Skip to main content
NCBI home page
Molecular Therapy logoLink to Molecular Therapy
. 2011 Mar 1;19(6):1079–1089. doi: 10.1038/mt.2011.3

A Potential Role of Distinctively Delayed Blood Clearance of Recombinant Adeno-associated Virus Serotype 9 in Robust Cardiac Transduction

Nicole M Kotchey 1, Kei Adachi 1, Maliha Zahid 1, Katsuya Inagaki 1,*, Rakshita Charan 1, Robert S Parker 2, Hiroyuki Nakai 1
PMCID: PMC3129802  PMID: 21364543

Abstract

Recombinant adeno-associated virus serotype 9 (rAAV9) vectors show robust in vivo transduction by a systemic approach. It has been proposed that rAAV9 has enhanced ability to cross the vascular endothelial barriers. However, the scientific basis of systemic administration of rAAV9 and its transduction mechanisms have not been fully established. Here, we show indirect evidence suggesting that capillary walls still remain as a significant barrier to rAAV9 in cardiac transduction but not so in hepatic transduction in mice, and the distinctively delayed blood clearance of rAAV9 plays an important role in overcoming this barrier, contributing to robust cardiac transduction. We find that transvascular transport of rAAV9 in the heart is a capacity-limited slow process and occurs in the absence of caveolin-1, the major component of caveolae that mediate endothelial transcytosis. In addition, a reverse genetic study identifies the outer region of the icosahedral threefold capsid protrusions as a potential culprit for rAAV9's delayed blood clearance. These results support a model in which the delayed blood clearance of rAAV9 sustains the capacity-limited slow transvascular vector transport and plays a role in mediating robust cardiac transduction, and provide important implications in AAV capsid engineering to create new rAAV variants with more desirable properties.

Introduction

Recombinant adeno-associated virus (rAAV) is among the most promising vectors for in vivo gene delivery. The usefulness of rAAV vectors has been increasingly expanded since a number of new serotypes were isolated from human and nonhuman primate tissues.1 Among the newly identified serotypes, AAV serotype 8 (AAV8) and AAV serotype 9 (AAV9) have gained much attention because rAAV vectors derived from these two serotypes can transduce various organs including the liver, heart, and skeletal muscles with extremely high efficiency following systemic administration via the periphery.1,2,3,4,5,6,7,8,9 Such robust transduction by rAAV8 and 9 vectors has been presumed to be ascribed to strong tropism for these cell types, efficient cellular uptake of vectors, and/or rapid uncoating of virion shells in cells.10 The ability of rAAV8 and 9 vectors to efficiently cross capillary endothelial cell barriers also appears to be responsible for their robustness.5,7 Thus, systemic gene delivery via the periphery using these robust rAAV serotype vectors has emerged as a new promising approach to treat various human diseases. Proof-of-concept for rAAV-mediated systemic gene delivery by peripheral vein injection has been shown in many preclinical studies to treat hemophilia,2,9,11 muscular dystrophy,6,7,12,13 cardiac failure,6,13,14 and metabolic diseases including glycogen storage disease15 and phenylketonuria.16

Despite the appreciation for the practical usefulness of the robust rAAV serotype vectors for systemic gene delivery, the scientific basis for rAAV-mediated systemic approach has not been fully established and the mechanism of their in vivo robustness remains largely unknown. Rapid clearance of exogenous agents from the bloodstream primarily by Kupffer cells in the liver has been a major hurdle in systemic approaches. This is particularly the case with adenoviral vectors and cationic nonviral vectors.17,18,19 In addition, for vectors to reach parenchymal cells from the bloodstream, they must cross the capillary walls, which may pose as a barrier in rAAV vector–mediated in vivo transduction. Transcytosis of rAAV across endothelial cells has been demonstrated in vitro;20 however, how rAAV vectors cross the capillary barriers in vivo remains unknown. Since systemic gene delivery approaches using rAAV vectors have become increasingly popular and could be superior to other systems in many applications, understanding of potential barriers that include blood clearance and transvascular transport of vectors will be essential to further improve the systems for more efficient and safer rAAV-mediated systemic gene delivery.

In the present study, to understand the roles of potential extracellular barriers in rAAV vector–mediated in vivo transduction, we comprehensively characterized pharmacokinetic profiles as well as in vivo transduction efficiencies of various rAAV serotypes and capsid-modified variants delivered via intravenous route in mice. Our study demonstrated that rAAV9 exhibits a distinctively slow blood clearance and remains infectious in the bloodstream, making this serotype very attractive for systemic gene delivery. Subsequently, we performed a series of mechanistic studies to understand the mechanisms of the in vivo robustness of rAAV9 vectors and the biological roles of hepatic and cardiac capillary walls in rAAV-mediated transduction. As a result, we identified a potential capsid region responsible for rAAV9's delayed blood clearance. In addition, we found that transvascular transport of rAAV9 is a capacity-limited caveolin-1–independent slow process forming a significant barrier in the heart, and this barrier has been overcome by rAAV9's distinctively delayed blood clearance. In the liver, such a barrier did not exist.

Results

rAAV9 exhibited very slow blood clearance distinct from other serotypes and maintained substantial infectivity in the bloodstream

To understand extracellular barriers in rAAV-mediated transduction in vivo following systemic vector administration, we determined blood clearance rates of rAAV1, 2, 8, and 9 vectors following intravenous administration of AAV-CMV-lacZ vectors into adult C57BL/6 male mice (n = 3–7 each). All the rAAV serotypes exhibited a biphasic blood clearance pattern showing a rapid decline during the first 30 minutes and a slower clearance thereafter (Figure 1a,b). Due to the biphasic nature of the blood vector concentration-time curves, two types of half-lives (distribution half-lives for the first 30 minutes, t1/2d:1–30 minutes, and clearance half-lives between 1 and 24 hours postinjection, t1/2c:1–24 hours) were determined (Table 1). The most striking finding was that rAAV9 exhibited a distinctively delayed blood clearance compared to the other serotypes. Blood concentrations of rAAV9 were significantly higher than rAAV1 (P < 0.01 at 4 hours and 8 hours, and P < 0.001 at 24 hours postinjection) and rAAV2 (P < 0.05 at 1 hour postinjection), and t1/2c:1–24 hours of rAAV9 was significantly prolonged compared to rAAV1 (P < 0.001) and rAAV8 (P < 0.001). None of rAAV vectors derived from serotypes 6, 7, and rh.10 exhibited a substantially delayed clearance as seen with rAAV9 (Supplementary Figure S1). Degrees of loss of infectivity of rAAV2, 8 and 9 in the bloodstream were assessed using rAAV particles in the blood collected from vector-injected animals 24 hours postinjection. rAAV8 and 9 retained readily measurable in vitro infectivity, while the infectivity of rAAV2 was abolished during the 24 hours following injection (Supplementary Figure S2a). In vivo infectivity of rAAV9 in the bloodstream was maintained for 24 hours at a level comparable to that of the rAAV9 vector stock (Supplementary Figure S2b).

Figure 1.

Figure 1

Open in a new tab

Blood vector concentration-time curves following intravenous injection of various serotype or variant AAV-CMV-lacZ vectors in adult C57BL/6 mice (n = 3–7 each). (a,b) Recombinant adeno-associated virus serotype 1 (rAAV1), rAAV2, rAAV8 or rAAV9 vector was injected into mice via the tail vein in bolus at a dose of 1.0 × 1013 vg/kg. Concentrations of rAAV particles in the blood are plotted as a function of time after injection on (a) a linear-linear scale or (b) a linear-log scale. Panel (a) highlights the clearance rates during the first 8 hours following vector administration. (c,d) Various AAV1 and AAV9 hybrid vectors and (e) rAAV variants with a modification at the C-terminus of the capsid were injected into mice (please also see Figure 2a). Vertical bars represent standard errors.

Table 1. Recombinant adeno-associated virus (rAAV) distribution and clearance half-lives (t1/2d:1–30 minutes and t1/2c:1–24 hours, respectively) and 72-hour AUCs following tail vein injection into mice.

graphic file with name mt20113t1.jpg

Open in a new tab

Construction of a panel of AAV1 and AAV9 hybrid vectors

To address the underlying molecular mechanisms for the delayed blood clearance of rAAV9, we investigated whether rAAV9's distinctive pharmacokinetic trait could be transferred, by swapping a domain, to other serotype vectors that exhibit a rapid blood clearance. As a recipient of individual AAV9 capsid domains, rAAV1 was chosen because rAAV1 exhibited the most rapid blood clearance among the serotypes we studied (Figure 1a,b). To this end, we generated a series of AAV-CMV-lacZ vectors carrying AAV1 and AAV9 hybrid capsids (Figure 2a). All the hybrid rAAV virions could be produced; however, large-scale production of rAAV1.9-4, rAAV1.9-5, and rAAV1.9-7 did not yield high titers sufficient for animal experiments. Therefore, these hybrid rAAV vectors were not used in the subsequent in vivo experiments except for rAAV1.9-7, which was injected into a mouse at a dose of 1.1 × 1012 vg/kg.

Figure 2.

Figure 2

Open in a new tab

Adeno-associated virus serotype 1 (AAV1) and AAV9 and other types of hybrid vectors. (a) Schematic representation of the hybrid capsids. Positions for five interstrand loops (loop I–V) are shown at the top. AAV2 capsid R585E mutation is indicated with a star. (b) AAV1.9-3 full capsid viewed down an icosahedral threefold symmetry axis at the center. The colored region indicates the 113-amino acid (a.a.). AAV9 capsid segment (a.a. 456–568) embedded in the AAV1.9-3 capsid. Each color represents the region from a different capsid monomer. (c) A closer view of (b). The 113-a.a. region from two different capsid monomers constitutes the middle and outer spikes of a protrusion as indicated with the red and green regions from the blue monomer and a gray monomer, respectively, in the circle. (d) AAV1.9-6 full capsid viewed down an icosahedral twofold symmetry axis at the center. The colored region indicates the 37-a.a. AAV9 capsid C-terminal segment (a.a. 699-735) embedded in the AAV1.9-6 capsid. Each color represents a region from a different capsid monomer. (e) The AAV9 capsid segments embedded in the AAV1.9-3-1.5 capsid viewed down from an outer spike of a capsid protrusion. Two separate AAV9 capsid regions from two different capsid monomers; i.e., a.a. 456-476 (the green region from a gray monomer in the circle) and a.a. 550–568 (the red region from a blue monomer in the circle) form the outer spike of the AAV capsid protrusion. UCSF Chimera45 was used to generate these figures. The inner, middle, and outer spikes that constitute the threefold capsid protrusion are indicated with black circles, triangles and crosses, respectively. Three and fivefold symmetry axes are shown with orange triangles and pentagons, respectively.

An AAV1-to-AAV9 exchange of the outer region of the AAV1 capsid protrusions located at the icosahedral threefold symmetry axes significantly delayed blood clearance

We first compared blood clearance rates between rAAV1, rAAV9, and rAAV1.9-1, -2, -3, and -6 (Figure 1c). rAAV1.9-3 exhibited a slowed blood vector concentration-time curve similar to that of rAAV9 while other hybrids were cleared rapidly. rAAV1.9-3 had the AAV1 capsid with a 113-amino acid (a.a.) stretch (a.a 455–568) replaced with a.a. 456–568 of the AAV9 capsid (Figure 2a). This modification of the AAV1 capsid substantially delayed the blood clearance and prolonged its half-life from t1/2c:1–24 hours = 3.3 to 23.8 hours. Next, to investigate whether a particular subregion within the 113-a.a. region has the ability to substantially delay blood clearance of rAAV, rAAV1.9-3-1 through AAV1.9-3-5 and AAV1.9-3-1.5 (Figure 2a) were injected into mice. None of these six hybrids exhibited substantially delayed clearance as seen with rAAV9 or rAAV1.9-3 (Figure 1d); therefore, the delayed blood clearance was not credited to a single smaller domain of the AAV9 capsid. rAAV9.1-3, in which the domains in the rAAV1.9-3 capsid were reciprocally swapped, exhibited rapid blood clearance (Figure 1c). Taken together, the relatively long stretch of 113 a.a. that constitutes the outer region of the AAV capsid protrusions located at the threefold symmetry axes (i.e., the threefold protrusions) (Figure 2b,c) is a potential determinant of the blood clearance rates of rAAV1 and rAAV9.

rAAV1.9-3-1.5 and rAAV1.9-6: the two rAAV1-derived variants that robustly transduced the liver and heart, respectively, with significantly accelerated blood clearance

Vector-injected animals were killed 11 days postinjection, and transduction efficiency in the liver and heart was determined (Figure 3 and Table 2). Although the primary objective in the creation of a panel of AAV1 and AAV9 hybrid vectors was to understand the molecular mechanisms for the delayed blood clearance of rAAV9, the array of the in vivo transduction data has provided us an opportunity to seek to identify potential capsid determinants for rAAV9's robustness in hepatic and cardiac transduction. The most informative hybrids are rAAV1.9-3-1.5 and rAAV1.9-6, which showed significantly enhanced transduction efficiency in the liver and the heart, respectively, compared to the parental rAAV1 (P < 0.001 for both variants) (Table 2, Figure 3). Transduction efficiencies in the liver with rAAV1.9-3-1.5 and in the heart with rAAV1.9-6 did not show any statistically significant difference compared to those with rAAV9. The AAV1.9-3-1.5 capsid differed from the AAV1 capsid by a total of 22 a.a. that reside in two capsid segments separated by a 73-a.a. stretch (Figure 2a). Importantly, these two segments associate closely in a quaternary structure, forming the outer spike and wall of the protrusions on the capsid (Figure 2e). This suggests that the outer region of the capsid protrusions plays a key role in rAAV9-like robust hepatic transduction. However, this region of rAAV9 should not be an absolute requirement for the hepatic robustness of rAAV9 because rAAV9.1-3 with the outer region being derived from rAAV1 also exhibited robust hepatic transduction. On the other hand, the AAV1.9-6 capsid had a 37-a.a. stretch of the AAV9 capsid on the AAV1 capsid backbone at the very C-terminus, which differs by 11 a.a. between AAV1 and AAV9. This C-terminal capsid segment contains a.a. exposed on the surface at the twofold symmetry axis of the capsid and forms the deepest valleys between the neighboring capsid protrusions (Figure 2d). A domain-swapping experiment revealed that the enhancing effect of AAV9's 37-a.a. C-terminal segment is observed only in the context of the AAV1 capsid, but not on AAV2, AAV2R585E or AAV1.9-3 (Figures 2a and 3). Injection of rAAV1.9-7 at a tenfold lower dose resulted in transduction of the liver and heart at efficiencies of 4.3% and 31.4%, respectively (data not shown), indicating that this variant retained the robustness of rAAV9 in transducing these organs. Other hybrids did not provide much information about the mechanism of rAAV9's robustness. Taken together, our observations clearly demonstrated that rAAV-mediated in vivo transduction efficiencies do not necessarily correlate with the blood clearance rates. The present study provided some insight into the capsid determinants of rAAV9's robustness; however, it did not provide conclusive results due to the context dependency of the observed effects caused by the domain swapping.

Figure 3.

Figure 3

Open in a new tab

Hepatic and cardiac transduction with various serotype and variant recombinant adeno-associated virus (rAAV) vectors. All mice were injected with AAV-CMV-lacZ vector intravenously at a dose of 1.0 × 1013 vg/kg except for AAV2.9-6, which was injected at a 1/5 dose, and AAV9SC, which was injected subcutaneously at a dose of 1.0 × 1013 vg/kg. Tissues were harvested 11 days postinjection, and analyzed by X-Gal staining. Tissue sections were counterstained with light hematoxylin.

Table 2. Transduction efficiencies in the liver and heart in mice injected with various serotype and variant AAV-CMV-lacZ vectorsa.

graphic file with name mt20113t2.jpg

Open in a new tab

Capillary walls in the heart form a capacity-limited transport bottleneck and this barrier is overcome by rAAV9's delayed blood clearance

To investigate the significance of rAAV9's substantially extended endothelial exposure in hepatic and cardiac transduction, we injected adult C57BL/6 male mice with AAV9-CMV-lacZ at a dose of 1.0 × 1013 vg/kg intravenously in bolus, and infused 100 µl of pooled mouse sera containing anti-AAV9 neutralizing antibodies 10 minutes, 1 hour, and 4 hours post-rAAV9 injection (n = 2 each). As predicted, rAAV9's pharmacokinetic profiles in the experimental groups were the same as those in the controls until anti-AAV9 sera injection, but they were altered thereafter, resulting in substantially shortened vector exposure time of the endothelium (Figure 4a). The 72-hour areas under the curve (AUCs) were significantly decreased (Figure 4b). As a result, cardiac transduction was substantially diminished with an hour latency (Figure 4c). By contrast, the nearly maximum hepatic transduction was achieved by a 4-hour endothelial exposure to the vector without a time lag (Figure 4c). These results demonstrate that a short endothelial exposure to a high concentration of rAAV9 leads to sufficient hepatic transduction but it is not sufficient to transduce cardiomyocytes efficiently, and extended vector exposure is essential for the robust cardiac transduction with rAAV9. This notion was also supported by an experiment in which rAAV9 was injected via subcutaneous route. The subcutaneous route was used as a means to decrease peak blood vector concentrations while sustaining vector circulation in the bloodstream (Supplementary Figure S3). Even with substantially lower blood vector concentrations during the first 24 hours after vector administration, sustained vector exposure by subcutaneous injection could mediate efficient heart transduction (Figure 3 and Table 2). To exclude an alternative possibility that rAAV9 virions were not fully capable of transducing cardiomyocytes in vivo at the time of vector injection, and became activated in the blood thereafter by a process that took a while, the experiment was repeated using rAAV9 particles that were preincubated in mouse blood in vivo or in mouse serum in vitro for 4 hours, showing no effects from the preincubation (data not shown). These results indicate that a substantial number of rAAV9 particles can cross liver sinusoidal endothelial cells in 4 hours, while transvascular transport of rAAV9 is a capacity-limited slow process, forming a vector transport bottleneck in the heart. This conforms to the difference in the anatomical structures of capillary walls in the liver and the heart. Liver sinusoidal endothelial cells form a discontinuous cell layer with many fenestrae of 100–200 nm in diameter.21 In contrast, capillary endothelium in the heart forms a tightly sealed continuous cell lining. Importantly, rAAV9 has effectively overcome this barrier in the heart by its delayed blood clearance.

Figure 4.

Figure 4

Open in a new tab

Relationship between the extent of endothelial vector exposure and transduction efficiency in the liver and heart. Mice were injected intravenously with 1.0 × 1013 vg/kg of (a–c) AAV9-CMV-lacZ or (d–f) AAV1.9-6-CMV-lacZ vector, then circulating vector particles were rapidly eliminated at designated time points by infusion of anti-AAV9 or AAV1.9-6 mouse sera (n = 2 each). The vector-injected control mice either did not receive any mouse sera, or received naive mouse sera 10 minutes postvector injection. (a,d) Blood vector concentration-time curves. Mouse sera containing anti-AAV9 or AAV1.9-6 capsid antibodies were infused 10 minutes, 1 hour or 4 hours postvector injection. (b,e) 72-hour areas under the curve (AUCs) in each group relative to those in the mice that received no mouse sera. (c,f) Liver and heart transduction efficiencies relative to those in the mice that received no mouse sera. Vertical bars represent differences between means and each value. (g,h) Cardiomyocyte transduction efficiencies with rAAV9 and rAAV1.9-6 in vitro. H9c2 cells were transduced with either AAV9-CMV-lacZ or AAV1.9-6-CMV-lacZ at a multiplicity of infection of 106, and transduction efficiencies were determined by (g) X-Gal staining and (h) β-galactosidase-specific enzyme-linked immunosorbent assay 9 days after infection. (i) Relative transduction efficiencies in the liver and heart by full or 8-hour endothelial exposure with either AAV9-CMV-lacZ or AAV1.9-6-CMV-lacZ. Anti-AAV9 or AAV1.9-6 sera were infused in mice 8 hours after rAAV9 or rAAV1.9-6 injection, respectively (n = 6 each). No sera control animals received only rAAV (n = 6 each). Vertical bars represent standard errors.

Capillary walls in the heart also pose a significant barrier to rAAV1.9-6 that mediates robust cardiac transduction with significantly limited endothelial vector exposure

rAAV1.9-6 exhibited rAAV9-like cardiac transduction efficiency with significantly shortened and reduced endothelial exposure to the vector. Therefore, we assumed that rAAV1.9-6 might efficiently cross the capillary walls in the heart. To test this assumption, we performed the same experiment as we did for rAAV9, in which we altered rAAV1.9-6's blood vector concentration-time curve by infusing anti-AAV1.9-6 neutralizing antibody (Figure 4d,e). Nearly maximum hepatic transduction could be attained with 4-hour exposure to rAAV1.9-6; however, 10-minute, 1-hour and 4-hour exposure resulted in significantly limited transduction in the heart (Figure 4f). These observations do not conform to our assumption described above. Rather, the results demonstrate that transvascular transport of rAAV1.9-6 takes at least 1 hour and has limited capacity as seen with rAAV9.

The robustness of rAAV1.9-6-mediated cardiac transduction is presumably due in part to its high capacity per unit of time to transduce cardiomyocytes

To address the mechanism of the cardiac robustness of rAAV1.9-6 with limited endothelial exposure, we investigated whether rAAV1.9-6 exhibits better transduction than rAAV9 when cardiomyocytes are directly exposed to rAAV particles to the same extent by taking an in vitro approach. H9c2 cardiomyocytes were exposed to AAV9-CMV-lacZ or AAV1.9-6-CMV-lacZ at the same dose and for the same duration of time. As a result, we found that rAAV1.9-6 transduced H9c2 cells more efficient than rAAV9 with a given extent of the vector exposure (P < 0.001) (Figure 4g,h). Although the biology of rAAV in vitro and in vivo could be different, this observation suggests that rAAV1.9-6 is more potent than rAAV9 in transducing cardiomyocytes once viral particles cross vascular wall barriers and reach cardiomyocytes, allowing efficient cardiac transduction with a limited vector exposure of rAAV1.9-6. To support this, when anti-AAV sera were infused 8 hours after rAAV injection in the same manner as described above, there emerged a trend that rAAV1.9-6 exhibits more accelerated cardiac transduction than rAAV9 (Figure 4i), although this trend was not statistically significant in the sample size we used (P = 0.11, n = 6).

Caveolin-1–independent transvascular transport of rAAV9 mediated robust cardiac transduction in mice

The results of the above experiments indicated significant involvement of a transendothelial transport system in rAAV9 and rAAV1.9-6–mediated cardiac transduction. Because transcytosis of rAAV has been demonstrated in vitro20 and caveolin-1–dependnent caveolae-mediated transcytosis is the major macromolecule transport pathway across the endothelial cell barrier,22,23 we investigated the role of the caveolin-1–dependent transport pathway in rAAV9-mediated in vivo transduction. To this end, we injected caveolin-1–deficient mice devoid of caveolae-mediated transcytosis and the control wild-type mice with rAAV9 via the tail vein in bolus. The result showed that caveolin-1–deficient mice cleared rAAV9 from the bloodstream faster than the wild-type controls during the first 4 hours following vector injection (Figure 5a), but the vector was cleared slowly in both groups thereafter with clearance half-lives t1/2c:1–24 hours of 15.8 ± 1.5 hours (R2 = 0.87) and 11.6 ± 1.2 hours (R2 = 0.95) in the caveolin-1–deficient mice and the wild-type control, respectively. The overall pharmacokinetic profiles looked similar in both groups with 32% decrease of 72-hour AUC in the caveolin-1–deficient mice. Besides the slight difference in the initial pharmacokinetic profile, neither blood clearance rate nor tissue transduction efficiency was substantially affected in the caveolin-1–knockout mice (Figure 5b). These results indicate that rAAV9 can traverse the endothelial cell barrier via a caveolin-1–independent pathway, resulting in robust cardiac transduction. Further studies will be needed to identify the transvascular transport pathway of rAAV9.

Figure 5.

Figure 5

Open in a new tab

Effects of the absence of caveolin-1 (Cav-1) on recombinant adeno-associated virus serotype 9's (rAAV9) blood vector clearance and hepatic and cardiac transduction in mice. (a) Blood vector concentration-time curves following intravenous injection of AAV9-CMV-lacZ at a dose of 1.0 × 1013 vg/kg in caveolin-1–deficient mice and wild-type controls (n = 3 each). Asterisks indicate the time points showing a difference with statistical significance (Student's t-test, P < 0.05). (b) Transduction efficiency in the liver and heart of wild-type and caveolin-1–deficient mice injected with AAV9-CMV-lacZ (n = 5 each). Transduction efficiencies were determined by X-Gal staining of tissue sections. Vertical bars represent standard errors. KO, knockout; WT, wild type.

Discussion

The primary objective in the present study was to understand the roles of potential extracellular barriers in rAAV vector–mediated in vivo transduction by a systemic approach. The barriers we focused on were (i) the clearance of infused vector particles from the bloodstream; and (ii) the capillary and sinusoidal endothelia that physically separate target parenchymal cells from the bloodstream. The study demonstrates that, although many rAAV serotypes and capsid-modified variants are rapidly cleared from the blood circulation, rAAV9 persistently circulates in the bloodstream. This conforms to the observation by Zincarelli et al. that rAAV9 was cleared from the blood more slowly than rAAV1, 4, 6, 7, and 8.24 Importantly, we showed indirect evidence suggesting that capillary wall in the heart forms a significant barrier even to rAAV9 and potentially to any rAAV serotypes. This barrier would limit the number of rAAV particles that cross the capillary wall in a given time and delays the onset of infection to parenchymal cells, causing 1-hour latency. We find that rAAV9 has conquered the barrier by prolonging the vector contact to endothelial cells and provides the vector sufficient time for passing through the transvascular transport bottleneck while rAAV1.9-6 appears to take advantage of its enhanced potency that enables efficient transduction with limited vector exposure. The observed barrier in cardiac transduction might be interpretable by alternative cardio-specific mechanisms that involve cell surface viral receptors, intracellular trafficking and metabolism of vectors; however, none of these mechanisms other than the capillary wall barrier model better explain our experimental observations.

Our study identified the 113-a.a. AAV9 capsid region that potentially delays blood clearance in the context of rAAV1. Interestingly, rAAV1.9-3, rAAV1.9-3-2, and rAAV1.9-3-4, which had a partial or total a.a exchange within the 113-a.a. capsid region, exhibited significantly impaired infectivity. This indicates that these capsid modifications somehow abolished the interaction between the viral capsid and its cognate cell surface receptor. a-2,3 and a-2,6 N-linked sialic acids have recently been identified as primary receptors for AAV1.25 Although the AAV1 capsid region responsible for sialic acid binding has yet to be identified, the modifications made to create rAAV1.9-3, rAAV1.9-3-2, and rAAV1.9-3-4 appear to have directly or indirectly disturbed the sialic acid binding motif leading to a significant loss of infectivity. In this regard, one may argue that rAAV1.9-3's delayed blood clearance would merely be a consequence of substantial reduction of viral particle sequestration by cellular receptor binding, and therefore the 113-a.a. region by itself would not account for the slow blood clearance. However, the fact that rAAV1.9-3-2 and rAAV1.9-3-4 showing significant loss of infectivity were cleared rapidly as rAAV1 does not fully support this argument. In addition, the corresponding region of the AAV1 capsid accelerated the blood clearance of rAAV9 in a reciprocal domain-swapping experiment. Taken together, it is plausible that the 113-a.a. region contains the primary culprit in accelerating or delaying blood clearance by enhancing or minimizing the contact to the host, respectively. This effect could be uncoupled with the potency of rAAV infection to target cells.

It should be noted that rAAV2-derived heparin-binding mutants, rAAV2R585E and rAAV2R585E.9-6, exhibited distinctively delayed blood clearance. A study by Asokan et al. has also shown that a rAAV2-derived variant with its heparin-binding motif replaced with an AAV8-derived segment devoid of heparin-binding ability exhibits markedly reduced blood clearance.26 Although the rapid blood clearance of rAAV2 has consistently been demonstrated in the present study and by others,24,27 we found that rAAV2's clearance half-life in the blood was paradoxically slow (t1/2c:1–24hours = 7.6 hours). The rapid blood clearance of rAAV2 was primarily due to its very short distribution half-life (t1/2d:1–30 minutes = 0.3 hour). rAAV2's primary receptor, heparan sulfate proteoglycan, is abundantly present in the extracellular matrix as well as on cell surface membranes; therefore, the initial rapid decline of rAAV2 concentrations in the blood following intravascular injection could be interpreted by rAAV2 sequestration by extracellular matrix.27,28 Based on these pharmacokinetic observations, we propose a hypothetical model in which the default blood clearance rate of rAAV should be slow unless rAAV capsid is given a trait that accelerates its clearance, such as the heparin-binding motif on the AAV2 capsid. In this model, the delayed blood clearance of rAAV1.9-3 could be explained by a removal of not-yet-defined blood clearance-accelerating motifs present in the outer region of the AAV1 capsid threefold protrusions, and the failure to delay the clearance in rAAV1.9-3-1, -2, -3, -4, -5, and -1.5 would suggest that such motifs reside at multiple locations within the outer region of the AAV1 capsid protrusions.

Di Pasquale and Chiorini have demonstrated that, using an in vitro transwell cell culture chamber system, AAV serotypes 4 and 5 and bovine AAV can penetrate polarized epithelial and endothelial monolayer cells via transcytosis in a cell-type specific manner, while AAV2 and AAV6 do not cross the barrier at appreciable levels.20 On the other hand, another group reported that AAV2, AAV8, and AAV9 could transcytose through human umbilical vein endothelial cell monolayer with AAV8 and 9 exhibiting more efficient penetration than AAV2.29 Virus transcytosis has been identified as a mechanism by which HIV type 1, hepatitis B virus, and poliovirus cross epithelial cell barriers and facilitate virus transmission.30,31,32 Thus, virus endothelial transcytosis does occur at least in vitro; however, how it occurs in vivo as well as its biological significance in animals and humans remain largely unknown. Our in vivo study revealed that transvascular rAAV transport takes an hour. Transcytosis has been reported as a rapid process but it still takes 10–30 minutes or longer depending on the studies.20,32,33,34 Although the present study by itself did not address whether transcytosis is the mechanism of rAAV transport across the capillary wall, our observation that there was a lag period of this time range indicated the involvement of the transcytosis pathway in rAAV transduction in mice.

Caveolin-1 is a key molecule in caveolae-mediated transcytosis;35 therefore, we investigated the role of caveolin-1 in rAAV9-mediated in vivo transduction using caveolin-1–deficient mice that lacked caveolae.36,37 In caveolin-1–deficient mice, transendothelial transport of macromolecules via transcytosis is significantly impaired; however, albumin, which is primarily transported by caveolae-mediated transcytosis,38 can still efficiently cross the endothelial barrier owing to the activation of an alternative transendothelial transport pathway, making these genetically deficient animals viable.36,37,39,39 The more rapid decline of rAAV9 concentrations we observed in the distribution phase in caveolin-1–deficient mice may indicate the presence of a paracellular transport pathway for rAAV9. However, the temporary nature of the observed accelerated clearance had a minimum effect on the overall pharmacokinetics of rAAV9. Although more animals will be needed to further investigate the small but potentially significant caveolin-1–associated differences in the pharmacokinetics, our observation that there was no substantial difference in the overall rAAV9 blood clearance rates and transduction efficiencies in tissues between wild-type and caveolin-1–deficient mice indicates that transvascular transport of rAAV9 is likely caveolin-1–independent in wild-type mice as well as in cavelin-1–deficient animals.

In summary, the results of our present study have advanced our understanding of the roles of extracellular barriers including blood vector clearance and capillary wall barriers in rAAV-mediated systemic gene delivery and provided a clue to elucidate how rAAV9 and other types of robust variants overcome the barriers. In addition, our observations may have important implication in the potential clinical use of plasmapheresis to reduce circulating anti-AAV neutralizing antibody levels41 because, unlike plasmapheresis in adenoviral vector–mediated gene delivery,42 more prolonged reduction of antibody titers would become essential for rAAV9-mediated nonhepatic gene transfer in anti-AAV9 antibody-positive patients.

Materials and Methods

DNA constructs and rAAV vector production. Production and purification of AAV-CMV-lacZ vectors were described previously.8 Various AAV1 and AAV9 hybrid helper plasmids were created by replacing the corresponding region of AAV1 helper plasmid DNA with either restriction enzyme-restricted AAV9 helper plasmid DNA or DNA fragments generated by PCR.

Animal studies. All the animal experiments were performed according to the guidelines for animal care at University of Pittsburgh. C57BL/6 male mice, caveolin-1–deficient male mice (Cav1tm1Mls/J) and their control male mice (B6129SF2/J) used for the study were purchased from the Jackson Laboratory (Bar Harbor, ME). rAAV vector in a final volume of 10 ml/kg of phosphate-buffered saline with 5% sorbitol was infused into adult animals via the tail vein in bolus at a dose of 1.0 × 1013 vg/kg. Unless otherwise noted, this dose was used in all the pharmacokinetic studies in which blood samples were collected at multiple time points. In some experiments, the vector preparation was injected subcutaneously under the dorsal skin at a dose of 1.0 × 1013 vg/kg in a final volume of 10 ml/kg of phosphate-buffered saline with 5% sorbitol, or the vector injection was followed by tail vein injection of 100 µl of either pooled naive mouse sera or sera containing neutralizing antibody against the AAV9 or AAV1.9-6 capsid.

The anti-AAV9 and AAV1.9-6 sera were obtained from mice intravenously injected with corresponding rAAV vectors at a dose of 1.0 × 1011 vg/mouse. To confirm the ability of the pooled anti-AAV9 and AAV1.9-6 sera to neutralize the corresponding rAAVs, 7.5 × 1011 vg of the corresponding AAV-CMV-lacZ vectors were incubated in the presence of 125 µl of either pooled naive mouse sera or anti-AAV sera at 37 °C for 1 hour in a final volume of 250 µl of the excipient, and then sera-treated rAAV particles in a final volume of 10 ml/kg of phosphate-buffered saline/5% sorbitol were infused into mice at a dose of 1.0 × 1013 vg/kg. This incubation completely blocked in vivo transduction in mice. In addition, we confirmed that infusion of 100 µl of the anti-AAV mouse sera can eliminate >98% of circulating viral particles within 20 minutes in mice injected with rAAV at a dose of 1.0 × 1013 vg/kg. Furthermore, we confirmed that infusion of antisera 4 days after rAAV injection had no effect on in vivo transduction. In one experiment, rAAV9 vector was incubated in the whole blood in C57BL/6 mice or in C57BL/6 mouse sera in vitro for 4 hours, and 100 µl of sera containing 7.8 × 1010 vg of AAV9-CMV-lacZ was infused via the tail vein immediately after the completion of the in vivo and in vitro incubation of rAAV9. Due to the nature of the experiment, the dose to be injected was retrospectively determined. For the in vivo incubation of rAAV9 in the whole blood, C57BL/6 mice were injected with rAAV9 at a dose of 1.0 × 1012 vg/mouse via the tail vein, and rAAV9-containing sera were collected 4 hours after injection. For the in vitro incubation, rAAV9 particles were incubated at 37 °C for 4 hours in a pooled naive mouse serum at a concentration of 7.8 × 1011 vg/ml. Blood samples were collected from saphenous vein or retro-orbital plexus 1 minute, 10 minutes, 30 minutes, 1 hour, 4 hours, 8 hours, 24 hours, and 3 days postinjection, using heparin-coated capillary tubes. For the vector bioactivity assay, plain capillary tubes were used for blood sampling. All the animals were killed 11 days postinjection for histological analysis as described.5 This time point was chosen based on our previous observations5,43 to maximize transduction efficiencies and avoid cytotoxic T lymphocyte responses against rAAV-transduced cells. It should be noted that this experimental design might have somewhat underestimated transduction efficiencies with certain serotypes or variants, particularly rAAV2 exhibiting a slow rise of transgene expression in vivo. In another experiment, following tail vein injection of rAAV9 at a dose of 3.0 × 1012 vg/mouse, the vector was incubated in the bloodstream in C57BL/6 mice for 24 hours, and 160 µl of sera containing 5.6 × 1010 vg of AAV9-CMV-lacZ (the titer was retrospectively determined) was infused via the tail vein immediately after the completion of the in vivo incubation. As a control, the same dose of rAAV9 from the vector stock was injected in mice with no prior incubation. The animals were killed 11 days postinjection to determine the transduction efficiency in the liver and heart.

Cell culture assays. Human embryonic kidney AAV293 cells and rat cardiomyocyte–derived H9c2 cells were purchased from Stratagene (La Jolla, CA) and ATCC (Manassas, VA), respectively. 84–31 cells, a 293-derived cell line constitutively expressing adenovirus E1 and E4 genes, were kindly provided by James M. Wilson, University of Pennsylvania. They were maintained in Dulbecco's Modified Eagle Medium supplemented with 10% fetal bovine serum. In vitro rAAV vector bioactivity assay was performed in the following manner. DNase I-resistant particle number in 25 µl of each mouse serum sample collected 24 hours postinjection of 4.0 × 1011 vg/mouse in 300 µl of phosphate-buffered saline/5% sorbitol was determined by a quantitative dot blot assay. The particle titers in 25 µl serum were 2.4 and 3.8 × 108 vg for rAAV2, 1.7 and 1.5 × 108 vg for rAAV8, and 2.0, 2.2, 2.8, 5.0, and 5.2 × 109 vg for rAAV9 (each value represents the titer in the samples obtained from different animals). 84-31 cells were seeded on a collagen-coated 96-well plate at 2.0 × 104 cells/well. Twenty-four hours later, 25 µl of rAAV-injected mouse serum samples or 25 µl of naive mouse sera plus vector activity standards were loaded in duplicate onto each well of 96-well plates, and incubated for 1 hour. Bioactivity standards were created such that, for 100% activity standards, cells were infected with the same amount of each rAAV vector from the stocks as the amount of rAAV in each 25 µl sample; therefore, the bioactivity standard series were generated for each sample. rAAV transduction efficiency was determined 48 hours postinfection by 5-bromo-4-chloro-3-indolylphosphate (X-Gal) staining. The assay was repeated at least twice. Heat-inactivated fetal bovine sera were used in the culture media. In vitro infectivity of rAAV9 and rAAV1.9-6 to cardiomyocytes was assessed as follows. H9c2 cells were seeded in 96-well plates at 10,000 cells per well and allowed to grow to confluence. The cells were exposed to rAAV vectors expressing the lacZ gene at a multiplicity of infection of ~106 for 3 days in quintuplicate. Nine days postinfection, transduction efficiencies were determined by X-Gal staining (two wells each) and quantification of the transgene products in cell lysates (three wells each) using a bacterial β-galactosidase-specific enzyme-linked immunosorbent assay kit (Roche Diagnostics, Indianapolis, IN) as previously described.44 Protein concentrations of the cell lysates were determined by DC Protein Assay Kit (Bio-Rad, Hercules, CA) and used to normalize the values obtained by the β-galactosidase-specific enzyme-linked immunosorbent assay.44

Blood assays. The number of DNase I-resistant rAAV particles in each blood sample was determined by a quantitative dot blot assay. Briefly, 14 µl of heparinized whole blood was diluted with DNase I digestion buffer in a total volume of 99 µl, underwent three freeze-thaw cycles, treated with 10 units of DNase I at 37 °C for 1 hour in a total volume of 100 µl, and then treated with 1 mg/ml Proteinase K at 55 °C for 1 hour in a total volume of 200 µl. Following phenol-chloroform extraction, 80% of aqueous phase was recovered, ethanol precipitated, and dissolved in 1 × TE (10 mmol/l Tris HCl pH 8.0, 1 mmol/l EDTA). The purified DNA was denatured in the presence of 0.4 N NaOH, blotted onto a Zeta-Probe membrane (Bio-Rad), and hybridized with a 32P-labeled lacZ probe. Serially diluted linearized plasmid DNA containing the lacZ gene was used as a standard. Dot blot signals were quantified as previously described.5 The sensitivity of our dot blot assay was 2 × 109 vector genomes per ml of blood. Vector half-lives were determined by log-linear regression of blood vector concentrations versus time through the corresponding time points using the least squares method. To do this, we used the Excel LINEST function. The clearance half-lives (t1/2c:1–24 hours) were determined in individual mice, while distribution half-lives (t1/2d:1–30 minutes) were determined using the average values of blood vector concentrations at each time point due to sample-to-sample variations in the early vector distribution phase. Area under the blood vector concentration-time curve (AUC) was defined as a value given by a definite integral between the blood vector concentration-time curve and the horizontal axis over a specific interval. We determined 72-hour AUCs (i.e., AUCs during the 72 hours following vector administration) using a trapezoidal rule. The trapezoidal rule is a numerical method to approximate a definite integral between two points by finding the sum of the areas of trapezoids. In the mice infused with neutralizing antibodies, the following consideration was taken into account in AUCs that over 98% circulating rAAV particles were eliminated from the bloodstream in 20 minutes following antibody infusion.

Analyses of tissue transduction. Frozen tissue sections were stained with X-Gal as previously described.5 Transduction efficiency in each sample was determined by counting X-Gal positive cells for the liver, and by an image analysis using MetaMorph software for the heart as previously described.5 rAAV vector genome copy numbers in the livers were determined by Southern blot analysis as previously described.8 Vector genome copy numbers in each cell were expressed as double-stranded vector genome copy numbers per diploid genomic equivalent.

Computer modeling of AAV capsids. Three-dimensional models of AAV viral capsid proteins were generated using the SWISS-MODEL program based on the known capsid structures obtained from the Protein Data Bank (AAV2, 1lp3). Sixty monomers of each modeled AAV variant capsid were assembled into a complete full capsid comprised of 60 monomers and visualized using the Multistage Model function of UCSF Chimera.45

Statistical analyses. All data were presented as means ± SEM unless otherwise noted. Statistical significance of differences in the observed values between two experimental groups was evaluated by two-tailed Student's t-test where appropriate. In comparison of blood clearance rates between rAAV1, rAAV2, rAAV8, and rAAV9, the significance of differences in the data was evaluated by a Kruskal–Wallis analysis followed by a Dunn's post hoc test using GraphPad RRISM software. A P value of <0.05 was considered to be significant.

SUPPLEMENTARY MATERIAL Figure S1. Pharmacokinetics of recombinant adeno-associated virus serotype 6 (rAAV6), 7, and rh10. Blood vector concentration-time curves following intravenous vector injection in mice. rAAV6, rAAV7, and rAAVrh.10 carrying the AAV-EF1α-nlslacZ vector genome [8] were injected into C57BL/6 male mice via the tail vein in bolus at a dose of 1.0 x 1013 vg/kg in a final volume of 10 ml/kg of PBS/5% sorbitol (n=2 each). Concentration of DNase I-resistant rAAV vector particles in whole blood were determined by a quantitative dot blot analysis and plotted as a function of time after injection. Vertical bars indicate the difference between mean value and each value. Figure S2. Infectivity of recombinant adeno-associated virus serotype 2 (rAAV2), 8, and 9 in mouse blood. Results of the assays to investigate the infectivity of rAAV vector particles in the vector-injected mouse blood recovered 24 h post-injection. (a) Results of the in vitro infectivity assay. rAAV2, rAAV8 and rAAV9 vectors carrying the AAV-CMV-lacZ vector genome were injected into C57BL/6 male mice via the tail vein in bolus at a dose of 4.0 x 1011 vg/mouse (n=2 each for rAAV2 and rAAV8, and n=5 for rAAV9). Sera were collected 24 h post-injection, and DNase I-resistant rAAV vector particles were quantified in each serum sample. Twenty-five μl of each sample was applied in duplicate on 84-31 cells that have grown on a 96-well plate, and transduction efficiency was determined 48 h post-infection by X-Gal staining. The infectivity standards were created such that if a sample showed the same transduction efficiency as that of the 100% bioactivity standard, there should be no loss of infectivity. Representative photomicrographs are shown. We found that adding mouse sera in the culture media (25 μl sera in 200 μl culture media) somewhat influenced rAAV transduction efficiency. The degrees of increase or decrease in transduction efficiency appeared to be consistent within the same control or sample serum but could vary to some extent among different controls or samples, as exemplified by the two experiments using different lots of control naïve mouse sera in the infectivity standards (Exp.3 and Exp.4 in the figure). Therefore, this in vitro infectivity assay only gives us a rough estimate. Nonetheless, it was obvious that rAAV8 and 9 maintained readily measurable infectivity, while the infectivity of rAAV2 was almost lost. (b) Results of the in vivo infectivity assay. C57BL/6 mice received intravenously either 24-h in vivo incubated AAV9-CMV-lacZ or the same vector from the vector stock at a dose of 5.6 x 1010 vg/mouse (n=3 each). Transduction efficiencies in the liver and heart were determined histologically 11 days post-injection. There is no statistically significant difference in transduction efficiencies between the two groups in this sample size by Student's t-test, indicating that infectivity of rAAV9 was substantially retained in the bloodstream. Vertical bars represent standard errors. Figure S3. Pharmacokinetics of subcutaneously injected recombinant adeno-associated virus serotype 9 (rAAV9). Pharmacokinetic profiles of subcutaneously injected rAAV9. Mice were injected with AAV9-CMV-lacZ subcutaneously under the dorsal skin at a dose of 1.0 x 1013 vg/kg in a final volume of 10 ml/kg of PBS/5% sorbitol (SC group, n=4). The data for intravenously injected rAAV9 (IV group, n=7) is included for comparison. Vertical bars represent standard errors.

Acknowledgments

We thank Guangping Gao and James M. Wilson for providing various AAV serotype packaging plasmids and 84-31 cells, Simon C. Watkins for fruitful discussions, Congrong Ma, Chuncheng Piao, Yunqing Kan and Marni Hershbain for technical assistance, and Christopher S. Naitza for assistance in the preparation of the manuscript. This work was supported by Public Health Service grants DK078388 (to H.N.) and AR050733 (to Joseph C. Glorioso) from National Institutes of Health and Career Development Award from National Hemophilia Foundation (to H.N.).

Supplementary Material

Figure S1.

Pharmacokinetics of recombinant adeno-associated virus serotype 6 (rAAV6), 7, and rh10. Blood vector concentration-time curves following intravenous vector injection in mice. rAAV6, rAAV7, and rAAVrh.10 carrying the AAV-EF1α-nlslacZ vector genome [8] were injected into C57BL/6 male mice via the tail vein in bolus at a dose of 1.0 x 1013 vg/kg in a final volume of 10 ml/kg of PBS/5% sorbitol (n=2 each). Concentration of DNase I-resistant rAAV vector particles in whole blood were determined by a quantitative dot blot analysis and plotted as a function of time after injection. Vertical bars indicate the difference between mean value and each value.

Click here for additional data file. (655.8KB, tiff)
Figure S2.

Infectivity of recombinant adeno-associated virus serotype 2 (rAAV2), 8, and 9 in mouse blood. Results of the assays to investigate the infectivity of rAAV vector particles in the vector-injected mouse blood recovered 24 h post-injection. (a) Results of the in vitro infectivity assay. rAAV2, rAAV8 and rAAV9 vectors carrying the AAV-CMV-lacZ vector genome were injected into C57BL/6 male mice via the tail vein in bolus at a dose of 4.0 x 1011 vg/mouse (n=2 each for rAAV2 and rAAV8, and n=5 for rAAV9). Sera were collected 24 h post-injection, and DNase I-resistant rAAV vector particles were quantified in each serum sample. Twenty-five μl of each sample was applied in duplicate on 84-31 cells that have grown on a 96-well plate, and transduction efficiency was determined 48 h post-infection by X-Gal staining. The infectivity standards were created such that if a sample showed the same transduction efficiency as that of the 100% bioactivity standard, there should be no loss of infectivity. Representative photomicrographs are shown. We found that adding mouse sera in the culture media (25 μl sera in 200 μl culture media) somewhat influenced rAAV transduction efficiency. The degrees of increase or decrease in transduction efficiency appeared to be consistent within the same control or sample serum but could vary to some extent among different controls or samples, as exemplified by the two experiments using different lots of control naïve mouse sera in the infectivity standards (Exp.3 and Exp.4 in the figure). Therefore, this in vitro infectivity assay only gives us a rough estimate. Nonetheless, it was obvious that rAAV8 and 9 maintained readily measurable infectivity, while the infectivity of rAAV2 was almost lost. (b) Results of the in vivo infectivity assay. C57BL/6 mice received intravenously either 24-h in vivo incubated AAV9-CMV-lacZ or the same vector from the vector stock at a dose of 5.6 x 1010 vg/mouse (n=3 each). Transduction efficiencies in the liver and heart were determined histologically 11 days post-injection. There is no statistically significant difference in transduction efficiencies between the two groups in this sample size by Student's t-test, indicating that infectivity of rAAV9 was substantially retained in the bloodstream. Vertical bars represent standard errors.

Click here for additional data file. (210KB, pdf)
Figure S3.

Pharmacokinetics of subcutaneously injected recombinant adeno-associated virus serotype 9 (rAAV9). Pharmacokinetic profiles of subcutaneously injected rAAV9. Mice were injected with AAV9-CMV-lacZ subcutaneously under the dorsal skin at a dose of 1.0 x 1013 vg/kg in a final volume of 10 ml/kg of PBS/5% sorbitol (SC group, n=4). The data for intravenously injected rAAV9 (IV group, n=7) is included for comparison. Vertical bars represent standard errors.

Click here for additional data file. (1.2MB, tiff)

REFERENCES

  1. Gao G, Vandenberghe LH, Alvira MR, Lu Y, Calcedo R, Zhou X.et al. (2004Clades of Adeno-associated viruses are widely disseminated in human tissues J Virol 786381–6388. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Vandendriessche T, Thorrez L, Acosta-Sanchez A, Petrus I, Wang L, Ma L.et al. (2007Efficacy and safety of adeno-associated viral vectors based on serotype 8 and 9 vs. lentiviral vectors for hemophilia B gene therapy J Thromb Haemost 516–24. [DOI] [PubMed] [Google Scholar]
  3. Ghosh A, Yue Y, Long C, Bostick B., and, Duan D. Efficient whole-body transduction with trans-splicing adeno-associated viral vectors. Mol Ther. 2007;15:750–755. doi: 10.1038/sj.mt.6300081. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Pacak CA, Mah CS, Thattaliyath BD, Conlon TJ, Lewis MA, Cloutier DE.et al. (2006Recombinant adeno-associated virus serotype 9 leads to preferential cardiac transduction in vivo Circ Res 99e3–e9. [DOI] [PubMed] [Google Scholar]
  5. Inagaki K, Fuess S, Storm TA, Gibson GA, Mctiernan CF, Kay MA.et al. (2006Robust systemic transduction with AAV9 vectors in mice: efficient global cardiac gene transfer superior to that of AAV8 Mol Ther 1445–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Zhu T, Zhou L, Mori S, Wang Z, McTiernan CF, Qiao C.et al. (2005Sustained whole-body functional rescue in congestive heart failure and muscular dystrophy hamsters by systemic gene transfer Circulation 1122650–2659. [DOI] [PubMed] [Google Scholar]
  7. Wang Z, Zhu T, Qiao C, Zhou L, Wang B, Zhang J.et al. (2005Adeno-associated virus serotype 8 efficiently delivers genes to muscle and heart Nat Biotechnol 23321–328. [DOI] [PubMed] [Google Scholar]
  8. Nakai H, Fuess S, Storm TA, Muramatsu S, Nara Y., and, Kay MA. Unrestricted hepatocyte transduction with adeno-associated virus serotype 8 vectors in mice. J Virol. 2005;79:214–224. doi: 10.1128/JVI.79.1.214-224.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Sarkar R, Mucci M, Addya S, Tetreault R, Bellinger DA, Nichols TC.et al. (2006Long-term efficacy of adeno-associated virus serotypes 8 and 9 in hemophilia a dogs and mice Hum Gene Ther 17427–439. [DOI] [PubMed] [Google Scholar]
  10. Thomas CE, Storm TA, Huang Z., and, Kay MA. Rapid uncoating of vector genomes is the key to efficient liver transduction with pseudotyped adeno-associated virus vectors. J Virol. 2004;78:3110–3122. doi: 10.1128/JVI.78.6.3110-3122.2004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Nathwani AC, Gray JT, Ng CY, Zhou J, Spence Y, Waddington SN.et al. (2006Self-complementary adeno-associated virus vectors containing a novel liver-specific human factor IX expression cassette enable highly efficient transduction of murine and nonhuman primate liver Blood 1072653–2661. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Gregorevic P, Allen JM, Minami E, Blankinship MJ, Haraguchi M, Meuse L.et al. (2006rAAV6-microdystrophin preserves muscle function and extends lifespan in severely dystrophic mice Nat Med 12787–789. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Bostick B, Yue Y, Lai Y, Long C, Li D., and, Duan D. Adeno-associated virus serotype-9 microdystrophin gene therapy ameliorates electrocardiographic abnormalities in mdx mice. Hum Gene Ther. 2008;19:851–856. doi: 10.1089/hum.2008.058. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Townsend D, Blankinship MJ, Allen JM, Gregorevic P, Chamberlain JS., and, Metzger JM. Systemic administration of micro-dystrophin restores cardiac geometry and prevents dobutamine-induced cardiac pump failure. Mol Ther. 2007;15:1086–1092. doi: 10.1038/sj.mt.6300144. [DOI] [PubMed] [Google Scholar]
  15. Sun B, Young SP, Li P, Di C, Brown T, Salva MZ.et al. (2008Correction of multiple striated muscles in murine Pompe disease through adeno-associated virus-mediated gene therapy Mol Ther 161366–1371. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Ding Z, Georgiev P., and, Thöny B. Administration-route and gender-independent long-term therapeutic correction of phenylketonuria (PKU) in a mouse model by recombinant adeno-associated virus 8 pseudotyped vector-mediated gene transfer. Gene Ther. 2006;13:587–593. doi: 10.1038/sj.gt.3302684. [DOI] [PubMed] [Google Scholar]
  17. Alemany R, Suzuki K., and, Curiel DT. Blood clearance rates of adenovirus type 5 in mice. J Gen Virol. 2000;81:2605–2609. doi: 10.1099/0022-1317-81-11-2605. [DOI] [PubMed] [Google Scholar]
  18. Tao N, Gao GP, Parr M, Johnston J, Baradet T, Wilson JM.et al. (2001Sequestration of adenoviral vector by Kupffer cells leads to a nonlinear dose response of transduction in liver Mol Ther 328–35. [DOI] [PubMed] [Google Scholar]
  19. Li SD., and, Huang L. Gene therapy progress and prospects: non-viral gene therapy by systemic delivery. Gene Ther. 2006;13:1313–1319. doi: 10.1038/sj.gt.3302838. [DOI] [PubMed] [Google Scholar]
  20. Di Pasquale G., and, Chiorini JA. AAV transcytosis through barrier epithelia and endothelium. Mol Ther. 2006;13:506–516. doi: 10.1016/j.ymthe.2005.11.007. [DOI] [PubMed] [Google Scholar]
  21. Jacobs F, Wisse E., and, De Geest B. The role of liver sinusoidal cells in hepatocyte-directed gene transfer. Am J Pathol. 2010;176:14–21. doi: 10.2353/ajpath.2010.090136. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Frank PG, Pavlides S., and, Lisanti MP. Caveolae and transcytosis in endothelial cells: role in atherosclerosis. Cell Tissue Res. 2009;335:41–47. doi: 10.1007/s00441-008-0659-8. [DOI] [PubMed] [Google Scholar]
  23. Predescu SA, Predescu DN., and, Malik AB. Molecular determinants of endothelial transcytosis and their role in endothelial permeability. Am J Physiol Lung Cell Mol Physiol. 2007;293:L823–L842. doi: 10.1152/ajplung.00436.2006. [DOI] [PubMed] [Google Scholar]
  24. Zincarelli C, Soltys S, Rengo G., and, Rabinowitz JE. Analysis of AAV serotypes 1-9 mediated gene expression and tropism in mice after systemic injection. Mol Ther. 2008;16:1073–1080. doi: 10.1038/mt.2008.76. [DOI] [PubMed] [Google Scholar]
  25. Wu Z, Miller E, Agbandje-McKenna M., and, Samulski RJ. Alpha2,3 and alpha2,6 N-linked sialic acids facilitate efficient binding and transduction by adeno-associated virus types 1 and 6. J Virol. 2006;80:9093–9103. doi: 10.1128/JVI.00895-06. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Asokan A, Conway JC, Phillips JL, Li C, Hegge J, Sinnott R.et al. (2010Reengineering a receptor footprint of adeno-associated virus enables selective and systemic gene transfer to muscle Nat Biotechnol 2879–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. White SJ, Nicklin SA, Büning H, Brosnan MJ, Leike K, Papadakis ED.et al. (2004Targeted gene delivery to vascular tissue in vivo by tropism-modified adeno-associated virus vectors Circulation 109513–519. [DOI] [PubMed] [Google Scholar]
  28. Pajusola K, Gruchala M, Joch H, Lüscher TF, Ylä-Herttuala S., and, Büeler H. Cell-type-specific characteristics modulate the transduction efficiency of adeno-associated virus type 2 and restrain infection of endothelial cells. J Virol. 2002;76:11530–11540. doi: 10.1128/JVI.76.22.11530-11540.2002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. He, B, Yuan, Z, Qiao, C, Madden, V, Thakker, D, Li, Jet al. (2009Transcytosis of AAV8 and AAV9 across endothelial barrierAbstr. Mol Ther 17S175 [Google Scholar]
  30. Bomsel M., and, Alfsen A. Entry of viruses through the epithelial barrier: pathogenic trickery. Nat Rev Mol Cell Biol. 2003;4:57–68. doi: 10.1038/nrm1005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Bhat P., and, Anderson DA. Hepatitis B virus translocates across a trophoblastic barrier. J Virol. 2007;81:7200–7207. doi: 10.1128/JVI.02371-06. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Ouzilou L, Caliot E, Pelletier I, Prévost MC, Pringault E., and, Colbère-Garapin F. Poliovirus transcytosis through M-like cells. J Gen Virol. 2002;83:2177–2182. doi: 10.1099/0022-1317-83-9-2177. [DOI] [PubMed] [Google Scholar]
  33. Oh P, Borgström P, Witkiewicz H, Li Y, Borgström BJ, Chrastina A.et al. (2007Live dynamic imaging of caveolae pumping targeted antibody rapidly and specifically across endothelium in the lung Nat Biotechnol 25327–337. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Kernéis S, Bogdanova A, Kraehenbuhl JP., and, Pringault E. Conversion by Peyer's patch lymphocytes of human enterocytes into M cells that transport bacteria. Science. 1997;277:949–952. doi: 10.1126/science.277.5328.949. [DOI] [PubMed] [Google Scholar]
  35. Rothberg KG, Heuser JE, Donzell WC, Ying YS, Glenney JR., and, Anderson RG. Caveolin, a protein component of caveolae membrane coats. Cell. 1992;68:673–682. doi: 10.1016/0092-8674(92)90143-z. [DOI] [PubMed] [Google Scholar]
  36. Drab M, Verkade P, Elger M, Kasper M, Lohn M, Lauterbach B.et al. (2001Loss of caveolae, vascular dysfunction, and pulmonary defects in caveolin-1 gene-disrupted mice Science 2932449–2452. [DOI] [PubMed] [Google Scholar]
  37. Razani B, Engelman JA, Wang XB, Schubert W, Zhang XL, Marks CB.et al. (2001Caveolin-1 null mice are viable but show evidence of hyperproliferative and vascular abnormalities J Biol Chem 27638121–38138. [DOI] [PubMed] [Google Scholar]
  38. Ghitescu L, Fixman A, Simionescu M., and, Simionescu N. Specific binding sites for albumin restricted to plasmalemmal vesicles of continuous capillary endothelium: receptor-mediated transcytosis. J Cell Biol. 1986;102:1304–1311. doi: 10.1083/jcb.102.4.1304. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Schubert W, Frank PG, Woodman SE, Hyogo H, Cohen DE, Chow CW.et al. (2002Microvascular hyperpermeability in caveolin-1 (-/-) knock-out mice. Treatment with a specific nitric-oxide synthase inhibitor, L-NAME, restores normal microvascular permeability in Cav-1 null mice J Biol Chem 27740091–40098. [DOI] [PubMed] [Google Scholar]
  40. Rosengren BI, Rippe A, Rippe C, Venturoli D, Swärd K., and, Rippe B. Transvascular protein transport in mice lacking endothelial caveolae. Am J Physiol Heart Circ Physiol. 2006;291:H1371–H1377. doi: 10.1152/ajpheart.01364.2005. [DOI] [PubMed] [Google Scholar]
  41. Jiang H, Couto LB, Patarroyo-White S, Liu T, Nagy D, Vargas JA.et al. (2006Effects of transient immunosuppression on adenoassociated, virus-mediated, liver-directed gene transfer in rhesus macaques and implications for human gene therapy Blood 1083321–3328. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Rahman A, Tsai V, Goudreau A, Shinoda JY, Wen SF, Ramachandra M.et al. (2001Specific depletion of human anti-adenovirus antibodies facilitates transduction in an in vivo model for systemic gene therapy Mol Ther 3768–778. [DOI] [PubMed] [Google Scholar]
  43. Inagaki K, Piao C, Kotchey NM, Wu X., and, Nakai H. Frequency and spectrum of genomic integration of recombinant adeno-associated virus serotype 8 vector in neonatal mouse liver. J Virol. 2008;82:9513–9524. doi: 10.1128/JVI.01001-08. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Nakai H, Storm TA., and, Kay MA. Increasing the size of rAAV-mediated expression cassettes in vivo by intermolecular joining of two complementary vectors. Nat Biotechnol. 2000;18:527–532. doi: 10.1038/75390. [DOI] [PubMed] [Google Scholar]
  45. Pettersen EF, Goddard TD, Huang CC, Couch GS, Greenblatt DM, Meng EC.et al. (2004UCSF Chimera–a visualization system for exploratory research and analysis J Comput Chem 251605–1612. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Figure S1.

Pharmacokinetics of recombinant adeno-associated virus serotype 6 (rAAV6), 7, and rh10. Blood vector concentration-time curves following intravenous vector injection in mice. rAAV6, rAAV7, and rAAVrh.10 carrying the AAV-EF1α-nlslacZ vector genome [8] were injected into C57BL/6 male mice via the tail vein in bolus at a dose of 1.0 x 1013 vg/kg in a final volume of 10 ml/kg of PBS/5% sorbitol (n=2 each). Concentration of DNase I-resistant rAAV vector particles in whole blood were determined by a quantitative dot blot analysis and plotted as a function of time after injection. Vertical bars indicate the difference between mean value and each value.

Click here for additional data file. (655.8KB, tiff)
Figure S2.

Infectivity of recombinant adeno-associated virus serotype 2 (rAAV2), 8, and 9 in mouse blood. Results of the assays to investigate the infectivity of rAAV vector particles in the vector-injected mouse blood recovered 24 h post-injection. (a) Results of the in vitro infectivity assay. rAAV2, rAAV8 and rAAV9 vectors carrying the AAV-CMV-lacZ vector genome were injected into C57BL/6 male mice via the tail vein in bolus at a dose of 4.0 x 1011 vg/mouse (n=2 each for rAAV2 and rAAV8, and n=5 for rAAV9). Sera were collected 24 h post-injection, and DNase I-resistant rAAV vector particles were quantified in each serum sample. Twenty-five μl of each sample was applied in duplicate on 84-31 cells that have grown on a 96-well plate, and transduction efficiency was determined 48 h post-infection by X-Gal staining. The infectivity standards were created such that if a sample showed the same transduction efficiency as that of the 100% bioactivity standard, there should be no loss of infectivity. Representative photomicrographs are shown. We found that adding mouse sera in the culture media (25 μl sera in 200 μl culture media) somewhat influenced rAAV transduction efficiency. The degrees of increase or decrease in transduction efficiency appeared to be consistent within the same control or sample serum but could vary to some extent among different controls or samples, as exemplified by the two experiments using different lots of control naïve mouse sera in the infectivity standards (Exp.3 and Exp.4 in the figure). Therefore, this in vitro infectivity assay only gives us a rough estimate. Nonetheless, it was obvious that rAAV8 and 9 maintained readily measurable infectivity, while the infectivity of rAAV2 was almost lost. (b) Results of the in vivo infectivity assay. C57BL/6 mice received intravenously either 24-h in vivo incubated AAV9-CMV-lacZ or the same vector from the vector stock at a dose of 5.6 x 1010 vg/mouse (n=3 each). Transduction efficiencies in the liver and heart were determined histologically 11 days post-injection. There is no statistically significant difference in transduction efficiencies between the two groups in this sample size by Student's t-test, indicating that infectivity of rAAV9 was substantially retained in the bloodstream. Vertical bars represent standard errors.

Click here for additional data file. (210KB, pdf)
Figure S3.

Pharmacokinetics of subcutaneously injected recombinant adeno-associated virus serotype 9 (rAAV9). Pharmacokinetic profiles of subcutaneously injected rAAV9. Mice were injected with AAV9-CMV-lacZ subcutaneously under the dorsal skin at a dose of 1.0 x 1013 vg/kg in a final volume of 10 ml/kg of PBS/5% sorbitol (SC group, n=4). The data for intravenously injected rAAV9 (IV group, n=7) is included for comparison. Vertical bars represent standard errors.

Click here for additional data file. (1.2MB, tiff)

Articles from Molecular Therapy are provided here courtesy of The American Society of Gene & Cell Therapy

RESOURCES