untreated control rat. (E and F) Oil red O staining of livers from control and AAV1-injected rats, respectively. The strong red staining in the liver from the AAV1-injected rat identifies the lesions as fatty deposits.

vector serotypes in rats seems to be different from that in mice since in our experiments AAV1 performs better than AAV8, which contrasts with the superior performance of AAV8 in mice [24]. Thus, since two closely related species show such distinct AAV tropisms, our findings underscore the importance of testing the efficiency of different AAV serotypes in human hepatocytes before starting clinical trials. In all animals injected with CMV-UGT1A1 vectors, antibodies to UGT1A1 were observed. The formation of UGT1A1 antibodies could be completely blocked by the use of an AAV vector (ALB-UGT1A1) with a liver-specific promoter. The most straightforward explanation of these results is that this immune response is caused by the transduction of antigen-presenting cells by the AAV vectors. An alternative but more unlikely mechanism for the generation of an immune response to UGT1A1 would be the transduction of nonhepatic cells with AAV vectors, which subsequently cross present UGT1A1 fragments to professional antigen-presenting cells. Because all AAV serotypes tested induced an immune response toward UGT1A1 it is likely they have comparable trop-

1090

isms for hematopoietic (antigen-presenting) cells. Indeed, quantitative PCR analysis showed that portal vein administration of AAV serotypes 1, 2, 6, and 8 resulted in equal levels of transduction to splenic cells. It is possible that the immune response to UGT1A1 is the cause of the decline in therapeutic effect we observed in AAV2- and AAV8-treated animals. However, we feel that this is not very likely because a humoral response does not usually affect expression of an intracellular protein such as UGT1A1. Furthermore, treatment of Gunn rats with ALB-UGT1A1 did not lead to the formation of antibodies but the therapeutic effect in these animals also slowly declined. In a previous study in which UGT1A1 lentiviral vectors were introduced in fetal Gunn rats, we also observed high titers of UGT1A1 antibodies, which did not preclude a long-term therapeutic effect [28]. Large macroscopic lipid lesions were observed in all AAV-treated animals. Tumorigenesis induced by AAV vectors is a controversial issue. An initial report documented a high incidence of hepatocellular carcinomas in AAV-injected mice [31]. However, a large study of 695

MOLECULAR THERAPY Vol. 13, No. 6, June 2006 Copyright C The American Society of Gene Therapy

doi:10.1016/j.ymthe.2006.01.014

mice injected with AAV vectors did not reveal an increased incidence of tumors compared to control mice [32]. Interestingly, the one AAV-treated animal with a liver tumor in this study developed a lipoma [32]. We did not observe any hepatocellular carcinomas in our AAVinjected rats. The nature of the lipid deposits observed in our experimental animals is uncertain; they did not contain abnormal amounts of dividing cells and were relatively small, which suggests that they might not be malignant. No AAV sequences were detected in the lesions which argues against a role for AAV in the generation of these structures. However, age-matched, sham-operated, control rats did contain areas of fatty liver but none of the large macroscopic fat deposits that were seen in the treated animals. At this point we do not know whether AAV vectors in general, our method of AAV vector preparation, or the transgene is responsible for the formation of the lipid deposits. Because the lipid lesions were associated with AAV vector administration, additional studies must be performed to determine whether this phenomenon is restricted to our experimental setup. We have shown that AAV1 vectors are more efficient for the transduction of rat hepatocytes in vivo than AAV2, 6, or 8. The level of UGT1A1 correction obtained in our study would be therapeutic in humans and our results therefore warrant the development of AAV gene therapy for the treatment of Crigler–Najjar disease.

MATERIALS

AND

METHODS

Construction and production of AAV vectors. An AAV vector with the CMV promoter driving UGT1A1 expression was constructed by cloning the UGT1A1 cDNA into the pTRCGW AAV vector backbone [33]. The inclusion of the woodchuck hepatitis virus posttranscriptional regulatory element in this vector ensures high expression levels. Control GFP expression vectors were also based on the pTRCGW backbone. To generate AAV with the liver-specific albumin promoter enhancer [27] driving UGT1A1 expression, UGT1A1 was inserted after the albumin promoter enhancer. This expression cassette was subsequently inserted into an AAV vector backbone with a SV40 polyadenylation signal [34]. Recombinant AAV was produced with AAV2 Rep and pseudotyped with capsids from AAV serotypes 1, 2, 6, and 8. The different AAV vectors are designated throughout this paper by the serotype of their capsid only. AAV1, 2, and 6 vectors were produced by transient transfection of the packaging constructs pDF1, 2, and 6 [35] and the AAV vector into 293T cells. AAV8 vectors were produced by transient transfection as described [21]. The average yields of the AAV serotypes were similar, between 3  1011 and 9  1011 genome copies per plate. AAV vector particles were purified by iodixanol gradient centrifugation as described [36]. Titration of AAV vectors was performed by quantitative PCR as described below. Average titers (genome copies/ml) of AAV vectors were 5.3  1012 for AAV1, 4.5  1012 for AAV2, 5.4  1012 for AAV6, and 19  1012 for AAV8. For this study two or three different virus preparations were made per AAV serotype. Qualitative and quantitative PCR. High-molecular-weight DNA from tissues and DNA from AAV vector preparations was isolated using proteinase K digestion as described [37]. In all PCRs, 250 ng of DNA was used. Qualitative PCR to determine the biodistribution of the AAV vectors was performed with primers specific for human UGT1A1 that did not

MOLECULAR THERAPY Vol. 13, No. 6, June 2006 Copyright C The American Society of Gene Therapy

ARTICLE

amplify rat UGT1A1: forward, TTCAGAGGACGTGCAGACAG; reverse, CAAGGTGGCACCTATGAAGC. Quantitative real-time PCR to determine transduction efficiency of rat tissue was performed on an ABI Prism 7000 using SYBR green to detect the amplification products. Primers were directed toward the CMV promoter: forward, ATGGGCGGTAGGCGTGTA; reverse, AGGCGATCTGACGGTTCACTAA. Sensitivity of this assay was 10 copies per reaction. Quantitative real-time PCR to determine AAV vector titer was performed on a Roche LightCycler using SYBR green detection. Primers directed at UGT1A1 were used: forward, TACACTGGAACCCGACCATC; reverse, AACAAGGGCATCATCACCA. Standard curves for real-time PCR were constructed by using dilutions of AAV vector DNA. Animal experiments. Gunn rats from our own breeding colony were used for all experiments and fed ad libitum. All animal experiments were performed in accordance with the Animal Ethical Committee guidelines of the Academic Medical Center of Amsterdam. Male Gunn rats, 6 to 8 weeks of age, 150 to 200 g, were used for all experiments. Each experimental group consisted of five animals. Rats were anesthetized with an intraperitoneal injection of KAR mix: 4 ml ketamine (100 mg/ml), 2 ml Rompun (xylazine; 20 mg/ml), 1 ml atropine (1 mg/ ml); dose of 0.1 ml/100 g body wt. Under deep anesthesia, the peritoneal cavity was opened and the rats were injected intraportally, using a 30-gauge needle, with AAV vector resuspended in a maximum volume of 500 Al. All animals received a dose of 2.5–5  1012 AAV vector genomes per kilogram. The animals were sutured and received the analgesic Temgesic subcutaneously following recovery from KAR mix. Control animals were either sham operated (opening of the peritoneal cavity and preparation of the portal vein) or untreated. For bile collection, rats were anesthetized by intraperitoneal injection of KAR mix as above and bile was collected by cannulation of the bile duct as described [14]. Blood was collected by tail vein puncture under gas anesthesia in pediatric heparin tubes. Histology. GFP was detected by immunohistochemistry as described [14]. Hematoxylin and azaphloxin staining was performed on formaldehyde and Paraplast-embedded tissue sections as described [14]. Oil red O staining was performed on formaldehyde-fixed cryosections as described [38]. ELISA for UGT1A1 antibodies. UGT1A1 and GFP were expressed in 293T cells by calcium phosphate coprecipitation. Expression of UGT1A1 was confirmed by Western blotting and GFP expression was confirmed by fluorescence microscopy. The cells were harvested with 5 mM EDTA in PBS, concentrated by centrifugation, and lysed by sonication. Protein was determined by Bio-Rad Bradford assay. ELISA plates (Nunc) were coated overnight with 5 Ag cellular protein per well in 96-well plates in 50 mM carbonate buffer, pH 9.6. The wells were blocked with 1% gelatin in phosphate-buffered saline, washed, and incubated with serial dilutions of Gunn rat plasma. After being washed the bound Gunn rat immunoglobulins were detected with anti-rat IgG peroxidase (Nordic) and o-phenylenediamine tablets (Sigma). Color development was measured at 490 nm in an ELISA reader. ELISAs were always performed in duplicate with the same samples applied on UGT1A1- and GFP-coated plates. The staining in the GFP plate was subtracted from the staining in the UGT1A1 plate to correct for background binding. Bilirubin quantification. Unconjugated bilirubin and bilirubin conjugates in bile were analyzed and quantified by HPLC as described [7] with the modification that an Omnisphere column (Varian, The Netherlands) was used. Bilirubin in serum was quantified by the hospital Routine Clinical Chemistry Department. Statistics. Average serum bilirubin levels and standard deviations per treatment group were calculated for every time point during 1 year. A mixed linear model analysis of variance (SPSS version 11.5) was used to test the differences between the treatment groups.

1091

ARTICLE

ACKNOWLEDGMENTS We thank Dr. Jan Ruijter for help with the statistical analysis and Lizzy Comijn and Sanne Derks of Amsterdam Molecular Therapeutics for producing AAV-GFP batches and for performing Q-PCR analysis of tissue samples. Dr. J. Kleinschmidt of the Deutsches Krebsforschungszentrum is acknowledged for providing the production systems for AAV-1, 2, and 6. Dr. J. M. Wilson and the Vector Core Facility of the University of Pennsylvania are acknowledged for providing the AAV-8 production system. This research was made possible by grants from the Dutch Crigler–Najjar Foundation and the Dutch Organization for Scientific Research (NWO) (016.026.012). RECEIVED FOR PUBLICATION NOVEMBER 16, 2005; REVISED JANUARY 25, 2006; ACCEPTED JANUARY 26, 2006.

REFERENCES 1. Ostrow, J. D., Mukerjee, P., and Tiribelli, C. (1994). Structure and binding of unconjugated bilirubin: relevance for physiological and pathophysiological function. J. Lipid Res. 35: 1715 – 1737. 2. Tukey, R. H., and Strassburg, C. P. (2000). Human UDP-glucuronosyltransferases: metabolism, expression, and disease. Annu. Rev. Pharmacol. Toxicol. 40: 581 – 616. 3. Jansen, P. L. (1999). Diagnosis and management of Crigler–Najjar syndrome. Eur. J. Pediatr. 158(Suppl. 2): S89 – S94. 4. Ostrow, J. D., Pascolo, L., and Tiribelli, C. (2002). Mechanisms of bilirubin neurotoxicity. Hepatology 35: 1277 – 1280. 5. Chowdhury, J. R., Kondapalli, R., and Chowdhury, N. R. (1993). Gunn rat: a model for inherited deficiency of bilirubin glucuronidation. Adv. Vet. Sci. Comp. Med. 37: 149 – 173. 6. Danko, I., Jia, Z., and Zhang, G. F. (2004). Nonviral gene transfer into liver and muscle for treatment of hyperbilirubinemia in the Gunn rat. Hum. Gene Ther. 15: 1279 – 1286. 7. Seppen, J., et al. (1997). Transplantation of Gunn rats with autologous fibroblasts expressing bilirubin UDP-glucuronosyltransferase: correction of genetic deficiency and tumor formation. Hum. Gene Ther. 8: 27 – 36. 8. Sauter, B. V., et al. (2000). A replication-deficient rSV40 mediates liver-directed gene transfer and a long-term amelioration of jaundice in Gunn rats. Gastroenterology 119: 1348 – 1357. 9. Askari, F. K., Hitomi, Y., Mao, M., and Wilson, J. M. (1996). Complete correction of hyperbilirubinemia in the Gunn rat model of Crigler–Najjar syndrome type I following transient in vivo adenovirus-mediated expression of human bilirubin UDP-glucuronosyltransferase. Gene Ther. 3: 381 – 388. 10. Takahashi, M., Ilan, Y., Chowdhury, N. R., Guida, J., Horwitz, M., and Chowdhury, J. R. (1996). Long term correction of bilirubin–UDP-glucuronosyltransferase deficiency in Gunn rats by administration of a recombinant adenovirus during the neonatal period. J. Biol. Chem. 271: 26536 – 26542. 11. Toietta, G., et al. (2005). Lifelong elimination of hyperbilirubinemia in the Gunn rat with a single injection of helper-dependent adenoviral vector. Proc. Natl. Acad. Sci. USA 102: 3930 – 3935. 12. Bellodi-Privato, M., Aubert, D., Pichard, V., Myara, A., Trivin, F., and Ferry, N. (2005). Successful gene therapy of the Gunn rat by in vivo neonatal hepatic gene transfer using murine oncoretroviral vectors. Hepatology 42: 431 – 438. 13. Tada, K., et al. (1998). Long-term reduction of serum bilirubin levels in Gunn rats by retroviral gene transfer in vivo. Liver Transplant. Surg. 4: 78 – 88. 14. Seppen, J., van der Rijt, R., Looije, N., van Til, N. P., Lamers, W. H., and Oude, R. P. (2003). Elferink, Long-term correction of bilirubin UDPglucuronyltransferase deficiency in rats by in utero lentiviral gene transfer. Mol. Ther. 8: 593 – 599. 15. Nguyen, T. H., et al. (2005). Therapeutic lentivirus-mediated neonatal in vivo gene therapy in hyperbilirubinemic Gunn rats. Mol. Ther. 12: 852 – 859. 16. Carter, B. J. (2005). Adeno-associated virus vectors in clinical trials. Hum. Gene Ther. 16: 541 – 550. 17. Kay, M. A., et al. (2000). Evidence for gene transfer and expression of factor IX in haemophilia B patients treated with an AAV vector. Nat. Genet. 24: 257 – 261.

1092

doi:10.1016/j.ymthe.2006.01.014

18. Nakai, H., et al. (2002). A limited number of transducible hepatocytes restricts a widerange linear vector dose response in recombinant adeno-associated virus-mediated liver transduction. J. Virol. 76: 11343 – 11349. 19. Xiao, W. D., Chirmule, N., Berta, S. C., McCullough, B., Gao, G. P., and Wilson, J. M. (1999). Gene therapy vectors based on adeno-associated virus type 1. J. Virol. 73: 3994 – 4003. 20. Rutledge, E. A., Halbert, C. L., and Russell, D. W. (1998). Infectious clones and vectors derived from adeno-associated virus (AAV) serotypes other than AAV type 2. J. Virol. 72: 309 – 319. 21. Gao, G. P., Alvira, M. R., Wang, L. L., Calcedo, R., Johnston, J., and Wilson, J. M. (2002). Novel adeno-associated viruses from rhesus monkeys as vectors for human gene therapy. Proc. Natl. Acad. Sci. USA 99: 11854 – 11859. 22. Davidoff, A. M., et al. (2005). Comparison of the ability of adeno-associated viral vectors pseudotyped with serotype 2, 5, and 8 capsid proteins to mediate efficient transduction of the liver in murine and nonhuman primate models. Mol. Ther. 11: 875 – 888. 23. Grimm, D., et al. (2003). Preclinical in vivo evaluation of pseudotyped adenoassociated virus vectors for liver gene therapy. Blood 102: 2412 – 2419. 24. Nakai, H., Fuess, S., Storm, T. A., Muramatsu, S., Nara, Y., and Kay, M. A. (2005). Unrestricted hepatocyte transduction with adeno-associated virus serotype 8 vectors in mice. J. Virol. 79: 214 – 224. 25. Rabinowitz, J. E., et al. (2002). Cross-packaging of a single adeno-associated virus (AAV) type 2 vector genome into multiple AAV serotypes enables transduction with broad specificity. J. Virol. 76: 791 – 801. 26. Loeb, J. E., Cordier, W. S., Harris, M. E., Weitzman, M. D., and Hope, T. J. (1999). Enhanced expression of transgenes from adeno-associated virus vectors with the woodchuck hepatitis virus posttranscriptional regulatory element: implications for gene therapy. Hum. Gene Ther. 10: 2295 – 2305. 27. Pinkert, C. A., Ornitz, D. M., Brinster, R. L., and Palmiter, R. D. (1987). An albumin enhancer located 10 kb upstream functions along with its promoter to direct efficient, liver-specific expression in transgenic mice. Genes Dev. 1: 268 – 276. 28. Seppen, J., van Til, N.P., van der Rijt, R., Hiralall, J.K., Kunne, C., and Oude-Elferink, R.P.J. (2006). Immune response to lentiviral bilirubin UDP-glucuronosyltransferase gene transfer in fetal and neonatal rats. Gene Ther., (electronic publication ahead of print). 29. Ohashi, K., Nakai, H., Couto, L. B., and Kay, M. A. (2005). Modified infusion procedures affect recombinant adeno-associated virus vector type 2 transduction in the liver. Hum. Gene Ther. 16: 299 – 306. 30. Vincent-Lacaze, N., Snyder, R. O., Gluzman, R., Bohl, D., Lagarde, C., and Danos, O. (1999). Structure of adeno-associated virus vector DNA following transduction of the skeletal muscle. J. Virol. 73: 1949 – 1955. 31. Donsante, A., et al. (2001). Observed incidence of tumorigenesis in long-term rodent studies of rAAV vectors. Gene Ther. 8: 1343 – 1346. 32. Bell, P., et al. (2005). No evidence for tumorigenesis of AAV vectors in a large-scale study in mice. Mol. Ther. 12: 299 – 306. 33. Ross, C. J., et al. (2004). Long-term correction of murine lipoprotein lipase deficiency with AAV1-mediated gene transfer of the naturally occurring LPL(S447X) beneficial mutation. Hum. Gene Ther. 15: 906 – 919. 34. Schimmenti, S., Boesen, J., Claassen, E. A., Valerio, D., and Einerhand, M. P. (1998). Long-term genetic modification of rhesus monkey hematopoietic cells following transplantation of adenoassociated virus vector-transduced CD34+ cells. Hum. Gene Ther. 9: 2727 – 2734. 35. Grimm, D., Kay, M. A., and Kleinschmidt, J. A. (2003). Helper virus-free, optically controllable, and two-plasmid-based production of adeno-associated virus vectors of serotypes 1 to 6. Mol. Ther. 7: 839 – 850. 36. Hermens, W. T., et al. (1999). Purification of recombinant adeno-associated virus by iodixanol gradient ultracentrifugation allows rapid and reproducible preparation of vector stocks for gene transfer in the nervous system. Hum. Gene Ther. 10: 1885 – 1891. 37. Maniatis, T., Fritsch, E. F., and Sambrook, J. (1989). Molecular Cloning: A Laboratory Manual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 38. Koopman, R., Schaart, G., and Hesselink, M. K. (2001). Optimisation of oil red O staining permits combination with immunofluorescence and automated quantification of lipids. Histochem. Cell Biol. 116: 63 – 68.

MOLECULAR THERAPY Vol. 13, No. 6, June 2006 Copyright C The American Society of Gene Therapy