Scholarly article on topic 'The Pleiotropic Effects of Natural AAV Infections on Liver-directed Gene Transfer in Macaques'

The Pleiotropic Effects of Natural AAV Infections on Liver-directed Gene Transfer in Macaques Academic research paper on "Biological sciences"

Share paper
Academic journal
Mol Ther
OECD Field of science

Academic research paper on topic "The Pleiotropic Effects of Natural AAV Infections on Liver-directed Gene Transfer in Macaques"

The Pleiotropic Effects of Natural AAV Infections on Liver-directed Gene Transfer in Macaques

Lili Wang1, Roberto Calcedo1, Huan Wang12, Peter Bell1, Rebecca Grant1, Luk H Vandenberghe1, Julio Sanmiguel1, Hiroki Morizono3, Mark L Batshaw3 and James M Wilson1

'Gene Therapy Program, Department of Pathology and Laboratory Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA; 2Vaccine Research Institute, Third Affiliated Hospital, Sun Yat-Sen University, Guangzhou, China;3Children's National Medical Center, Department of Pediatrics, George Washington University School of Medicine and Health Sciences, Washington, DC, USA

Adeno-associated viral (AAV) vectors hold great potential for liver-directed gene therapy. Stable and high levels of transgene expression have been achieved in many murine models. Systemic delivery of AAV vectors in nonhuman primates (NHPs) that are natural hosts of AAVs appear to be challenging due to the high prevalence of pre-existing neutralizing antibodies (NAbs). This study evaluates the performance of AAV8, hu.37, and rh.8 vectors expressing green fluorescent protein (GFP) from a liver-specific promoter in rhesus macaques. Two of the animals that received AAV8 showed transduction of 24 and 40% of hepatocytes 7 days after systemic vector delivery. Importantly, expression was detected in several animals after 35 days despite the elevation of liver enzymes and development of transgene-specific T cells in liver. Pre-existing low levels of NAbs profoundly impacted the outcome of gene transfer and redirected vector DNA to spleen. We developed a sensitive in vivo passive transfer assay to detect low levels of NAbs to these novel AAV serotypes. Other strategies need to be developed to reduce immune response to the transgene in order to maintain long-term gene expression.

Received 15 July 2009; accepted 21 September 2009; published online 3 November2009. doi:10.1038/mt.2009.245


Vectors based on adeno-associated viruses (AAVs) show promise for in vivo applications of gene therapy. Successful gene transfer has been demonstrated without dose-limiting toxicities in several recent clinical trials. In some cases, such as gene transfer to the eye for Leber's congenital amaurosis and gene transfer to the central nervous system for Parkinson's disease, there is evidence for clinical efficacy.1-5 However, systemic administration of AAV to target the liver has led to liver inflammation and loss of transgene expression.6'7

Critical to the success of gene therapy is the availability of animal models that accurately predict outcomes in humans. AAV-mediated gene transfer to liver yields impressive results in mouse models, achieving efficient and long-term transgene expression in

the absence of toxicity.8,9 Studies in larger animal models, however, have not been as encouraging, with examples of lower transduction efficiencies and toxicity due to T-cell responses against foreign transgene products.10

One factor that may influence outcome of in vivo gene transfer is prior exposure to viruses that are similar to the virus used to create the vector. Viruses of the parvovirus family, of which AAV is a member, cause natural infections in many species including mice, dogs, nonhuman primates (NHPs), and humans. We have shown that AAVs persist as latent genomes that are widely distributed in primate tissues with substantial structural homology across macaques, great apes, and humans.11-13 These infections result in complex profiles of serum antibodies capable of binding and/or neutralizing various AAV serotypes.14,15 T-cell responses to AAV capsids are surprisingly low in primates.10,16

This study evaluates the performance of AAV vectors for liver-directed gene transfer in rhesus macaques using capsids that have shown promise in mouse studies.


In the accompanying paper, we evaluated a large number of AAV vectors based on natural isolates for liver-directed gene transfer in mice.17 Based on criteria of safety and efficacy in mice, using green fluorescent protein (GFP) as a reporter gene, we identified three different vectors with attractive performance profiles. These vectors [based on capsids from one human isolate (AAVhu.37) and two rhesus macaque isolates (AAV8 and AAVrh.8)], were selected for a thorough evaluation in macaques, the topic of this paper.

The basic study design was as follows: vectors were created with each capsid using an AAV2-based, single-stranded genome expressing the GFP transgene from a liver-specific thyroid hormone-binding globulin promoter. GFP was selected in order to permit a quantitative assessment of transduction efficiency and transgene-specific T-cell responses. This was important because in the treatment of recessive diseases part or all of the transgene product may be nonself. We showed in a previous study that expression of GFP from the ubiquitously expressed cytomegalovirus enhancer/chicken P-actin promoter in AAV vectors does lead to destructive cytotoxic T lymphocytes (CTL) in macaques.10 Previous studies suggested that these T-cell

Correspondence: James M Wilson, Gene Therapy Program, Department of Pathology and Laboratory Medicine, University of Pennsylvania, Philadelphia, Pennsylvania 19104, USA. E-mail:

responses may be diminished using a cell-specific promoter.18,19 As this is the approach we would expect to use in the clinic, it was the approach we used in this study. A total of 18 macaques were enrolled in this study: each vector was injected into 6 animals at a dose of 3 x 1012 genome copies/kg and groups of three were killed at days 7 and 35 for necropsy. Studies during the in-life phase of the experiment included analysis of serum for AAV-neutralizing antibody (NAb), an assessment of abnormalities in hematologic and chemical parameters of the blood, and analysis of peripheral blood mononuclear cells (PBMCs) for capsid and GFP-specific T-cell responses. Tissues obtained at necropsy were analyzed for GFP expression, histopathology, and vector genomes. In addition, tissue-derived mononuclear cells were evaluated for GFP and capsid T cells.

rhesus macaques were evaluated for the presence of NAb to the respective AAV capsids before selection, and only those that were sero-negative at the sensitivity of our assay (<1:20) were selected for participation. The titer of NAb was determined initially by the highest dilution that showed a 50% reduction of transduction in vitro. Analysis of day-7 liver tissues did reveal impressive GFP expression in some monkey livers. Figure 1 shows representative fluorescent micrographs, and Table 1 presents a quantitative mor-phometric analysis of the efficiency based on percent transduction

of hepatocytes and relative GFP intensity. For example, two of the animals that received AAV8 showed transduction of 24 and 40% of hepatocytes. Importantly, expression was detected in most animals even after 35 days. This contrasts to similar studies with AAV7 using the chicken P-actin promoter in which expression was absent in all animals at this time point.10 An unanticipated finding was the substantial animal-to-animal variation in trans-duction efficiency observed with all three vectors at both time points; some of the animals had no detectable GFP expression even at day 7.

We hypothesized that this may be due to the presence of NAb that is not detected by our standard in vitro NAb assay. Evidence in support of this was obtained by analyzing the animals for vector-specific NAb at day 7 (Table 1). We reasoned that a rapid and substantial increase in vector NAb at this short-term time point would be indirect evidence for the presence of memory B cells making antibodies that were reactive to the vector. There was a good correlation between the rise of day-7 NAb and gene transfer for the AAV8 vectors. Those animals showing good transduc-tion had NAbs of 1:20 to 1:40 at day 7, while those that performed less well had NAbs of 1:80 to 1:2,560 at day 7. This was more difficult to evaluate for AAVhu.37 and AAVrh.8 as all but one animal showed a substantial increase in NAb at day 7. Animal 608137 did

Table 1 Summary of gene transfer in rhesus macaques following systemic vector administration

Animal no. (category of outcome) AAV NAb (1:dilution) GFP transduction in liver GFP DNA copy/ diploid genomed Spleen/ liver GFP DNA ratio GFP T cells in liver (d35)

Vector Prea Day 7 Efficiency (%)b (mean ± SD) Intensityc (mean ± SD) Liver Spleen Lymph LFT elevation CD4+ CD8+

AAV8 605045 1:10 1:2,560 1.6 ± 0.7 6.1 ± 5.4 X 104 0.78 0.88 1.46 1.13 n.a. n.a. n.a.

(day 7) 607213 <1:5 1:40 23.7 ± 12.3 3.6 ± 1.5 X 106 17.46 3.90 1.18 0.22 n.a. n.a. n.a.

608059 <1:5 1:20 40.1 ± 11.9 6.9 ± 3.4 X 106 28.90 2.46 0.98 0.09 n.a. n.a. n.a.

AAV8 605067(C) 1:10 1:80 0 1.1 ± 1.2 X 101 0.005 15.44 0.20 3,207.95 No + -

(day 35) 605103(A) <1:5 1:20 11.9 ± 7.3 8.0 ± 7.9 X 105 16.36 0.30 0.13 0.02 Yes - -

606183(A) <1:5 1:40 1.6 ± 0.6 1.0 ± 2.0 X 103 8.09 0.29 0.27 0.04 Yes - -

AAVhu.37 607079 <1:5 1:1,280 1.3 ± 0.9 8.1 ± 9.5 X 104 1.58 4.55 0.73 2.89 n.a. n.a. n.a.

(day 7) 608171 <1:5 1:320 1.7 ± 1.2 1.0 ± 1.3 X 105 13.58 1.46 1.34 0.11 n.a. n.a. n.a.

608173 <1:5 1:160 2.4 ± 2.3 3.9 ± 3.0 X 104 7.36 0.40 0.008 0.05 n.a. n.a. n.a.

AAVhu.37 510213 (B) 1:5 1:2,560 0.1 ± 0.1 3.2 ± 7.1 X 102 1.12 7.16 0.42 6.40 Yes + +

(day 35) 603103(A) <1:5 1:1,280 1.7 ± 1.0 8.0 ± 3.0 X 104 6.25 4.44 1.12 0.71 No - -

607211 (C) <1:5 1:1,280 0 6.9 ± 3.7 0.006 7.04 0.23 1,103.69 No + -

AAVrh.8 606145 1:80 1:640 0 2.0 ± 3.3 X 101 0.0007 11.61 0.14 16,108.30 n.a. n.a. n.a.

(day 7) 607177 1:5 1:160 0.2 ± 0.1 5.2 ± 6.0 X 103 0.22 10.08 2.17 45.56 n.a. n.a. n.a.

608137 <1:5 1:40 9.0 ± 2.1 2.2 ± 1.2 X 105 5.12 1.19 1.00 0.23 n.a. n.a. n.a.

AAVrh.8 512099 (A) <1:5 1:320 0.5 ± 0.7 5.6 ± 3.9 0.83 0.30 1.14 0.37 Yes + -

(day 35) 601133 (B) <1:5 1:640 0 6.7 ± 3.6 1.33 3.06 0.56 2.29 Yes + +

603017(A) <1:5 1:1,280 2.3 ± 1.1 1.0 ± 8.3 X 103 3.37 0.17 0.34 0.05 Yes + -

Abbreviations: AAV, adeno-associated virus; GFP, green fluorescent protein; LFT, liver function test; lymph, mesenteric lymph nodes; n.a., not applied; NAb, neutralizing antibody.

aPre-samples were obtained from animals at various days before gene transfer: same day for 605045, 607212, and 608059; 12 days for 605067, 605103, and 606183; 11 days for 607079, 608171, and 608173; 20 days for 510213, 603103, and 607211; 14 days for 606145, 607177, and 608137; and 28 days fro 512099, 601133, and 603017. bMeasured as the percentage of GFp-positive area per liver section regardless of brightness. Ten sections per animal were analyzed. cDetermined as the sum of the brightness values of all pixels of an image above the background level (see Material and Methods). Ten sections per animal were analyzed. Intensity for an uninjected control monkey liver was 9.3 ± 0.8. dDetermined by quantitative PCR of transgene EGFP, according to a standard curve generated with linearized plasmid DNA pAAV.CMV.EGFP. The limit of detection is 7 x 10-5 copies/diploid genome.

Day 35

A * B * * * * 007076 600145

> 608171 607177

# ■ 808173 . ■ - X »» » - ^ 111» • .

60506? 510213 ! 512099

' ' w i,.' •V * • ■ <> * * * J * 60*103 A 003103 601133

606183 007211 - 6030IT

3.0 x 104 2.5 x 104 2.0 x 104 ■ 1.5 x 104 1.0 x 104 5.0 x 103 0

10 100 1,000 Reciprocal of serum dilution

Figure 1 GFP expression in rhesus macaque livers after intravenous infusion of 3 x 1012 GC/kg of AAV.TBG.EGFP pseudotyped with AAV8, hu.37, and rh.8 capsids. (a) Livers are harvested at day 7. (b) Livers are harvested at day 35. Representative images of each animal are shown. Bar = 100 jjm.

show the highest transduction of the remaining animals and had day-7 NAb of only 1:40.

Two experimental strategies were pursued to improve the detection of low NAb levels to the AAVs. One problem that limits the sensitivity of the in vitro assay, which is based on inhibition of transduction, is that most of the novel AAVs poorly transduce cells in vitro. They require high multiplicity of infection to get a quantifiable transgene expression in the absence of NAb. We also observed a nonspecific enhancement of transduction in the presence of high quantities of naive serum, complicating the assessment of transduction inhibition. The enhancement is not detectable at serum dilutions >1:20. The baseline serum samples, which originally titered to <1:20, were reevaluated at titers of 1:5 and 1:10, using identical dilutions of serum from naive mice to control for the nonspecific enhancement of transduction.

Figure 2 presents data from baseline sera of select animals that eventually received AAVrh.8 and AAV8. In both panels, there is one animal indistinguishable from the mouse serum even at 1:5 dilution (assigned titer <1:5). Altho ugh Animal 607177 shows 50% reduction at 1:5 dilution (titer = 1:5), and 605067 shows over 50% reduction at 1:10 dilution (titer = 1:10), a much higher titer of 1:80 was shown for 606145. It turns out that the original baseline serum sample obtained from this animal 125 days before gene transfer was consistent with the enrollment criteria of <1:20. The animal

5.0 x 103-

10 100 1,000 Reciprocal of serum dilution

Figure 2 In vitro neutralizing antibody assay on preinjected monkey sera. Serial twofold dilutions of monkey and naive mouse serum were incubated with 1 x 109 genome copies of (a) AAVrh.8.CMV.LacZ or (b) AAV8.CMV.LacZ. Transduction efficiency at each serum dilution was measured by P-galactosidase levels (relative light unit per second, RLU/ second) using a luminometer 24 hours after infection. AAV neutralization titer for each sample was determined by the highest serum dilution that inhibited AAV.CMV.LacZ transduction by >50%, compared with the mouse serum control. Pre-samples for 608137, 606145, and 607177 were obtained 14 days before gene transfer, and 12 days before gene transfer for 605103 and 605067. AAV, adeno-associated virus.

apparently seroconverted during the 111 days between the original sampling date and the second sampling date for pre-NAb (14 days before gene transfer), explaining the titer of 1:80 in the later pre-sample. Baseline serum samples were also analyzed for AAV NAb in vivo by passively transferring 200 |l of serum into naive mice immediately before intravenous injection of the respective AAV vectors expressing canine factor IX (cFIX). Evidence for NAb was demonstrated by showing a decrease in plasma cFIX levels relative to control animals, i.e., animals received passive transfer of naive mouse serum before vector administration (Figure 3). High tro-pism of these capsids for liver allowed the assay to be performed with low doses of vector, thereby maximizing its sensitivity.

There was an excellent correlation between transduction efficiency of AAV8 and the three measures of baseline NAb: (i) in vitro transduction inhibition with improved sensitivity (Table 1), (ii) day-7 memory B-cell response (Table 1), and (iii) in vivo passive transfer (Figure 3). Results with the other two vectors were not as informative: 9/12 animals had NAbs <1:5 (Table 1), and all showed significant inhibition of transduction in the in vivo passive transfer assay (Figure 3). These data suggest that the in vitro assay is still of insufficient sensitivity to detect levels of NAb that are capable of diminishing transduction of AAVhu.37 and AAVrh.8 in NHP liver. It also suggests that the in vivo passive transfer assay, as designed, is too sensitive and incapable of discriminating among animals within the AAVhu.37 and AAVrh.8 groups in terms of transduction in NHPs. We did not have sufficient baseline serum

1.5 x 10

1,200 1,000 800 g 600 400 200 0

—O- 605067 605103 -- - Naive mouse serum -r

-0-510213^—603103-■■»■■ Naive mouse serum

14 21 Time (days)

-0-512099^—601133-Naive mouse serum


Figure 3 Detection of neutralizing AAV antibodies in preinjected monkey sera by in vivo passive transfer experiment. C57BL/6 mice were infused with of 200 jjl of serum sample from an individual NHP or naive mouse serum. Two hours after the passive transfer, mice received an intravenous injection of 1 x 109 genome copies of AAV.LSP.cFIX-W vector packaged with the respective capsids (AAV8, hu.37, or rh.8). cFIX expression levels in the plasma at days 7, 21, and 28 after vector administration are shown. Data are presented as mean ± SD (n = 3). AAV, adeno-associated virus; cFIX, canine factor IX; NHP, nonhuman primate.



400 300 200 100 0

200' 100 0

400 300 200 100 0







Pre1 Pre2 0 1 2 3 4 5

Pre1 Pre2 0 1 2 3 4 5 Time (weeks)

Pre1 Pre2 0 1 2 3 4 5

Figure 4 Time course of liver enzyme levels (ALT and AST) in monkeys before and after intravenous infusion of 3 x 1012 genome copies/kg of AAV.TBG.EGFP pseudotyped with AAV8, hu.37, and rh.8. Prel samples were taken 71, 79, and 87 days before vector administration of AAV8, hu.37, and rh.8, respectively. Pre2 samples were taken 12, 20, and 35 days before vector administration of AAV8, hu.37, and rh.8, respectively. AAV, adeno-associated virus; ALT, alanine aminotransferase; AST, aspartate aminotransferase; LFT, liver function test.






= 300-



to repeat the passive transfer studies under less sensitive conditions (i.e., more vector or less serum).

The day-7 animals from all groups demonstrated no abnormalities in blood chemistry/hematology or evidence of T cells against the vector capsid or GFP in peripheral blood or tissues (data not shown). However, this was not the case for animals necropsied at day 35. The only consistent laboratory abnormalities were elevations in serum transaminases in 6/9 animals, first detected 3-4 weeks after gene transfer (Figure 4 and Table 1 ). Analysis of PBMCs failed to reveal T cells to AAV capsids; and significant elevations in GFP T cells were observed in only two animals at weeks 4 and 5 (Figure 5).

Analysis of mononuclear cells from tissues harvested at day 35 after gene transfer was more informative in assessing T-cell responses to GFP (Figure 6). Enzyme-linked immunosorbent spot assay (ELISPOT) analysis revealed interferon-y-expressing GFP-specific T cells in cells harvested from 7/9 livers, with 4 showing frequencies between 300 and 1,250 spots/106 cells. Spleen showed detectable but lower responses in 5/9 animals. PBMCs and bone marrow showed positive responses in only one animal.

Intracellular cytokine staining was performed on liver-derived mononuclear cells. In this study, we measured the magnitude of GFP-specific CD4+ and CD8+ T cells expressing interferon-^ or tumor necrosis factor-a (Table 2). GFP-specific CD4+ T cells

- AAV poolA

- AAV poolB

- AAV poolC

200 150 100 50 0

250 200 150 100 50 0

250 200 150 100 50 0

5 Pre 1 2 3 4 Time (weeks)

Figure 5 Time course of T-cell responses to AAV capsid and GFP in the PBMCs of monkeys after vector administration. PBMCs, isolated before and weekly after vector injection, were stimulated with AAV8 capsid and GFP-peptide libraries and assayed by IFNy ELISPOT. Background from unstimulated control (<30 SFU/106 lymphocytes) was subtracted from each sample. Asterisk indicates a positive response, which is defined as more than threefold of background and >55 SFU/106 lymphocytes. AAV, adeno-associated virus; ELISPOT, enzyme-linked immunosorbent assay; GFP, green fluorescent protein; IFNy, interferon--/; PBMC, peripheral blood mononuclear cell; SFU, spot forming units.


Table 2 ICS analysis of liver T-cell responses to GFP at day 35

Vector Animal CD4+ (% Total)" CD8+ (% Total)a

Contro l GFP Control GFP

AAV8 605067 0.32 0.47 0.16 0.16

605103 0.97 0.97 0.37 0.37

606183 0.63 0.63 0.33 0.35

AAVhu.37 510213 0.23 1.48 0.09 0.83

603103 0.66 0.66 0.23 0.24

607211 0.41 0.58 0.25 0.25

AAVrh.8 512099 0.81 1.21 0.24 0.25

601133 0.34 1.33 0.12 0.39

603017 0.18 0.34 0.28 0.29

Abbreviations: AAV, adeno-associated virus; GFP, green fluorescent protein; ICS, intracellular cytokine staining; IFNy, interferon-Y TNFa, tumor necrosis factor-a. "Magnitude of T-cell responses includes TNFa+IFNy-, TNFa-IFNy+, and TNFa+IFNy+ responses.

were found in 6/9 animals representing all three vectors, while CD8+ T cells specific to GFP were observed in only two animals. These were the two animals with the highest frequencies of GFP-specific CD4+ T cells (510213 from AAVhu.37 and 601133 from AAVrh.8).

Tissues recovered at days 7 and 35 were also analyzed for the presence of vector genomes. Table 1 summarizes vector abundance in liver, spleen, and mesenteric lymph nodes. A broader analysis of vector biodistribution is presented in Supplementary Table S1. There was a good correlation between the efficiency of transduction and the presence of vector in liver (p = 0.9). The

highest levels, reaching 14-29 vectors/diploid genome, were seen in those animals with transduction efficiencies ranging from 2 to 40%. The presence of vector genomes in liver was reduced 4-5 logs in most GFP-negative animals, with the exception of two animals (512099 and 601133) that had ~1 vector copy/diploid genome in liver but undetectable gene expression. The reason for this is unclear although it may be that transcriptionally dormant vector genomes reside to a variable extent in cells of the liver other than hepatocytes. GFP-specific RNA levels in liver, as measured by real-time reverse transcriptase-PCR (Supplementary Table S2), showed a good correlation with liver GFP expression (p = 0.9) and vector copy data (p = 0.8). Analysis of spleen revealed an interesting inverse relationship between transduction and vector DNA abundance. Low transduction and low vector DNA in liver were associated with higher vector genomes in spleen. The mechanism by which enhanced splenic uptake occurs is unclear, although it could be mediated by Fc receptor interactions with complexes of vector and antibody.20 GFP-specific RNA levels in spleen, as measured by real-time reverse transcriptase-PCR (Supplementary Table S2), did not show any correlation with the spleen vector copy data (p = -0.2). All animals showed very low but similar levels of GFP-specific RNA in spleen despite high vector copies in the spleen of some animals, indicating the liver-specific thyroid hormone-binding globulin promoter was not active in spleen.


The original intent of this study was to evaluate the relative performance of three vectors (AAVs 8, hu.37, and rh.8), in achieving safe and efficient liver-directed gene transfer in macaques. We learned during the execution of the study, however, that low levels

500 400

ÍS 1,000

f 500 .§ 400

1 300 о 200

500 400

500 400

* — 512099

1,000 *

500 *

500 400

500 400

PBMCs Spleen BM Liver

500 400

PBMCs Spleen BM Liver

500 400

PBMCs Spleen BM Liver







Figure 6 T-cell response to GFP in different tissues at day 35 after vector administration. PBMCs and lymphocytes isolated from spleen, bone marrow (BM), and liver were stimulated with GFP-peptide library and assayed by IFNy ELISPOT. Background from unstimulated control (<30 SFU/106 PBMCs, <65 SFU/106 splenocytes, <70 SFU/106 bone marrow lymphocytes, and <50 SFU/106 liver lymphocytes) was subtracted from each sample. Asterisk indicates a positive response, which is defined as more than threefold of background and >55 SFU/106 lymphocytes. AAV, adeno-associated virus; ELISPOT, enzyme-linked immunosorbent assay; GFP, green fluorescent protein; IFNy interferon-y; PBMC, peripheral blood mononuclear cell; SFU, spot forming units.

of AAV NAbs generated in the context of natural infections can have profound consequences on vector performance.

The relationship between AAV8 transduction and pre-existing NAb was quite clear, with NAbs as low as 1:5 significantly interfering with gene transfer in macaques. Previous reports of the frequency of pre-existing NAbs in NHPs and humans based on standard in vitro transduction inhibition assays probably represent underestimates. Re-examination of sera from human subjects with the more sensitive assay indicated that 55% of volunteers had detectable NAb to AAV8 at 1:5 dilution (data not shown).

Extrapolating from our macaque data, we hypothesize that AAV8 vectors would achieve therapeutic levels of gene transfer with modest doses of vector in about half of humans, i.e., those who are truly AAV8 NAb-negative. Strategies to treat those resistant to AAV8 could include using a different AAV serotype or engineering AAV8 to escape neutralization. The former strategy may be difficult, however, because humans who are sero-positive tend to have NAbs across a broad array of serotypes.14 We were unable to determine a relationship between NAbs and transduc-tion efficiency with AAVhu.37 and AAVrh.8. This was the result of an apparent lack of sensitivity of the in vitro assay (essentially 9/12 were <1:5 despite significant variation in transduction) combined with an overly sensitive in vivo passive transfer assay (sera from all animals completely inhibited in vivo transduction).

An encouraging result of this study is the finding that expression of the antigenic transgene GFP persisted for the duration of the study, 35 days. This differs from our recent study that showed the emergence of GFP-specific CTLs and extinction of expression over a similar time frame in macaques injected with AAV2/7

vectors expressing GFP from the cytomegalovirus-enhanced chicken P-actin promoter.10 Additional studies with longer time points are warranted to better define duration of transgene expression beyond 35 days and to rule out the late emergence of destructive CTLs. An important aspect of this result, therefore, is the avoidance of destructive CTLs to GFP in most animals, which we believe is achieved in part through the use of a liver-specific promoter. Two animals did generate CD8+ T cells in liver to GFP although it is hard to evaluate the impact of this on expression since pre-existing NAb may have blocked transduction. Several previous studies in mice and dogs demonstrated a similar phenomenon of diminished transgene immune responses when the expression of the transgene was restricted to non-immune-cells through the use of cell-specific promoters or microRNA targets sequences.18,19,21-23 This is the first demonstration, however, of the impact of hepatocyte-specific expression from AAV vectors on immune responses in NHPs and supports the use of this strategy in future clinical trials.

Pre-existing NAbs impacted vector biology beyond the trans-duction efficiency. One observation of concern was the impact of NAb on biodistribution (i.e., increased targeting of immune organs such as spleen) and its relationship to toxicity and immune responses. Our analysis of the total data set of day 35 animals in all vector groups indicates three scenarios based on level of measured or surmised pre-existing NAb. Table 1 summarizes features of these groups. In group A animals (605103, 606183, 603013, 512099, and 603017), vector DNA was primarily distributed to liver. All animals in this group had undetectable pre-NAb (<1:5) and retained detectable levels of GFP expression, although GFP

intensity in the liver of 512099 was much weaker than the others. T-cell responses to GFP in this group were either completely absent or limited to CD4+ T cells. All animals in this group except 603103 demonstrated some level of transaminase elevation without extinction of GFP expression, which could be due to GFP toxicity or low level CTLs. Group C (605067 and 607211) represents the other extreme where vector was significantly redirected to spleen. In the two animals of this group, pre-NAb titers were 1:10 (605067) and <1:5 (607211). No transgene expression was detected, and only limited GFP-specific CD4+ T cells were found in liver. No CD8+ T cells were activated to GFP, and there were no transaminase elevations. In this scenario, we speculate that retargeting to spleen prevented transduction of hepatocytes leading to no efficacy and low liver toxicity.

The third scenario is the most interesting. Group B (510213 and 601133) had intermediate redirecting of vector to spleen. This group had NAb at the limit of detection (1:5 and <1:5), produced low to undetectable GFP expression, and had vibrant liver CD4+ and CD8+ T-cell responses to GFP associated with transaminase elevations. The intermediate levels of NAb are presumably sufficient to diminish but not block hepatocyte transduction, while facilitating dendritic cell transduction and/or activation. The association between NAbs, spleen targeting and T-cell responses needs to be confirmed in larger macaque studies. However, there is ample precedent in the literature for antibody-enhanced activation of dendritic cells by viruses, including adenovirus, much of which has been ascribed to Fc receptor-mediated uptake of virus-Ab complexes.24,25 Similar antibody-dependent enhancement has been reported in infection of Fcy-R-positive cells with human parvovirus B19 (ref. 26), AAV2, and AAV10 (ref. 20).

T-cell responses to the AAV2 capsid have been implicated as the cause for loss of transgene (h.FIX) expression in the first AAV2-mediated, liver-directed clinical trial for hemophilia B.6,7 Higher than therapeutic level of factor IX expression was achieved in one subject but it was only short-lived, and the loss of transgene expression was concurrent with self-limiting transaminase elevation. CD8+ T cells specific to AAV2 capsid but not to human factor IX were detected in the PBMCs following vector administration. Although we and other groups have demonstrated the generation of CTL response to AAV capsid following vector administration,16,27-29 these capsid-specific CTLs do not eliminate AAV-transduced hepatocytes in mouse models29-31 although CTL killing can be showing in cultured human hepatocytes.32 Recently, Li et al. proposed an alternative mechanism for eliciting a CTL response against the therapeutic transgene through generation of cryptic epitopes; they demonstrated that such CTLs could eliminate transduced hepatocytes in mice.33 In this NHP study, capsid T cells were only detected in 1/18 animals. Animal 510213, injected with AAVhu.37, showed a T-cell response on day-7 PBMCs to the AAV8 pool-A, which covers the first 260 amino acids of VP1 that are completely conserved between AAV8 and AAVhu.37 (Figure 5). For AAVhu.37- and rh.8-injected animals, we could have missed the detection of some serotype-specific T cells. This is because we only used an AAV8 peptide library for stimulation in the ELISPOT assay, while AAVhu.37 and rh.8 share 93 and 91% sequence homology to AAV8 in the VP1 region, respectively. In gene therapy or vaccine clinical trials, PBMCs would be

a convenient source to monitor T-cell response following vector treatment. However, our study showed that peripheral T cells cannot predict local, i.e., liver, T-cell responses. While only 2/9 animals (510213 and 601133) were positive for GFP-specific T cells in their PBMCs (Figure 5), 7/9 animals had GFP-specific T cells in their livers with much higher magnitude (Figure 6); and 5/9 animals showed GFP-specific T cells in spleen.

This study illustrates how difficult it is to evaluate the performance of systemically administered AAV vectors in macaques. Pre-existing humoral immunity likely confounded previous studies of AAV gene transfer to target organs such as liver and heart in large animals. For example, the enhanced transduction of murine liver with AAV8 as compared to AAV2 has not been consistently demonstrated in macaques.34,35 This is likely to be the result of a higher prevalence of NAb to AAV8 in macaques as compared to AAV2 and the greater sensitivity of the AAV2 NAb assay contributed by its higher transduction in vitro requiring less vector in the transduction inhibition assay. Our comparative studies of the three candidate AAVs indicate that AAV8 is the most attractive vector for progressing into clinical trials as it: (i) yielded the highest levels of transduction and (ii) current assays are capable of detecting levels of NAb that would interfere with transduction.

In moving new AAV vectors into clinical trials that use systemic routes of administration, several issues need to be considered. First, in determining research subject eligibility it would seem prudent to measure baseline NAb and select those subjects who are truly NAb-negative. This approach will maximize efficacy, minimize subject-to-subject variation, and potentially diminish toxicity due to T-cell responses to transgene and extra-hepatic vector distribution. Assays to evaluate baseline NAbs need to be improved and validated going forward. Second, it appears that transgene-specific T-cell responses cannot be completely avoided even with a liver-specific promoter, although it should pointed out that our studies were performed with a highly immunogenic transgene product. This will have implications for trials involving recessive genetic diseases. Until this aspect of vector biology is better understood and circumvented, it may be best to limit participation to those subjects with minimal differences between the transgene product and the endogenous mutant gene product.


Vector construction, production, and purification. Recombinant AAV vectors expressing enhanced GFP or cFIX driven by the liver-specific thyroid hormone-binding globulin promoter36 and packaged with viral capsids from AAV8, hu.37, and rh.8 used in this study were produced by the Penn Vector Core at the University of Pennsylvania as described pre-viously.36 Novel AAV capsid sequences of natural isolates were described previously.11,13,37 All vectors used in this study were purified by two rounds of cesium chloride-gradient centrifugation, buffer-exchanged with phosphate-buffered saline, and concentrated using Amicon Ultra 15 centrifugal filter devices-100K (Millipore, Bedford, MA). Genome titer (genome copies/ml) of AAV vectors were determined by real-time PCR using a primer/ probe set corresponding to the polyA region of the vector and linearized plasmid standards. All vectors used in the NHP studies were subjected to extensive quality control tests including four repeated vector genome titrations, sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis for vector purity, Limulus amebocyte lysate for endotoxin detection (Cambrex Bio Science, Walkersville, MD), and transgene expression analysis in mice.

NHP experiments. Rhesus macaques (male Chinese origin and captive bred, 2.80-4.10 kg) were treated and cared for at the Nonhuman Primate Research Program facility of the Gene Therapy Program of the University of Pennsylvania (Philadelphia, PA) during the study. The study was performed according to a study protocol approved by the Environmental Health and Radiation Safety Office, the Institutional Biosafety Committee, and the Institutional Animal Care and Use Committee of the University of Pennsylvania. Vectors (3 x 1012 genome copies/kg) were administered to the study animals via the saphenous vein. Blood samples were taken prestudy and weekly during the study via venipuncture of the femoral vein. All clinical pathology tests on blood samples were conducted by Antech Diagnostics (Irvine, CA), including complete blood counts and differentials, and complete clinical chemistries.

PBMCs were isolated from whole blood collected in EDTA-containing Vacutainer tubes after Picoll density-gradient centrifugation. At the time of necropsy, lymphocytes were also isolated from bone marrow and liver using Ficoll and Percoll, respectively, by density-gradient centrifugation and from spleen and mesenteric lymph nodes by crushing through 40-|lm strainers (BD Biosciences, San Jose, CA). At the time of necropsy (days 7 or 35), 16 tissues, including the target organ liver and 15 distant tissues (brain, bone marrow, diaphragm, heart, kidney, lung, mesenteric lymph nodes, pancreas, seminal vesicles, skeletal muscle, spinal cord, spleen, stomach, testicles, and urinary bladder) were collected for histopathology and vector biodistribution analysis.

Passive transfer experiments. C57BL/6 male mice (6-8 weeks old) were purchased from Charles River Laboratories (Wilmington, MA) and kept at the Animal Facility of the Translational Research Laboratories at the University of Pennsylvania. All animal procedures were performed in accordance with protocols approved by the Institutional Animal Care and Use Committee at the University of Pennsylvania. A group of three mice each received an intravenous injection of 200 |il of serum sample from an individual NHP. Two hours after the passive transfer, mice received an intravenous injection of 1 x 109 genome copies of AAV.LSP.cFIX-W vector packaged with the respective capsids (AAV8, hu.37, or rh.8). cFIX expression levels in the plasma at days 7, 14, 21, and 28 after vector administration were determined by enzyme-linked immunosorbent assay as described previously.38

Quantification of GFP expression in liver. To visualize GFP expression in liver, tissues were fixed overnight in formalin, washed in phosphate-buffered saline for 30 minutes, and frozen in O.C.T. compound (Sakura Finetek USA, Torrance, CA). Cryosections were prepared at 8 |im. GFP transduction in liver was evaluated by two aspects: the percentage of area expressing GFP and the intensity of GFP, each were examined on 10 randomly chosen images from liver sections (2-3 images from each lobe) that were taken at identical camera and microscope settings, with a fluorescence microscope equipped with a digital camera. To quantify the percentage of area expressing GFP, images were analyzed with ImageJ software (Rasband 1997-2006; National Institutes of Health, Bethesda, MD, to determine the GFP-positive area of each image. Briefly, images were subjected to thresholding (i.e., determining the minimal brightness value that represents true GFP fluorescence) using GFP-negative control sections from untreated rhesus macaques as reference for background levels. The percentage of areas with brightness values equal to or exceeding the threshold value within each image was then calculated and averaged for all 10 images per animal. To quantify GFP intensity of each image, the brightness values were measured for all pixels with ImageJ and then for each brightness value between the background level (8) and the maximum value (255) the number of corresponding pixels was determined. The number of pixels was then multiplied with their brightness value and the products were added to give a final value for GFP intensity for each image.

Interferon-y ELISPOT and intracellular cytokine staining assays.

Interferon-y ELISPOT assay were performed using previously described protocols.39 Lymphocytes (2 x 105) isolated from PBMCs or tissues were stimulated with 2 |lg/ml of GFP-peptide library (46-peptide pool containing 15-mers with a 10-amino acid overlap with the preceding peptide; Mimotopes, Clayton, Australia) or AAV8 capsid peptide library (3 peptide pools-A, B, and C, spanning the entire VP1 region).16 Spots were counted with an AID ELISPOT reader system (Autoimmun Diagnostika, Strassberg, Germany). Peptide-specific cells were represented as spot forming units per 106 lymphocytes and were calculated by subtracting spot numbers in media only wells from spot numbers in peptide-containing wells. Intracellular cytokine staining assays to measure the cytokine production by liver lymphocytes to GFP were performed as previously described39 by combined surface and intracellular staining with monoclonal antibodies and subsequent five-color flow cytometric analysis. Data were analyzed using FlowJo software (Treestar, OR).

AAV NAb assay. NHP serum samples were heat inactivated at 56 °C for 30 minutes. NAb assays were performed on Huh7 cells as previously described.14

Vector biodistribution analysis. Tissue DNA was extracted using a QIAamp DNA Mini Kit (Qiagen, Valencia, CA). Detection and quantification of vector genomes in extracted DNA were performed by real-time PCR as described previously.40

Real-time reverse transcriptase-PCR. Total RNAs were extracted from macaque tissues with TRIzol reagent (Invitrogen, Carlsbad, CA), treated with RNase-free DNase I (Roche, Indianapolis, IN), and purified with a RNeasy Plus mini kit (Qiagen). Total RNA was reverse transcribed with a high capacity complementary DNA reverse transcription kit (Applied Biosystems, Foster City, CA) according to the manufacturer's instructions. A combined absolute and relative quantification strategy as described previously10 was employed to assess the presence and level of GFP sequences. Fifty nanograms each of total RNA equivalent complementary DNA (5 |il of RT reaction) was used for TaqMan reactions.


Table SI. Biodistribution of vector DNA in rhesus macaques following intravenous vector administration.

Table S2. GFP transcript levels in rhesus macaques following intravenous vector administration.


We thank Julie Johnston and Arbans Sandhu (Penn Vector) for supplying vectors; Erin Bote for invaluable assistance with macaque studies (Gene Therapy Program); Qiuyue Qin, Surina Boyd, and Jennifer Miliaresis (Immunology Core, Gene Therapy Program) for lymphocyte isolation and NAb assays; and Hongwei Yu (Gene Therapy Program) for tissue sectioning. This work was supported in part by the Kettering Family Foundation and the following grants to J.M.W.: P01-HD057247, P01-HL059407, P30-DK047757 and GlaxoSmithKline. H.W. was supported by a scholarship from China Scholarship Council. L.H.V. is an inventor on patents licensed to various biopharmaceutical companies, including ReGenX. J.M.W. is a consultant to ReGenX Holdings, and is a founder of, holds equity in, and receives a grant from affiliates of ReGenX Holdings; in addition, he is an inventor on patents licensed to various biopharmaceutical companies, including affiliates of ReGenX Holdings.


1. Maguire, AM, Simonelli, F, Pierce, EA, Pugh, EN Jr, Mingozzi, F, Bennicelli, J et a!. (2008). Safety and efficacy of gene transfer for Leber's congenital amaurosis.

N Eng! j Med 358: 2240-2248.

2. Bainbridge, JW, Smith, AJ, Barker, SS, Robbie, S, Henderson, R, Balaggan, K et a!. (2008). Effect of gene therapy on visual function in Leber's congenital amaurosis. N Eng! j Med 358: 2231-2239.

3. Hauswirth, WW, Aleman, TS, Kaushal, S, Cideciyan, AV, Schwartz, SB, Wang, L et a!. (2008). Treatment of Leber congenital amaurosis due to RPE65 mutations by ocular

subretinal injection of adeno-associated virus gene vector: short-term results of a phase I trial. Hum Gene Ther 19: 979-990.

4. Kaplitt, MG, Feigin, A, Tang, C, Fitzsimons, HL, Mattis, P, Lawlor, PA et al. (2007). Safety and tolerability of gene therapy with an adeno-associated virus (AAV) borne GAD gene for Parkinson's disease: an open label, phase I trial. Lancet 369: 2097-2105.

5. Eberling, JL, Jagust, WJ, Christine, CW, Starr, P, Larson, P, Bankiewicz, KS et al. (2008). Results from a phase I safety trial of hAADC gene therapy for Parkinson disease. Neurology 70: 1980-1983.

6. Manno, CS, Pierce, GF, Arruda, VR, Glader, B, Ragni, M, Rasko, JJ et al. (2006). Successful transduction of liver in hemophilia by AAV-factor IX and limitations imposed by the host immune response. Nat Med 12: 342-347.

7. Mingozzi, F, Maus, MV, Hui, DJ, Sabatino, DE, Murphy, SL, Rasko, JE et al. (2007). CD8+ T-cell responses to adeno-associated virus capsid in humans. Nat Med 13: 419-422.

8. Snyder, RO, Miao, CH, Patijn, GA, Spratt, SK, Danos, O, Nagy, D et al. (1997). Persistent and therapeutic concentrations of human factor IX in mice after hepatic gene transfer of recombinant AAV vectors. Nat Genet 16: 270-276.

9. Cunningham, SC, Spinoulas, A, Carpenter, KH, Wilcken, B, Kuchel, PW and Alexander, IE (2009). AAV2/8-mediated correction of OTC deficiency is robust in adult but not neonatal Spfash mice. Mol Ther 17: 1340-1346.

10. Gao, G, Wang, Q, Calcedo, R, Mays, L, Bell, P, Wang, L et al. (2009). AAV-mediated gene transfer to nonhuman primate liver can elicit destructive transgene-specific T cell responses. Hum Gene Ther 20: 930-942.

11. Gao, GP, Alvira, MR, Wang, L, Calcedo, R, Johnston, J and Wilson, JM (2002). Novel adeno-associated viruses from rhesus monkeys as vectors for human gene therapy. Proc Natl Acad Sci USA 99: 11854-11859.

12. Gao, G, Alvira, MR, Somanathan, S, Lu, Y, Vandenberghe, LH, Rux, JJ et al. (2003). Adeno-associated viruses undergo substantial evolution in primates during natural infections. Proc Natl Acad Sci USA 100: 6081-6086.

13. Gao, G, Vandenberghe, LH, Alvira, MR, Lu, Y, Calcedo, R, Zhou, X et al. (2004). Clades of adeno-associated viruses are widely disseminated in human tissues.

J Virol 78: 6381-6388.

14. Calcedo, R, Vandenberghe, LH, Gao, G, Lin, J and Wilson, JM (2009). Worldwide epidemiology of neutralizing antibodies to adeno-associated viruses. J Infect Dis 199: 381-390.

15. Murphy, SL, Li, H, Mingozzi, F, Sabatino, DE, Hui, DJ, Edmonson, SA et al. (2009). Diverse IgG subclass responses to adeno-associated virus infection and vector administration. J Med Virol 81: 65-74.

16. Vandenberghe, LH, Wang, L, Somanathan, S, Zhi, Y, Figueredo, J, Calcedo, R et al. (2006). Heparin binding directs activation of T cells against adeno-associated virus serotype 2 capsid. Nat Med 12: 967-971.

17. Wang, L, Wang, H, Bell, P, McCarter, RJ, He, J, Calcedo, R et al. (2009). Systematic evaluation of AAV vectors for liver-directed gene transfer in murine models. Mol Ther, this issue.

18. Cordier, L, Gao, GP, Hack, AA, McNally, EM, Wilson, JM, Chirmule, N et al. (2001). Muscle-specific promoters may be necessary for adeno-associated virus-mediated gene transfer in the treatment of muscular dystrophies. Hum Gene Ther 12: 205-215.

19. Franco, LM, Sun, B, Yang, X, Bird, A, Zhang, H, Schneider, A et al. (2005). Evasion of immune responses to introduced human acid a-glucosidase by liver-restricted expression in glycogen storage disease type II. Mol Ther 12: 876-884.

20. Mori, S, Takeuchi, T and Kanda, T (2008). Antibody-dependent enhancement of adeno-associated virus infection of human monocytic cell lines. Virology 375: 141-147.

21. Ziegler, RJ, Lonning, SM, Armentano, D, Li, C, Souza, DW, Cherry, M et al. (2004). AAV2 vector harboring a liver-restricted promoter facilitates sustained expression of therapeutic levels of a-galactosidase A and the induction of immune tolerance in Fabry mice. Mol Ther 9: 231-240.

22. Brown, BD, Venneri, MA, Zingale, A, Sergi Sergi, L and Naldini, L (2006). Endogenous microRNA regulation suppresses transgene expression in hematopoietic lineages and enables stable gene transfer. Nat Med 12: 585-591.

23. Brown, BD, Cantore, A, Annoni, A, Sergi, LS, Lombardo, A, Della Valle, P et al. (2007). A microRNA-regulated lentiviral vector mediates stable correction of hemophilia B mice. Blood 110: 4144-4152.

24. Meier, O, Gastaldelli, M, Boucke, K, Hemmi, S and Greber, UF (2005). Early steps of clathrin-mediated endocytosis involved in phagosomal escape of Fcy receptor-targeted adenovirus. J Virol 79: 2604-2613.

25. Clyde, K, Kyle, JL and Harris, E (2006). Recent advances in deciphering viral and host determinants of dengue virus replication and pathogenesis. J Virol80: 11418-11431.

26. Munakata, Y, Kato, I, Saito, T, Kodera, T, Ishii, KK and Sasaki, T (2006). Human parvovirus B19 infection of monocytic cell line U937 and antibody-dependent enhancement. Virology 345: 251-257.

27. Sabatino, DE, Mingozzi, F, Hui, DJ, Chen, H, Colosi, P, Ertl, HC et al. (2005). Identification of mouse AAV capsid-specific CD8+ T cell epitopes. Mol Ther 12: 1023-1033.

28. Chen, J, Wu, Q, Yang, P, Hsu, HC and Mountz, JD (2006). Determination of specific CD4 and CD8 T cell epitopes after AAV2- and AAV8-hF.IX gene therapy. Mol Ther 13: 260-269.

29. Wang, L, Figueredo, J, Calcedo, R, Lin, J and Wilson, JM (2007). Cross-presentation of adeno-associated virus serotype 2 capsids activates cytotoxic T cells but does not render hepatocytes effective cytolytic targets. Hum Gene Ther 18: 185-194.

30. Li, C, Hirsch, M, Asokan, A, Zeithaml, B, Ma, H, Kafri, T et al. (2007). Adeno-associated virus type 2 (AAV2) capsid-specific cytotoxic T lymphocytes eliminate only vector-transduced cells coexpressing the AAV2 capsid in vivo. J Virol 81: 7540-7547.

31. Li, H, Murphy, SL, Giles-Davis, W, Edmonson, S, Xiang, Z, Li, Y et al. (2007). Pre-existing AAV capsid-specific CD8+ T cells are unable to eliminate AAV-transduced hepatocytes. Mol Ther 15: 792-800.

32. Pien, GC, Basner-Tschakarjan, E, Hui, DJ, Mentlik, AN, Finn, JD, Hasbrouck, NC et al. (2009). Capsid antigen presentation flags human hepatocytes for destruction after transduction by adeno-associated viral vectors. J Clin Invest 119: 1688-1695.

33. Li, C, Goudy, K, Hirsch, M, Asokan, A, Fan, Y, Alexander, J et al. (2009). Cellular immune response to cryptic epitopes during therapeutic gene transfer. Proc Natl Acad Sci USA 106: 10770-10774.

34. Davidoff, AM, Gray, JT, Ng, CY, Zhang, Y, Zhou, J, Spence, Y 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.

35. Jiang, H, Couto, LB, Patarroyo-White, S, Liu, T, Nagy, D, Vargas, JA et al. (2006). Effects of transient immunosuppression on adenoassociated, virus-mediated, liver-directed gene transfer in rhesus macaques and implications for human gene therapy. Blood 108: 3321-3328.

36. Gao, G, Lu, Y, Calcedo, R, Grant, RL, Bell, P, Wang, L et al. (2006). Biology of AAV serotype vectors in liver-directed gene transfer to nonhuman primates. Mol Ther 13: 77-87.

37. Vandenberghe, LH, Breous, E, Nam, HJ, Gao, G, Xiao, R, Sandhu, A et al. (2009). Naturally occurring singleton residues in AAV capsid impact vector performance and illustrate structural constraints. Gene Ther (epub ahead of print).

38. Wang, L, Calcedo, R, Nichols, TC, Bellinger, DA, Dillow, A, Verma, IM et al. (2005). Sustained correction of disease in naive and AAV2-pretreated hemophilia B dogs: AAV2/8-mediated, liver-directed gene therapy. Blood 105: 3079-3086.

39. Calcedo, R, Vandenberghe, LH, Roy, S, Somanathan, S, Wang, L and Wilson, JM (2009). Host immune responses to chronic adenovirus infections in human and nonhuman primates. J Virol 83: 2623-2631.

40. Bell, P, Moscioni, AD, McCarter, RJ, Wu, D, Gao, G, Hoang, A et al. (2006). Analysis of tumors arising in male B6C3F1 mice with and without AAV vector delivery to liver. Mol Ther 14: 34-44.