Introduction

Hemophilia B, the deficiency in clotting factor IX (FIX), is an X-linked recessive disorder that occurs in about one in 25 000 males in the US. Of these patients, approximately 40% are characterized as having 'severe' hemophilia with serum levels of FIX below 100 ng/ml. Although the use of plasma-derived or recombinant FIX has greatly extended the lifespan of these patients, they remain afflicted by a variety of sequellae of the disease including retinopathies, joint dysfunction, vascular occlusive disease and internal bleeding events that can be life threatening. Moreover, through the use of human plasma-derived coagulation factors, patients have been inadvertently infected with a variety of infectious agents including hepatitis B virus, hepatitis C virus and human immunodeficiency virus. The development of recombinant FIX was considered a breakthrough in treatment and has virtually eliminated the possibility of transmitting infectious agents; however, the use of recombinant FIX does not preclude adverse events in recipients. In addition, the half-life and cost of these products limit their use to treatment on demand, medical emergencies and surgical necessity. The use of recombinant adeno-associated virus (rAAV) vectors to deliver the human FIX gene offers the opportunity to achieve continuous, stable therapeutic levels of FIX in hemophilia patients and the potential to avoid problems associated with replacement therapy.

The stable expression of human FIX protein in mice and canine FIX protein in hemophilic B dogs has been demonstrated following gene transfer with rAAV vectors derived from serotype 2^1,2,3,4,5,6 and more recently with vectors derived from alternative serotypes.⁷ Hepatic portal vein (HPV) delivery of rAAV FIX vectors in mice, non-human primates and hemophilic B dogs is well tolerated without significant increases in markers of toxicity, no induction of anti-FIX inhibitors or non-neutralizing antibodies.^2,5,6,8 Although vector could be detected in the circulation immediately following administration, there has been no evidence of germline transmission. In murine models of hemophilia B, rAAV gene transfer of human FIX corrected the clotting deficiency as determined by functional correction of the bleeding disorder. In hemophilia B dogs, rAAV canine FIX gene therapy achieved partial correction of the bleeding disorder. Successful results in animal models of hemophilia B have prompted the initiation of human clinical trials using rAAV vectors encoding human FIX.⁹ Initial trials used intramusclar (i.m.)-based rAAV vector administration; however, i.m. administration in animal models has yielded low circulating FIX levels and induced a robust anti-FIX antibody response in some immunocompetent hosts.^10,11,12 Preclinical studies in which gene delivery was targeted to the liver, the natural site of FIX production, resulted in higher levels of FIX expression with few immunological complications. Currently, direct rAAV delivery to the liver via a catheter to the hepatic artery is being investigated in human trials.¹³

This study describes the evaluation of an AAV vector for the expression of FIX with the aim of developing a vector that can be delivered intravenously (i.v.). Although portal vein administration of rAAV vectors appears to be the optimal route of administration to achieve gene transfer to the liver, avoidance of such invasive procedures in hemophiliac patients is highly desirable. We used a liver-specific expression cassette for FIX expression, the rationale being that the use of a tissue-specific promoter/enhancer offers the potential for greater safety by ensuring that gene expression will be occurring in the appropriate cell types.¹⁰ This is particularly important for a vector delivered systemically where numerous tissues could be transduced. Previous studies indicated that lower levels of transgene expression were seen following i.v. delivery compared to HPV delivery.^11,14,15 In order to achieve higher levels of gene expression, we took a systematic approach to evaluate cis-acting regulatory sequences, such as an intron and a post-transcriptional regulatory element (PRE). Introns have previously been demonstrated to potentiate gene expression by various mechanisms including increased mRNA stability and transport.^16,17,18 Rodriguez et al¹⁹ and Miao et al²⁰ demonstrated that the addition of a truncated human FIX intron 1 to the FIX-coding sequence greatly increased gene expression in vitro and in vivo, respectively. PREs such as that derived from the woodchuck hepatitis virus²¹ have also been incorporated into transgene expression cassettes to optimize protein production. The woodchuck post-transcriptional regulatory element (WPRE) increases transgene expression from retroviral,²² and AAV²³ vectors through mechanisms that affect mRNA processing, including the enhancement of polyadenylation, mRNA export and the inhibition of splicing.^24,25 Here, we investigate a side-by-side comparison of these regulatory elements and show that an intron has the greatest impact on transgene expression in these vectors. Once a strong liver-specific expression cassette was identified, we compared i.v. and HPV delivery of the same vector. We demonstrate that i.v. delivery yielded expression levels that were 70–80% of those achieved from HPV delivery. Finally, expanding on the results seen in mice, we demonstrate for the first time canine FIX expression following i.v. delivery of AAV to a canine model of hemophilia B.

Results

Selection of an AAV-2 expression cassette for the expression of human FIX

Our previous studies using an AAV-2 vector for the expression of human FIX in animal models of hemophilia B^2,3 used the vector rAAV-MFG-hFIX (Figure 1a). This vector employs a modified Maloney murine leukemia virus (MMLV) retroviral LTR as its enhancer promoter and, as such, gene expression is not tissue specific. In order to allow i.v. delivery of an AAV vector encoding FIX, we modified the expression cassette to restrict expression of FIX to the liver and we improved transgene expression to overcome the less efficient transduction efficiency seen with i.v. delivery compared to HPV delivery. Our rationale for a liver-specific expression cassette was to limit expression to cells that normally express FIX and to reduce the potential of inducing an anti-FIX immune response. The liver-specific promoter (LSP) originally described by Ill et al²⁶ is a chimeric promoter comprised of two copies of the 1-microglobulin/bikunin enhancer element placed upstream of a minimal thyroid hormone-binding globulin promoter and followed by an optimized leader sequence. The LSP promoter was placed upstream of human FIX to create the vector rAAV-LSP-hFIX (Figure 1b). In order to evaluate the contributions of different regulatory elements in an optimized vector, we constructed the vectors shown in Figure 1c–e. An intron derived from the human -globin intervening sequence II (IVS-II) sequence was placed within the optimized leader sequence to create rAAV-LSP-gb-hFIX. In a separate vector, the WPRE was inserted between the 3' terminus of the human FIX transgene and bovine growth hormone polyadenylation sequence producing rAAV-LSP-hFIX-WPRE. Finally, both the -globin intron and the WPRE element were introduced into the same vector rAAV-LSP-gb-hFIX-WPRE.

Figure 1.

rAAV vectors encoding human FIX. (a) rAAV-MFG-hFIX, encodes the human FIX cDNA (hFIX) under the control of a modified MMLV retroviral LTR promoter (MFG) linked to a Maloney murine leukemia virus intron (MuLV IVS), with a bovine growth hormone polyadenylation sequence 3' (bGHpA). (b) rAAV-LSP-hFIX, encodes the hFIX cDNA under the control of an LSP and (c) rAAV-LSP-gb-hFIX, with the human -globin intron II (gb IVS-II) inserted between the LSP promoter and human FIX cDNA. (d) rAAV-LSP-hFIX-WPRE, with the WPRE inserted 3' of the hFIX cDNA and (e) rAAV-LSP-gb-hFIX-WPRE, a combination of (c) and (d) constructs. ITRs represent the AAV-2 inverted terminal repeats.

Full figure and legend (88K)

Expression of human FIX in vivo

The impact of these regulatory elements on human FIX expression in vivo was examined following rAAV administration to C57Bl/6 mice. Vector was delivered to the liver via the HPV at a dose of 3.0 10¹¹ vector genomes (vgs) per animal. Expression of FIX from the rAAV constructs increased steadily following the first 5 weeks postinjection (p.i.), then leveled off to a steady state (Figure 2). Animals that received the rAAV-MFG-hFIX vector expressed 103 ng/ml serum FIX at week 5. Mice that were given the rAAV-LSP-hFIX vector without additional regulatory elements expressed 5 ng/ml serum FIX at the same time point. The addition of the -globin IVS-II sequence and separately the WPRE element to this vector increased expression approximately 85-fold (430 ng/ml) and 25-fold (138 ng/ml), respectively. Constructs that contained both the -globin IVS-II and the WPRE resulted in the strongest expression in vivo, with serum FIX levels reaching 516 ng/ml, although this increase in the expression level was statistically insignificant when compared to the -globin IVS-II alone vector. These levels of transgene expression remained stable for more than 90 weeks (data not shown).

Figure 2.

Transgene expression following HPV administration of rAAV vectors encoding human FIX in C57Bl/6 mice. Mice (n=5) were administered 3.0 10¹¹ vector particles per animal of one of the following constructs – rAAV-LSP-hFIX (filled diamonds), rAAV-LSP-gb-hFIX (filled triangles), rAAV-LSP-hFIX-WPRE (open circles), rAAV-LSP-gb-hFIX-WPRE (open squares), rAAV-MFG-hFIX (crosses) or PBS (filled squares) and serum human FIX monitored over time. Each data time point represents the meanss.e.m.

Full figure and legend (28K)

Liver specificity of rAAV-LSP-gb-hFIX-WPRE transgene expression in vitro

To confirm liver-specific expression from the rAAV-LSP-gb-hFIX-WPRE vector construct, FIX levels were measured in vitro following infection of a variety of cell types (Figure 3). Transgene expression under the LSP promoter was demonstrated in the hepatoma cell lines HuH7 (human hepatoma) and HepG2 (human hepatocellular carcinoma liver), but not in nonliver-derived cell lines, including HEK293 (human kidney/neural progenitor), HeLa (human adenocarcinoma cervix), A549 (human lung carcinoma), Cos-1 (monkey kidney) or RF/6A (Rhesus monkey choroid retina endothelial). FIX transgene expression was observed in all the cell lines that were transduced with the rAAV-MFG-FIX vector. Although FIX expression under the MFG promoter was greater in the hepatoma cell lines than the LSP promoter, expression from the latter promoter was clearly liver specific.

Figure 3.

Liver specificity of rAAV-LSP-gb-hFIX-WPRE transgene expression in cell lines. Cells were infected at a concentration of 1 10⁵ vg/cell with either rAAV-MFG-hFIX (filled bars) or rAAV-LSP-gb-hFIX-WPRE (lightly shaded bars) viral vectors and human FIX secretion monitored for a 48 h time period. Secretion levels were normalized for transduction efficiency with a GFP-encoding rAAV vector. ^*Indicates <0.05 showing a significant difference in human FIX secretion comparing rAAV-LSP-gb-hFIX-WPRE to rAAV-MFG-hFIX groups. Data are presented as meanss.e.m.

Full figure and legend (55K)

Influence of administration route on rAAV-mediated human FIX expression in vivo

We examined the efficacy of i.v. (tail vein) administration of rAAV vector as a less-invasive route than HPV injection for achieving transgene expression. C57Bl/6 mice were administered rAAV-LSP-gb-hFIX-PRE or rAAV-MFG-hFIX vector at a dose of 2 10¹¹ vg/animal by both routes of administration (Figure 4). Delivery of rAAV-LSP-gb-hFIX-PRE by portal vein resulted in serum FIX levels greater than 400 ng/ml, whereas expression following i.v. administration resulted in serum FIX levels of 300–400 ng/ml or about 70–80% that seen with HPV delivery. The same trend was observed for the rAAV-MFG-hFIX construct, with i.v. administration giving 70–80% the level of transgene expression observed following HPV administration. These levels of FIX expression remained stable for up to 60 weeks p.i. (data not shown). We also examined FIX expression following direct liver injection of vector. Expression levels were 25% of that achieved by HPV administration but were highly variable between animals (data not shown).

Figure 4.

Effect of administration route on rAAV-mediated human FIX expression in C57Bl/6 mice. rAAV-LSP-gb-hFIX-WPRE and rAAV-MFG-hFIX vector at a dose of 2 10¹¹ vg/animal (n=4) were administered by either HPV (crosses and open triangles, respectively) or by i.v. injection via tail vein (filled squares and open circles, respectively) and human FIX expression monitored over time. PBS controls for both i.v. (filled diamonds) and HPV (filled triangles) are included. Data are presented as meanss.e.m.

Full figure and legend (23K)

rAAV vector biodistribution and FIX immunohistochemistry following delivery in vivo

Biodistribution of vector transduction was determined by analyses of tissue for the presence of the human FIX-coding region using quantitative PCR following rAAV administration by the HPV and tail vein (Figure 5). Tissues analyzed included liver, lung, heart, kidney, spleen, pancreas, duodenum, quadriceps, skin and ovaries 10 weeks following vector injection. The primary target organ for transduction in all animals was the liver with about 1.3 copies per cell detected. Small numbers of viral genomes (<0.1 copies/cell) were also detected in the spleen following HPV but not i.v. administration. Following i.v. injection, very low but detectable amounts of rAAV viral genomes were detected in the lung and heart; however, the amount detected was <5% of that observed in the liver. Tissues were also examined immunohistochemically for the expression of human FIX (Figure 6). FIX immunostaining could be observed in 25.6% (12.6%) of hepatocytes of animals that were given rAAV-LSP-glob-hFIX-WPRE, whereas 5.2% (3.2%) of hepatocytes were positive for hFIX following administration of rAAV-MFG-hFIX vector. No significant differences in the transduction of hepatocytes were seen between portal vein and i.v. administration of either vector (data not shown). FIX staining was absent in the spleens of animals injected with either vector regardless of administration route. No obvious inflammation or abnormal tissue morphology was detected by light microscopy of hemotoxylin- and eosin-stained sections.

Figure 5.

Biodistribution of vgs by quantitative PCR following i.v. and HPV administration in C57Bl/6 mice. Distribution of vgs within animals (n=4) was evaluated using quantitative PCR with primer probe sets specific for the human FIX cDNA following i.v. delivery (filled bars) and HPV administration (open bars) 10 weeks postadministration of vector. PBS control injection mice levels are indicated (shaded bars). Data are presented as meanss.e.m.

Full figure and legend (15K)

Figure 6.

Immunohistochemical detection of human FIX expression in the hepatocytes in rAAV-hFIX-injected mice. Representative photomicrographs of immunohistochemical staining for hFIX (left figures) and hemotoxylin/eosin (right figures) in the liver following rAAV injection for the following experimental groups: (a and b) PBS buffer injected alone; (c and d) rAAV-LSP-gb-hFIX-WPRE injected i.v.; (e and f) rAAV-LSP-gb-hFIX-WPRE intraportally injected; (g and h) rAAV-MFG-hFIX injected i.v.; and (i and j) rAAV-MFG-hFIX injected intraportally.

Full figure and legend (937K)

Intravenous administration of rAAV-LSP-gb-hFIX-WPRE in a canine model of hemophilia B

Following the demonstration of efficacy of human FIX expression after i.v. administration of the rAAV-LSP-gb-hFIX-WPRE vector in mice, we examined this expression cassette and delivery route in hemophilia B dogs. An rAAV-LSP-gb-cFIX-WPRE expression vector encoding canine FIX was generated and demonstrated to be equivalent in expression level and cell specificity to the human FIX cassette in vitro and in vivo in C57Bl/6 mice (data not shown). rAAV-LSP-gb-cFIX-WPRE vector was infused i.v. into the forelimb vein of two hemophilia B dogs from the UNC Chapel Hill colony. The first dog treated in this experiment, F56, received 2.2 10¹² vg/kg and a second dog, G10, was injected at a log higher dose, 1.9 10¹³ vg/kg. Following vector infusion, toxicity markers remained within the normal range for both dogs (Table 1). Alkaline phosphatase was slightly elevated for both animals both before and after rAAV injection and thus considered unrelated to vector administration. The induction of a neutralizing antibody (NAB) immune response against AAV-2 was also monitored in F56 and G10 following administration of vector using an infectious tissue culture assay. Within 8 days of rAAV injection, both animals had developed a strong NAB response, which remained for the duration of the experiment (data not shown).

Table 1 - Liver enzyme levels following i.v. rAAV administration in hemophilia B dogs.

Full table

Canine FIX levels, whole blood clotting time (WBCT) and activated partial thrombosplasin time (aPPT) were monitored following injection. In dog F56, WBCT rapidly decreased from a baseline level of >60 to 23.5 min by 14 days postadministration and remained stable at an average of 24.9 min for up to 315 days postadministration (Table 2 and Figure 7). WBCT in a normal dog is approximately 8 min. Canine FIX antigen in F56 dog plasma over this period, however, remained <5 ng/ml, the limit of detection of the ELISA. aPPT levels that are only detected when FIX activity >1% normal (50 ng/ml) was not significantly changed following vector treatment (Table 2). Dog G10 also demonstrated a decrease in WBCT from 51.0 min preinjection to an average of 16.8 min following vector administration (Figure 7a). Unlike dog F56, dog G10 had measurable expression of plasma canine FIX antigen with levels steadily increasing following rAAV injection reaching a plateau at an average of 30 ng/ml (Figure 7b). Expression of cFIX has been stable for the course of the study (170 days). Despite this level of canine FIX expression, aPPT levels were unaltered in comparison to prebleed levels (Table 2). No antibodies against canine FIX as detected by ELISA were seen in either dog (data not shown).

Figure 7.

WBCT and canine FIX expression in hemophilia B canines following i.v. administration of rAAV-LSP-gb-cFIX-WPRE vector. Dogs received 2.2 10¹² (F56) and 1.9 10¹³ (G10) vg/kg via i.v. injection into the forelimb vein at day 0. Following injection, (a) WBCT and (b) plasma canine FIX antigen were monitored in F56 (filled circles) and G10 (filled squares) over time. The arrow represents coverage in dog F56 with normal canine plasma on day 14 due to a cephalic hematoma.

Full figure and legend (30K)

Table 2 - Summary of hemophilia B dogs injected i.v. with rAAV encoding cFIX.

Full table

Hemophilia B dogs from the UNC Chapel Hill colony have on average six bleeds per year that require treatment with normal canine plasma (T Nichols, unpublished data). These bleeds have been reduced in both dogs following the treatment with rAAV-LSP-gb-cFIX-WPRE vector. F56 had suffered a cephalic hematoma on day 14 postvector infusion requiring treatment (Figure 7b), but this is the only clinically obvious bleed since injection more than 1 year ago. Dog G10 has had no bleeding events during the 7 months since vector injection. Thus, both dogs have shown phenotypic improvement in the bleeding disorder from rAAV-mediated canine FIX expression following i.v. administration.

Discussion

In these studies, we examined the feasibility of i.v. delivery of an AAV-2 vector for the expression of FIX in both murine and canine models. Initially, we carried out studies to optimize the vector cassette to achieve liver-specific gene expression through the use of the LSP described by Ill et al.²⁶ In addition, we examined FIX expression in conjunction with an intron and/or WPRE. Use of the promoter alone in the construct rAAV-LSP-hFIX resulted in very low human FIX expression in C57Bl/6 mice in the absence of these cis-acting regulatory sequences. Addition of the WPRE element increased gene expression approximately 25-fold, in agreement with previous publications,^22,23 while addition of the -globin IVS-II intron sequence enhanced this level approximately 85-fold. Interestingly, the combination of both WPRE and intron in the same vector yielded no significant increase in gene expression in comparison to the vector containing only the intron, perhaps signifying a redundancy in these two elements in their ability to increase mRNA stability and transport within the host cell. Indeed, PREs are usually found in the context of genes that lack introns and are thought to bypass the requirement for splicing by specific interactions with viral or cellular factors for nuclear export and stability.²⁴ Comparison of FIX levels from the rAAV-MFG-hFIX construct previously used in animal studies^2,3 with the rAAV-LSP-gb-hFIX-WPRE vector showed a five-fold increase in expression for the liver-specific vector following i.v. and HPV administration in mice. In comparison, cell culture analysis of expression levels from the two vectors showed that while the LSP promoter showed a liver-specific expression pattern, expression from the MFG promoter/intron was higher even in those liver-derived cell lines. This divergence between cell culture expression and in vivo expression levels has been observed in a number of other vector systems including hydrodynamic-based gene delivery,²⁰ retroviral²⁷ and rAAV vectors,²³ and underscores the need for careful interpretation of vector design.

FIX expression following i.v. delivery of the rAAV-LSP-gb-hFIX-WPRE vector in mice resulted in approximately 70% of the expression levels observed from the same dose delivered via HPV. These results are in agreement with Nathwani et al,¹¹ who delivered two different rAAV expression cassettes for the expression of human FIX by tail vein and observed 60–80% of the levels of expression achieved by intraportal injection in C57Bl/6 SCID and C.B-17 SCID mice. These results are consistent with our quantitative PCR results, demonstrating that peripheral vascular administration of the rAAV-LSP-gb-hFIX-WPRE vector produces a similar pattern of vector biodistribution as portal vein delivery of the vector in mice. The liver is the overwhelmingly dominant site of gene transfer by both routes. This tropism for liver, as suggested by previous work on biodistribution,^23,28 may depend on a number of attributes of the rAAV vector interaction with the cells of the liver, including primary and secondary receptor expression, intracellular trafficking, second-strand synthesis, integration and/or concatermization, all factors known to effect transduction efficiency (reviewed in Sanlioglu et al²⁹). It should be noted that some previous studies exploring the efficacy of i.v. delivery for AAV-2 vector-mediated transduction have reported a much lower efficiency of gene transfer in comparison to more direct hepatic delivery routes.^14,15

Following the demonstration of efficacy of i.v. administration of rAAV vectors in murine studies, we explored this mode of delivery in a canine model of hemophilia B. rAAV-LSP-gb-cFIX-WPRE vector was administered to two dogs at doses of 2.2 10¹² and 1.9 10¹³ vg/kg. The less-invasive nature of this injection route represents a distinct advantage in comparison to previous studies using intraportal injection to deliver rAAV vectors encoding FIX.³ Following treatment, no signs of toxicity were observed in the animals based on observation of a panel of liver enzyme markers in addition to markers of kidney and muscle function, complete blood count and electrolytes (data not shown), demonstrating the safety of this route of administration. Similar safety profiles for AAV vectors expressing FIX have been demonstrated following intraportal and i.m. administration.^3,30 Evidence of canine FIX expression as assessed by a reduction in WBCT was clearly observed in both dogs following injection. Canine FIX levels, as determined by ELISA, were detected in the dog that received the higher AAV dose but not in the dog that received the lower dose. While these results demonstrate that i.v. administration of an rAAV vector is feasible to obtain canine FIX expression, it is interesting to compare expression levels with alternative administration routes for efficacy. An rAAV-LSP-gb-cFIX-WPRE vector that is similar to the vector employed in our studies has been previously developed and tested in both hemophilia B murine and canine models by Wang et al.^5,6 Following portal vein administration of the vector at doses of 2.8 10¹² and 4.6 10¹² vg/kg, average canine FIX levels were 31.7 and 218.1 ng/ml level, respectively, suggesting a 10-fold higher level of canine FIX expression for a given rAAV vector dose by intraportal delivery compared to i.v. delivery. However, it is difficult to make direct comparisons between our results and those of Wang et al,^5,6 because the vectors were made differently and were not compared side-by-side in the same animal study. We have previously treated two other hemophilic dogs with a rAAV-MFG-cFIX construct.³ These dogs received 1–2 10¹¹ vg/kg via the portal vein and expressed up to 18 ng/ml over the course of the experiment. Our rAAV-LSP-gb-cFIX-WPRE vector in dog F56 failed to demonstrate FIX antigen above 5 ng/ml, suggesting that i.v. delivery of AAV in dogs is less efficient than portal vein in dogs. At a higher dose, i.v. treatment of dog G10 with the AAV-cFIX vector resulted in long-term expression of cFIX and partial correction of the bleeding disorder, demonstrating the efficacy of i.v. administration in a large animal albeit at a higher vector dose.

Following injection of rAAV vector, in both dogs there was rapid increase in NAB titer that persisted for the duration of the experiment. Previous studies have also shown the induction of a strong anti-AAV response following AAV administration to mice, dogs and monkeys, suggesting that readministration of AAV serotype 2 vectors may not be possible^3,8,30 by any route of administration. The use of alternative serotypes that are immunologically distinct to AAV-2, such as AAV-8, may allow repeat administration of a vector for increased therapeutic effect.³¹ Furthermore, the use of other serotypes of AAV with differing receptor specificity and tissue tropism⁷ will be interesting to examine in the context of peripheral vein administration, particularly those that appear to demonstrate an increased transduction within the liver. Recent data have demonstrated improved efficacy of hepatic gene transfer with the AAV-8 serotype in murine models³¹ and the AAV-5 serotype in both murine¹⁵ and canine models³² of hemophilia.

In summary, we report the feasibility of using i.v. delivery of an rAAV-2 vector for the production of FIX in mice and a canine model of hemophilia B. In order to address issues of lower transduction efficiency with i.v. delivery, we evaluated the roles of different regulatory elements to increase transgene expression. Clearly, the -globin intron had the greatest impact on gene expression. Expression levels of canine FIX following i.v. administration of rAAV were lower than values that have been reported by intraportal delivery, but the reduction in WBCT and spontaneous bleed events clearly showed an improvement in the clotting abilities of these two dogs presented here. Our studies demonstrate advancement in the development of a safe and easy administration of AAV to hemophilia patients.

Materials and methods

Recombinant AAV vector construction

Standard cloning methods were used for rAAV vector construction. The construction of the plasmid pAAV-MFG-hFIX has been described previously.² Briefly, the AAV-MFG-hFIX vector contains the MMLV 5'LTR, adjacent splice donor/acceptor sequences, the human FIX cDNA sequence and the polyadenylation site of bovine growth hormone. Construction of the parental vector pAAV-LSP-hFIX, from which all the additional LSP vectors were derived, began with the plasmid pABTTBG-mFIX (a gift from Inder Verma, Salk Institute) that contains the LSP promoter driving the mouse FIX gene. pABTTBG-mFIX was cut with NcoI, NgoMI and HindIII to extract the AAV-2 5'ITR, LSP promoter and bacterial vector backbone. This fragment was then ligated with an NgoMI, BsrGI fragment, encompassing the hFIX cDNA, bovine growth hormone poly-A tail and 3'AAV ITR derived from the plasmid pSSV9-MFG-A148T-hFIX using a synthetic oligonucleotide to create the pAAV-LSP-FIX plasmid. To construct the plasmid pAAV-LSP-gb-hFIX, the plasmid pAd5tetMDRev was cut with BamHI and MluI to release the -globin intron IVS-II and the ends were blunted using Klenow fragments. The version of gb IVS-II used within these constructs has a 370 bp deletion 107 bp after the 5' splice donor. This blunted intron fragment was then introduced into a blunted BglII site present in the LSP promoter leader sequence of the plasmid pAAV-LSP-hFIX to produce pAAV-LSP-gb-hFIX. To generate the plasmid pAAV-LSP-hFIX-WPRE containing the woodchuck hepatitis virus post-transcriptional element between the 3' end of the human FIX cDNA and the 5' end of the bovine growth hormone polyadenylation sequence, the plasmid pSSV9-MFG-A148T-hFIX-WPRE was digested with BsaBI and SphI to extract a fragment of the hFIX cDNA with the WPRE element 3' and ligated into identical sites within pAAV-LSP-hFIX to create pAAV-LSP-hFIX-WPRE. The final construct pAAV-LSP-gb-hFIX-WPRE was then generated by AflII and XbaI digestion of pAAV-LSP-hFIX-WPRE and pAAV-LSP-gb-hFIX followed by ligation of selected fragments. The pAAV-LSP-gb-cFIX-WPRE construct was obtained in two stages. The first intermediary vector, pAAV-LSP-cFIX-WPRE, was constructed by ligation of a BssSI/NcoI fragment of pABTTBGmFIX, encompassing the 5'ITR and the LSP promoter, with a BamHI/BssSI fragment containing WPRE and the 3'ITR from pAAV-LSP-gb-hFIX-WPRE, linked by an NcoI/BamHI cFIX fragment from pAAV-MFG-cFIX.³ The final vector was made by inserting the BamHI/MluI fragment of pAd5tetMD-RV into a blunted BglII digest of pAAV-LSP-cFIX-WPRE. All constructs were fully sequenced.

Recombinant AAV vector preparation

Recombinant AAV vectors were prepared according to Snyder et al³³ with a few modifications. Briefly, for murine studies, subconfluent human embryonic kidney 293 cells were cotransfected with the vector plasmid and the AAV helper plasmid pUC-ACG using the calcium phosphate method. At 8 h following transfection, cells were infected with adenovirus Addl312 (an E1A^- deletion mutant) at an MOI=2 and the infection was allowed to proceed for 72 h. Following incubation, cells were harvested and lysed by three freeze/thaw cycles. Lysates were treated with benzonase for 15 min at 37°C and then centrifuged to remove the cellular debris. The cleared cell lysate was fractionated by ammonium sulfate precipitation and the rAAV virions were isolated on two sequential CsCl gradients. The gradient fractions containing rAAV were dialyzed against sterile PBS containing CaCl₂ and MgCl₂, heated for 10 min at 56°C to inactivate any residual adenovirus and stored at -80°C. Vector rAAV-LSP-gb-cFIX-WPRE used for canine studies was produced in an identical manner to the human FIX-encoding vector used for the murine studies with the exception that 293 cells were transfected in Cell factories (Nunc, Roskilde, Denmark) followed by Addl312 infection. Following production, rAAV-infected cells were harvested by centrifugation and lysed by three freeze–thaw cycles. The lysate was purified by a serial filtration, followed by column chromatography (D Frey, unpublished results) and heat inactivation at 56°C for 30 min. rAAV vector was 0.2 M filtered and stored at -70°C.

Characterization of rAAV preparations

The recombinant AAV vector particles were characterized by several criteria as described previously.² The rAAV banded with a density of 1.42 g/ml on CsCl gradients, similar to the density for wild-type AAV (wtAAV), and the rAAV preparations were stable to the heat treatment used to inactivate any remaining adenovirus. The absence of contaminating infectious adenovirus was confirmed by an assay to detect cytopathic effect. The presence of contaminating wtAAV was determined by a quantitative PCR assay with oligonucleotide primers that were specific to wtAAV and phenotypically wtAAV generated during production by recombination of helper and vector plasmids.

Determination of AAV titers

The rAAV preps were treated with DNaseI to degrade any unencapsidated DNA and then treated with proteinase K (0.25 mg/ml) in the presence of 0.5% SDS and 10 mM EDTA to liberate the rAAV genomes, followed by phenol–chloroform extraction and ethanol precipitation. Viral DNA was then denatured in alkali and applied to a nylon membrane. Dilutions of the corresponding vector plasmid were used as standards to determine the rAAV virion copy number. A radioactive probe specific for the human FIX transgene was hybridized to DNA on the filter and the filter was exposed to film followed by quantification of radioactivity by -counter (1450 Micobeta Trilux, Perkin-Elmer, Inc., Wellesley, MA, USA). Titer was also determined via quantitative PCR analysis using vector-specific FIX primers. Briefly, DNA was extracted from AAV virions as above but the phenol/chloroform and ethanol precipitation steps omitted. Samples were heated at 90°C for 10 min, serially diluted and then placed into a PCR assay. Primer and probe sequence are as follows: LSP-hFIX Fwd, 5'-GAT-AAC-AAG-AAC-GAA-ACA-ATA-ACA-GCC-3', and LSP-hFIX Rev, 5'-TCC-TAA-AAG-GCA-GAT-GGT-GAT-GA-3'; LSP-hFIX probe, 5'-6FAM-CCA-TGC-AGC-GCG-TGA-ACA-TGA-TC-TAMRA-3'. PCR was carried out under standard conditions for 40 cycles and data analyzed using Sequence Detection System version 1.6.3 (Applied Biosystems, Foster City, CA, USA).

Cells

A549, HEK 293, HuH7, Cos-1 and HeLa cell lines were grown in DMEM high-glucose medium (JRH Biosciences, Lenexa, KS, USA) supplemented with 10% fetal bovine serum (FBS) (JRH Biosciences, Lenexa, KS, USA; irradiated), 2 mM L-glutamine (JRH Biosciences, Lenexa, KS, USA) and penicillin/streptomycin (Gibco BRL, Rockville, MD, USA). HepG2 were grown in DMEM low-glucose supplemented with 10% FBS, 1% nonessential amino acids, 1 mM sodium pyruvate (JRH Biosciences, Lenexa, KS, USA), 2 mM L-glutamine and penicillin/streptomycin. RF/6A cells were grown in F12 media (JRH Biosciences, Lenexa, KS, USA) supplemented with 10% FBS, 2 mM L-glutamine and penicillin/streptomycin.

In vitro infectivity

Cells were seeded at 5 10⁵/well on six-well plates and the following day infected with purified rAAV preparations at an MOI of 1 10⁵ viral particles/cell in conjunction with the adenovirus helper virus Addl309 at an MOI of 5. Cells were incubated in the presence of 1 g/ml vitamin K for 48 h following infection before human FIX expression was determined by ELISA (see below). FIX expression levels were normalized for transduction efficiency using an rAAV-EF1á-GFP construct.

Animal studies

C57Bl/6 mice were obtained from Taconic (Germantown, NY, USA) and housed under SPF conditions. Animals were treated according to the ILAR Guide for the care and use of laboratory animals. The rAAV vectors were injected through a portal vein catheter in the case of HPV administration and via direct tail vein injection for i.v. administration. The 300 l injection volume was infused over a 30 s period. Mice were periodically bled by the retro-orbital technique. Hemophilia B dogs used in these studies were bred at the Francis Owen Blood Research Laboratory at the University of North Carolina, Chapel Hill, USA and treated according to NIH guidelines for animal care (details of dogs are presented in Table 2). Before rAAV injection, animals were placed under anesthesia using either acepromazine/torbugesic or domitor. The rAAV-LSP-gb-cFIX-WPRE vector was administered to the forelimb of the animals at a flow rate of 2 ml/min.

Immunoassays for FIX antigen

The human FIX ELISA was performed as previously described.¹⁰ Briefly, a mouse monoclonal anti-human FIX antibody (Boehringer Manneheim, Indianapolis, IN, USA) was used for capture and a horseradish peroxidase (HRP)-linked goat anti-human FIX antibody (Affinity Biologicals, Hamilton, Ontario, Canada) for detection. Canine FIX antigen was detected by capture with the monoclonal anti-human FIX antibody followed by detection with an HRP-linked sheep anti-cFIX antibody (Affinity Biologicals, Hamilton, Ontario, Canada).

AAV vector biodistribution following delivery in vivo

AAV vector biodistribution was determined using quantitative PCR analysis on specific tissues. Animals were killed and liver, duodenum, skin, lung, kidney, muscle (quadriceps), spleen, pancreas, heart and ovaries were harvested and instantly frozen on dry ice followed by storage at -80°C. For DNA extraction from the organs, 20 mg of tissue was dissected from the sample while still frozen and processed using a QIAamp DNA extraction kit (Qiagen GmbH, Hilden). Following extraction, 500 ng of tissue DNA was used as a template in a PCR reaction with human FIX specific primers and probe to quantitate the distribution of vgs following rAAV administration to mice. Primers and probe sequence are as follows: hFIX Fwd, 5'-TGA-AAC-CAT-TTT-GGA-TAA-CAT-CAC-TC-3', and hFIX Rev, 5'-GAA-TTG-ACC-TGG-TTT-GGC-ATC-T-3'; hFIX probe 5'-6FAM-CAC-CCA-ATC-ATT-TAA-TGA-CTT-CAC-TCG-GGT-T-TAMRA-3'. PCR was carried out under standard conditions for 40 cycles and data analyzed using Sequence Detection System version 1.6.3 (Applied Biosystems, Foster City, CA, USA).

Human FIX immunohistochemistry

Tissue for human FIX immunohistochemistry was processed as follows: tissue harvested from animals was instantly frozen in dry ice, sectioned on a cryostat and placed on glass slides. Tissue was then fixed in 1% PFA, endogenous peroxidase activity quenched using a peroxidase block and blocked using a 5% normal donkey serum (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA, USA). Tissue was then exposed to an avidin block followed by a biotin block (Biotin Blocking System, Dako Corporation, Carpinteria, CA, USA). Samples were incubated with a sheep anti-FIX primary antibody (Affinity Biologicals, Hamilton, Ontario, Canada) and then primary antibody detected using donkey anti-sheep IgG biotinylated secondary antibody (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA, USA) followed by standard HRP streptavidin complex and DAB substrate-chromogen incubation (Dako Corporation, Carpinteria, CA, USA). Tissue was then dehydrated while frozen through serial alcohol reduction ending with a xylene incubation. Slides were cleared using Histoclear (National Diagnostics, Atlanta, GA, USA) and mounted. To exclude background stain, test livers were also incubated with nonimmune pooled rabbit sera instead of primary antibody. To determine transduction efficiency, images of human FIX staining were captured using a Spot Camera (Diagnostic Instruments Inc., Burlingame, CA, USA) attached to a Zeiss Axioplan microscope and the images analyzed using Image Pro Plus (Media Cybernetics, Carlsbad, CA, USA).

Statistical analysis and data presentation

The figures show data from a single representative experiment or the meanss.e.m. of data pooled from 'n' independent experiments (data normalized as described in the figure legends). Data are typically reported in the text as meanss.e.m. and statistical analysis was by two-tailed Student's t-tests, accepting P<0.05 as statistically significant.

References

1. Herzog RW et al. Stable gene transfer and expression of human blood coagulation factor IX after intramuscular injection of recombinant adeno-associated virus. Proc Natl Acad Sci USA 1997; 94: 5804–5809. | Article | PubMed | ChemPort |
2. Snyder RO et al. Persistent and therapeutic concentrations of human factor IX in mice after hepatic gene transfer of recombinant AAV vectors. Nat Genet 1997; 16: 270–276. | Article | PubMed | ISI | ChemPort |
3. Snyder RO et al. Correction of hemophilia B in canine and murine models using recombinant adeno-associated viral vectors. Nat Med 1999; 5: 64–70. | Article | PubMed | ISI | ChemPort |
4. Chao H et al. Persistent expression of canine factor IX in hemophilia B canines. Gene Therapy 1999; 6: 1695–1704. | Article | PubMed | ChemPort |
5. Wang L et al. Sustained correction of bleeding disorder in hemophilia B mice by gene therapy. Proc Natl Acad Sci USA 1999; 96: 3906–3910. | Article | PubMed | ChemPort |
6. Wang L et al. Sustained expression of therapeutic level of factor IX in hemophilia B dogs by AAV-mediated gene therapy in liver. Mol Ther 2000; 1: 154–158. | Article | PubMed | ISI | ChemPort |
7. Chao H et al. Several log increase in therapeutic transgene delivery by distinct adeno-associated viral serotype vectors. Mol Ther 2000; 2: 619–623. | Article | PubMed | ISI | ChemPort |
8. Nathwani AC et al. Sustained high-level expression of human factor IX (hFIX) after liver-targeted delivery of recombinant adeno-associated virus encoding the hFIX gene in rhesus macaques. Blood 2002; 100: 1662–1669. | Article | PubMed | ISI | ChemPort |
9. Kay MA et al. Evidence for gene transfer and expression of factor IX in haemophilia B patients treated with an AAV vector. Nat Genet 2000; 24: 257–261. | Article | PubMed | ISI | ChemPort |
10. Ge Y, Powell S, Van Roey M, McArthur JG. Factors influencing the development of an anti-factor IX (FIX) immune response following administration of adeno-associated virus-FIX. Blood 2001; 97: 3733–3737. | Article | PubMed | ISI | ChemPort |
11. Nathwani AC et al. Factors influencing invivo transduction by recombinant adeno-associated viral vectors expressing the human factor IX cDNA. Blood 2001; 97: 1258–1265. | Article | PubMed | ISI | ChemPort |
12. Herzog RW et al. Muscle-directed gene transfer and transient immune suppression result in sustained partial correction of canine hemophilia B caused by a null mutation. Mol Ther 2001; 4: 192–200. | Article | PubMed | ISI | ChemPort |
13. Kay MA et al. A proposed rAAV-liver-directed clinical trial for hemophilia B. Mol Ther 2001; 5: S33 (abstr. 85).
14. Nakai H, Iwaki Y, Kay MA, Couto LB. Isolation of recombinant adeno-associated virus vector-cellular DNA junctions from mouse liver. J Virol 1999; 73: 5438–5447. | PubMed | ISI | ChemPort |
15. Mingozzi F et al. Improved hepatic gene transfer by using an adeno-associated virus serotype 5 vector. J Virol 2002; 76: 10497–10502. | Article | PubMed | ISI | ChemPort |
16. Brinster RL et al. Introns increase transcriptional efficiency in transgenic mice. Proc Natl Acad Sci USA 1988; 85: 836–840. | Article | PubMed | ChemPort |
17. Palmiter RD et al. Heterologous introns can enhance expression of transgenes in mice. Proc Natl Acad Sci USA 1991; 88: 478–482. | Article | PubMed | ChemPort |
18. Choi T, Huang M, Gorman C, Jaenisch R. A generic intron increases gene expression in transgenic mice. Mol Cell Biol 1991; 11: 3070–3074. | PubMed | ISI | ChemPort |
19. Rodriguez MH et al. Expression of coagulation factor IX in a haematopoietic cell line. Thromb Haemost 2002; 87: 366–373. | PubMed | ISI | ChemPort |
20. Miao CH et al. Inclusion of the hepatic locus control region, an intron, and untranslated region increases and stabilizes hepatic factor IX gene expression in vivo but not in vitro. Mol Ther 2000; 1: 522–532. | Article | PubMed | ISI | ChemPort |
21. Donello JE, Loeb JE, Hope TJ. Woodchuck hepatitis virus contains a tripartite posttranscriptional regulatory element. J Virol 1998; 72: 5085–5092. | PubMed | ISI | ChemPort |
22. Zufferey R, Donello JE, Trono D, Hope TJ. Woodchuck hepatitis virus posttranscriptional regulatory element enhances expression of transgenes delivered by retroviral vectors. J Virol 1999; 73: 2886–2892. | PubMed | ISI | ChemPort |
23. Loeb JE et al. Enhanced expression of transgenes from adeno-associated virus vectors with the woodchuck hepatitis virus posttranscriptional regulatory element: implications for gene therapy. Hum Gene Ther 1999; 10: 2295–2305. | Article | PubMed | ISI | ChemPort |
24. Huang Y, Wimler KM, Carmichael GG. Intronless mRNA transport elements may affect multiple steps of pre-mRNA processing. EMBO J 1999; 18: 1642–1652. | Article | PubMed | ISI | ChemPort |
25. Choi EA, Hope TJ. The role of poly(A) tail stability in the ability of the woodchuck hepatitis virus posttranscriptional regulatory element to stimulate gene expression. Mol Ther 2002; 5: S142 (abstr. 439).
26. Ill CR et al. Optimization of the human factor VIII complementary DNA expression plasmid for gene therapy of hemophilia A. Blood Coagul Fibrinolysis 1997; 8 (Suppl 2): S23–S30. | PubMed | ISI | ChemPort |
27. Hafenrichter DG et al. Quantitative evaluation of liver-specific promoters from retroviral vectors after in vivo transduction of hepatocytes. Blood 1994; 84: 3394–3404. | PubMed | ChemPort |
28. Ponnazhagan S et al. Adeno-associated virus 2-mediated gene transfer in vivo: organ-tropism and expression of transduced sequences in mice. Gene 1997; 190: 203–210. | Article | PubMed | ISI | ChemPort |
29. Sanlioglu S et al. Rate limiting steps of AAV transduction and implications for human gene therapy. Curr Gene Ther 2001; 1: 137–147. | Article | PubMed | ChemPort |
30. Herzog RW et al. Long-term correction of canine hemophilia B by gene transfer of blood coagulation factor IX mediated by adeno-associated viral vector. Nat Med 1999; 5: 56–63. | Article | PubMed | ISI | ChemPort |
31. Gao GP et al. Novel adeno-associated viruses from rhesus monkeys as vectors for human gene therapy. Proc Natl Acad Sci USA 2002; 99: 11854–11859. | Article | PubMed | ChemPort |
32. Wang L et al. High level expression of canine factor IX in hemophilia B dogs after readministration of AAV2/5 vectors. Mol Ther 2002; 5: S83 (abstr. 250).
33. Snyder RO, Xiao X, Samulski RJ. Production of recombininant adeno-associated viral vectors. In: Dracapoli NC, Haines JL, Korf BR, Moir DT, Mortan CC, Seidman CE, Seidman JG, Smith DR (eds). Current Protocols in Human Genetics. John Wiley & Sons: New York, 1996.

Acknowledgements

We thank Sandra Powell and Tammy Langer for technical assistance.