Introduction

Human adenoviral (Ad) serotypes 2 and 5 (Ad2 and Ad5) bind and enter cells in vitro through a multistep process usually involving the combined interactions of the fiber and penton base proteins with their cellular receptors [the coxsackie-adenovirus receptor (CAR) and _v integrins, respectively]. Ad5 vectors efficiently transduce the liver after intravenous (IV) injection. Instead of relying on CAR interactions, Ad5 appears to bind a number of blood factors that retarget the vector to the liver.^1,2 Binding of protein C (PC), factor VII (FVII), factor IX (FIX), and factor X (FX) appear to increase in vitro transduction by Ad5 (ref. 2) suggesting that all or some of these may play a role in in vivo liver transduction. These factors have recently been shown to bind to the hexon protein and act as a bridge to retarget the virus to hepatocytes.^3,4 While blood factor binding does mediate liver transduction, Ad5 still transduces many other sites. Therefore, efforts to reduce nonspecific transduction while retaining liver specificity may improve liver gene therapy. Conversely, liver tropism is problematic for delivering the vector to other sites. Therefore, efforts to reduce liver uptake may improve vector delivery to other sites.

In contrast to IV injection, intraperitoneal (IP) Ad injection mediates poor transduction of the liver, but extensive transduction of the peritoneum by CAR-dependent interactions.⁵ This simple vector introduction site therefore has utility to modify the peritoneum for certain applications, but is a poor route if one aims to mediate liver-directed or systemic gene therapy, because CAR interactions will deplete the vector. On this basis, efforts to detarget CAR binding by Ad may improve the ability of the vector to distribute systemically after IP delivery.

Polyethylene glycol (PEG) is a hydrophilic polymer used to improve the pharmacokinetics of a variety of protein therapeutics. In these applications, the hydrophilic PEG molecule is cross-linked to the therapeutic agent to "shield" or reduce interactions of it with proteins and cells. PEG has also been applied to improve the pharmacology of Ad vectors. PEG molecules bearing reactive groups are chemically conjugated to free amine groups on the virion surface. PEGylation of Ad protects it from pre-existing neutralizing antibodies to allow multiple administration into immune recipients.^6,7,8 PEGylation also reduces the production of new antibody and cellular immune responses against Ad proteins.⁷ Furthermore, PEG reduces innate immune responses against the first generation and helper-dependent Ad5 vectors^9,10 and thromobocytopenia induced by the vector in mice.^10,11

These observations suggest that PEG blocks binding of Ad5 to a number of different cells and proteins. While one might expect this level of shielding to negatively affect Ad5 transduction, results vary in vitro and in vivo. Conjugation of Ad5 with 5 kd linear PEG molecules reduces in vitro transduction by blocking interactions of the virus with the cellular CAR.^9,12 However, while the 5 kd PEG ablates in vitro transduction, it surprisingly had no effect on in vivo transduction and vector distribution.⁹ In contrast to the effects of small 5 kd PEG, it has recently been reported that larger 20 kd PEG conjugation to Ad can reduce transduction in the liver and at other sites.¹³

In this article, we have studied how conjugation with PEGs of different size alters the liver-specific transgene expression after IV injection in mouse models. We have also tested the ability of 5 kd PEG to alter the tropism of Ad5 in vivo following IP injection via the blood factor–dependent pathway in contrast to the nonspecific CAR-dependent pathway. Finally, we investigate how PEG affects interactions of the vector with blood factors thought to mediate in vivo liver transduction, and how warfarin, a chemical that has been shown to globally deplete these blood factors,² reduced liver transgene expression following IV administration of Ad5 and PEGylated Ad5 in vivo.

Results

PEGylation increases the diameter and molecular weight of Ad

CsCl-purified Ad5 expressing luciferase and renilla green fluorescent protein, Ad-cmv-Luc-IRES-hrGFP (Ad), was conjugated to a series of amine-reactive PEG molecules, and virus size and their function were evaluated. Conjugation with each of the PEGs resulted in the modification of 65–74% of the available amines on the virus (Table 1). As expected, increasing the size of PEG also increased both the effective diameter and molecular weight of the viruses when assessed by light scattering (Table 1). Unmodified Ad had a diameter of 112 nm, whereas modification with 5, 20, and 35 kd PEG increased particle diameter to 132, 145, and 158 nm, respectively. This increase in apparent diameter paralleled increases in calculated particle molecular weight with unmodified Ad: 5, 20, and 35 kd PEGylated vectors have predicted weights of 37, 55, 69, and 84 Md, respectively. Therefore, PEGylation appears to increase viral size by up to 40% and viral mass by up to twofold.

Table 1 - Characterization of adenovirus conjugated to different PEG molecules.

Full table (21K)

PEGylation decreases the in vitro activity of Ad5

PEGylation of Ad5 vectors using a 5 kd PEG has been shown to completely ablate in vitro activity. When unmodified and PEGylated viruses were bound to A549 cells for 20 hours and luciferase activity was measured, PEGylation reduced transduction by 99.8–99.99% (Table 1). These data demonstrate substantial reduction in viral activity in vitro by different-sized PEGs. Previous work has shown that this effect is primarily due to blocking interactions of the virus with CAR.^9,12

Effects of different-sized PEGs on in vivo liver transduction

To test the effect of the different-sized PEGs on in vivo liver transduction, we injected groups of four BALB/c mice IV with 5 10¹⁰ Ad virus particles (v.p.) of either mock PEGylated Ad or virus modified with 5, 20, and 35 kd PEG. The mice were then imaged at days 1, 3, and 7 for luciferase expression (Figure 1). One day after injection, stronger liver expression was observed in the Ad and Ad-5 kd PEGylated groups (P 0.05 by one-way ANOVA followed by Tamhane post-hoc test) (Figure 1a and b). This is consistent with previous observations of equal to slightly higher transduction by the 5 kd–modified virus.⁹ In contrast, both the 20 and 35 kd groups showed reduced luciferase expression when compared with the unmodified Ad group (P 0.05) (Figure 1b). These data agree with a recent study by Wortmann et al., which demonstrated that a 20 kd PEG severely decreases liver transduction in vivo. ¹³ When the kinetics of luciferase expression were compared over 7 days, highest expression was consistently observed in the 5 kd PEG and mock groups, with approximately tenfold lower expression in the 20 and 35 kd PEG groups (Figure 1c).

Figure 1.

In vivo transduction by unmodified and PEGylated Ad5. Groups of four mice were injected intravenously with 5 10¹⁰ virus particles (v.p.) of unmodified, 5, 20, and 35 kd PEGylated Ad-LucIREShrGFP. Luciferase expression was imaged 24, 72, and 168 hours later. (a) Pseudocolor image of luciferase activity at 24 hours postinjection overlaid on while light images of mice injected with different unmodified and PEGylated Ad-LucIREShrGFP. (b) Comparison of luciferase activity (in photons/second) for mice. The asterisks designate that the mean measurements of four mice each were significantly different than the unmodified Ad group. Significance was determined using one-way ANOVA, and because the variance was found not to be homogeneic by a Levene analysis (P 0.05), a Tamhane post-hoc test was performed and P 0.05 determined significance. (c) Kinetics of luciferase activity (in photons/second) for 24, 72, and 168 hours postinjection. Quantitative real-time PCR of Ad genomes/100 ng liver DNA using Ad5 Hexon primers. (d) Quantitative PCR of livers of mice removed 7 days postinjection expressed as number of v.p./100 ng total liver DNA. Using the same statistical analysis used for b, there is no significant difference between the means of the different groups (Ad-5 kd, Ad-20 kd, and Ad-35 kd) compared to the unmodified Ad group. There was a significant difference for the Ad-20 kd and Ad-35 kd compared to the Ad-5 kd (P = 0.042). Error bars in c represent SEM. Ad, adenoviral; PEG, polyethylene glycol.

Full figure and legend (73K)

After 7 days, the animals were killed and quantitative PCR for Ad5 virus genomes was performed on liver DNA (Figure 1d). PCR showed that there were approximately threefold higher amounts of Ad5 genomes in the liver after injection of Ad-5 kd versus mock PEGylated Ad (P = 0.07 by one-way ANOVA and Tamhane post-hoc test). Vector genomes for the 20 and 35 kd PEG groups were 28-fold lower than in the 5 kd group (P = 0.042). These genome data are generally consistent with the transduction data (Figure 1b). These data suggest that the use of different-sized PEGs influences delivery of virions to the liver and transduction of liver cells.

To explore why the larger PEG molecules reduce liver transduction, mice were injected IV with 5 10¹⁰ genomes of unmodified Ad and Ad modified with the 35 kd PEG. Blood was collected from the mice at 10 minutes and 24 hours after injection to determine whether PEGylation modified the distribution of the virus from the blood into tissues. When quantitated by real-time PCR, this experiment demonstrated that only 5% of the injected dose of Ad genomes was present in the blood at 10 minutes and that this amount fell tenfold by 24 hours (Figure 2). In contrast, Ad-35 kd PEG genome levels were more than tenfold higher at both time points (P < 0.0001 and P = 0.035 at 10 minutes and 24 hours, respectively by two-tailed t-test). Ad-35 kd PEG genome levels were 50% of injected dose at 10 minutes and 5% at 24 hours. These data suggest that the larger PEG molecules increase the amount of virus held in the circulatory system over at least 24 hours. When liver, kidney, and spleens were assayed for vector deposition at 24 hours, the Ad and Ad-35 kd PEG genome levels were not statistically different at this short time point (data not shown). In contrast, testing at 7 days indicated that modification of the virus with the larger PEGs reduced ultimate deposition of the virus in the liver (Figure 1d).

Figure 2.

Effect of 35 kd polyethylene glycol (PEG) on blood levels of Ad5. Groups of five mice were injected intravenously with 5 10¹⁰ virus particles of unmodified or 35 kd PEGylated Ad-LucIREShrGFP. Ten minutes and 24 hours later, blood was collected and DNA was prepared for quantitative PCR. Data are expressed as the number of adenoviral genomes/37.2 ng blood DNA. Error bars represent SEM. Ad versus Ad-35 kd PEG genome levels at 10 minutes and 24 hours were more than tenfold higher (P < 0.0001 and P = 0.035) by two-tailed t-test. The percentage of injected Ad dose discussed in the text is based on the number of genomes from the volume sampled extrapolated to 2 ml of total blood volume in a mouse. *P < 0.0001, **P < 0.05.

Full figure and legend (8K)

In vitro transduction by unmodified and PEGylated Ad is enhanced with blood factors

These data indicate that PEGylated Ad5 can still transduce the liver in vivo despite the fact that in vitro transduction is markedly reduced. This was not entirely surprising given recent data suggesting that liver transduction is not mediated by CAR interactions, but is instead mediated by binding of blood factors to the virus.^1,2 Given this, we tested whether PEGylated Ad can still interact with blood clotting factors. Unmodified and 5 and 35 kd PEGylated Ad5 were incubated with PC, FVII, FIX,¹⁴ and FX for 10 minutes as described in ref. 2. The viruses were then added to HepG2 cells for 3 hours and luciferase expression was measured 24 hours later (Figure 3). In the absence of added blood factors, this short-term binding of unmodified Ad mediated five- to tenfold higher transduction than the 5 and 35 kd vectors (Figure 3a and b). Addition of FIX and FX increased transduction for all of the vectors with FX generating the 7- to 70-fold increases in transduction. ANOVA and Levene followed by Bonferroni post-hoc test (if homogeneic) or Tamhane post-test (if not homogeneic) of the samples demonstrated that FVII, FIX, and FX mediated significant increases for unmodified Ad. In comparison, it also demonstrated that the addition of FIX and FX to the 5 and 35 kd PEGylated Ad significantly increased transduction (P 0.05).

Figure 3.

In vitro effect of blood factors on HepG2 transduction. Mock PEGylated Ad (Ad), 5 kd PEGylated Ad (Ad-5 kd), and 35 kd PEGylated Ad (Ad-35 kd) were incubated with different blood factors, no factor (0), protein C (PC), factor VII (FVII), factor IX (FIX), and factor X (FX). The virus blood factor complex was added to HepG2 cells and a luciferase activity assay was performed. (a) Effect of different factors on different Ad groups. Asterisks denote a significant difference between the different factor groups compared to the no factor (0) group by one-way ANOVA, and if the variance was found to be homogeneic by a Levene analysis (P 0.05), a Bonferroni post-hoc test was performed with a resulting P 0.05. If after a one-way ANOVA was performed the variance was found not to be homogeneic by a Levene analysis (P 0.05), a Tamhane post-hoc test was performed and P 0.05 determined significance. (b) Fold increase in luciferase activity relative to the no factor (0) group for Ad, Ad-5 kd, and Ad-35 kd for each individual factor. To remove the effects of polyethylene glycol (PEG) on vector activity, the luciferase activity produced by addition of each factor was normalized by dividing it by the luciferase activity mediated by the vector alone. This ratio is designated on the y-axis as "Normalized increase in transduction". Ad, adenoviral.

Full figure and legend (24K)

Effect of warfarin on in vivo liver transduction and transgene activity

The data above suggested that the PEGylated vectors can still interact with blood factors and that this binding may enable in vivo liver transduction. To test whether these blood factors influence liver transduction by the PEGylated vectors, mice were warfarinized as described in ref. 2 to inhibit these vitamin K–dependent blood clotting factors in vivo. Groups of four mice were warfarinized before IV injection with 5 10¹⁰ v.p. of mock PEGylated Ad, Ad-5 kd, Ad-20 kd, or Ad-35 kd (Figure 4). The mice were then imaged at days 1, 3, and 7. Imaging of all of the warfarinized groups showed little or no detectable luciferase expression (Figure 4a and data not shown). Comparison of luciferase expression at day 1 demonstrated that warfarinization reduced transduction significantly in the Ad, Ad-5, and Ad-35 kd PEG groups (P 0.05 t-test for each + and – warfarin group), but the Ad-20 kd PEG group did not reach significance. These data are in accordance with those of Parker et al.,² and support our hypothesis that the PEGylated Ad retains its in vivo activity due to maintenance of its interactions with blood factors.

Figure 4.

In vivo effect of warfarin depletion of blood factors for unmodified (Ad), 5 kd PEGylated (Ad-5 kd), 20 kd PEGylated (Ad-20 kd), and 35 kd PEGylated (Ad-35 kd) Ad-LucIREShrGFP. Groups of four mice were warfarinized and then injected intravenously with 5 10¹⁰ virus particles of unmodified, 5, 20, and 35 kd PEGylated Ad-LucIREShrGFP. Luciferase expression was imaged 24, 72, and 168 hours later. (a) Pseudocolor image of luciferase activity at 24 hours postinjection overlaid on while light images of mice injected with different unmodified and PEGylated Ad-LucIREShrGFP. (b) Comparison of luciferase activity (in photons/second) for mice. Asterisks designate a significant difference between the means of the two columns for each group (+ and warfarin) by two-tailed Student's t-test with a P 0.05. Ad, adenoviral; PEG, polyethylene glycol.

Full figure and legend (60K)

Effects of PEGylation on Ad transduction after IP injection

For some applications, IP injection is the simplest route of vector administration. However, this route may not efficiently transduce the liver, which is generally thought to be the optimal target for gene therapy of metabolic diseases. Previous work has shown that unmodified Ad transduces much of the squamous epithelial cells of the mesothelium in the peritoneal cavity in a CAR-dependent fashion.⁵ Given that PEGylation blocks CAR binding, but not blood factor interactions, we tested whether PEGylation could retarget Ad for better liver transduction after IP injection. To test this, adult ICR mice were injected IP with 1 10¹¹ v.p. of luciferase-expressing Ad-cmv-Luc-IRES-hrGFP (Ad) (Figure 5a). At 24 hours, luciferase expression was observed to spread diffusely throughout the entire peritoneum. In contrast, IP injection of 5 kd PEGylated Ad produced a markedly different pattern of expression. Luciferase expression was no longer observed throughout the peritoneum, but was focused almost exclusively in the liver. When the animals were imaged over 24 days after single IP injection, expression by the unmodified vector appeared to diminish within 3 days of injection. In contrast, expression by the PEGylated vector was detectable over the time course, but with marked reduction after day 8.

Figure 5.

In vivo liver targeting by 5 kd PEGylated adenovirus following intraperitoneal (IP) injection. Groups of four mice were injected intraperitoneally with 1 10¹¹ virus particles (v.p.) of unmodified (Ad) and 5 kd PEGylated (Ad-PEG) Ad-LucIREShrGFP. Luciferase expression was imaged at 72 hours postinjection. (a) Pseudocolor image of luciferase activity at 72 hours postinjection overlaid on while light images of mice injected with unmodified (Ad) and 5 kd PEGylated Ad-LucIREShrGFP (Ad-PEG). (b) Real-time PCR quantification of Ad genomes/500 ng liver DNA using Ad5 Hexon primers. Asterisk designates that the mean measurements of four mice each were significant by two-tailed Student's t-test with a P 0.05. Ad, adenoviral; IV, intravenous; PEG, polyethylene glycol.

Full figure and legend (85K)

To quantify viral genome delivery to the liver after IP injection, adult ICR mice were injected IP with 1 10¹¹ v.p. of mock PEGylated Ad and 5 kd PEGylated Ad, and livers were harvested 72 hours later for quantitative PCR (Figure 5b). Using this assay, it was found that IP injection of PEGylated Ad delivered 2.6-fold more viral genomes to the liver than unmodified Ad (Figure 5b; 5 10⁴ and 1.3 10⁵ genomes of Ad per 500 ng of liver DNA, respectively, P < 0.05). For comparison, IV injection of 2 10¹⁰ to 2 10¹¹ v.p. of Ad delivered 10⁴ to 2 10⁵ genomes, consistent with previous results.^9,11 These data suggest that PEGylation of Ad detargets CAR-dependent transduction of the peritoneum to enable better liver transduction after injection by the IP route.

Discussion

One goal of this work was to determine why PEGylation affects transduction by Ad differently in vitro and in vivo. Another goal was to determine how the use of different sizes of PEG influences the in vitro and in vivo transduction by Ad5. These data show that PEGylation with 5, 20, or 35 kd PEG drastically reduces in vitro transduction. In contrast, in vivo liver transduction after IV injection is highly dependent on PEG size.

We previously showed that 5 kd PEG can completely ablate in vitro activity and have no effect when IV injected in vivo. Data indicated that the in vitro effect was due to blocking of CAR interactions that might not be important in vivo in the liver. We speculated wildly that the normal transduction in vivo might be due to pressure effects driving the virus onto cells from blood.⁹ Since then, a new paradigm has been suggested for Ad biology where it appears that liver transduction is mediated not by ligands evolved by the virus, but instead by interactions with endogenous blood factors that retarget the virus to the liver.^1,2 With this new perspective, we tested here whether transduction by unmodified and PEGylated viruses is increased by blood factor interactions. Consistent with previous work,² we found that the addition of FVII, FIX, and FX to unmodified Ad5 significantly increased transduction in vitro, but PC did not. In contrast, for the PEGylated viruses, addition of FIX and FX had a significant increase with FX having the strongest effect. When we globally decreased the activity of these blood factors in vivo with warfarin, both unmodified and PEGylated Ad transgene expression was decreased. These in vitro and in vivo data suggest that PEGylated Ad remains active for in vivo transduction by maintaining blood factor interactions.

We also looked at the effect of using larger PEG molecules on liver transduction and transgene expression in vivo. Using in vivo luciferase imaging, we demonstrate that the larger PEGs significantly reduce transgene expression in the liver. Our findings agree with recent data in ref. 13 that show that a 20 kd PEG significantly reduces liver transgene expression. We hypothesize that this decrease in in vivo activity may be due to the significantly larger diameter and molecular weight of Ad5 modified with the 20 and 35 kd PEGylated versus unmodified and 5 kd PEG-modified Ad. Ad, Ad-5 kd, Ad-20 kd, and Ad-35 kd had estimated diameters of 112, 132, 145, and 158 nm, respectively, determined by dynamic light scattering in aqueous buffer. The size of liver fenestrations allowing access from the blood to the liver parenchyma is 100–110 nm in mice.¹⁵ While both the unmodified and 5 kd PEGs are larger than this size, dynamic scattering is an artificial measurement made in buffer rather than blood. Therefore, the virus particles may be smaller when injected into the bloodstream. In addition, PEG is highly flexible and may allow virions to squeeze through the fenestrations by bending the PEGs.

We hypothesize that the size of the PEGs could be mediating this effect. However, it is formally possible that the small variations in linkers between the different-sized PEGs and the virus may have unexpected effects. The observation that the large 35 kd PEG increased the amount of virus in the blood from 5 to 50% at 10 minutes after injection is consistent with this fenestra-PEG size hypothesis. This observation is consistent with previous work coating Ad with poly-(N-(2-hydroxypropyl)methacrylamide).¹⁶ In this case, poly-(N-(2-hydroxypropyl)methacrylamide) polymer shielding increased Ad titers in the blood approximately tenfold at 10 minutes. While it is possible that the higher genome levels observed in the blood with polymer-modified viruses could be due to increased stability of the virus and its genome due to these modifications, this would not explain why liver transduction is lower rather than higher and why viral genome levels in the liver are lower for viruses modified with the large PEGs.

An additional goal of this study was to determine whether PEGylation of Ad vectors could assist in retargeting the vector to the liver after IP injection. This speculation was based on work showing that IP injection of Ad vectors ablated for CAR binding detargeted the virus from the peritoneum and increased liver transduction.⁵ This finding, coupled to our work and other's work showing that PEGylation blocks CAR-mediated transduction in vitro,^9,12 suggested that PEGylation might also retarget IP-injected virus. This prediction appears to be accurate, because we observed that PEGylated Ad mediated substantially reduced expression in the peritoneum, and substantially increased expression in the liver. Given that PEGylated viruses can still bind FIX and FX, we speculate that the PEGylated viruses that are liberated from the peritoneum are leached into the bloodstream from the peritoneum via the thoracic duct, and that these factor-bound virions can then efficiently transduce the liver. Therefore, PEG can be used to detarget Ad, at least from CAR binding.

In summary, PEGylation can change the tropism of adenovirus by blocking CAR interactions and perhaps by changing virus diameter. This in addition to PEG's effects on reducing thrombocytopenia, innate immune responses, and anti-vector immune responses make this simple chemical modification a viable approach to improve and modify the pharmacology of Ad and other viral vectors for human gene therapy.

Materials and Methods

Adenoviruses. First-generation replication-defective (E1 deleted) Ad5 vectors carrying luciferase and renilla green fluorescent protein, Ad-cmv-Luc-IRES-hrGFP (Ad), were produced by the AdEasy System (Qbiogene, Montreal, Québec, Canada) in 293A cells as described in ref. 9.

Animals. All animal experiments were carried out according to the provisions of the Animal Welfare Act, Public Health Service Animal Welfare Policy, the principles of the National Institutes of Health Guide for the Care and Use of Laboratory Animals, and the policies and procedures of Mayo Clinic. Mice were purchased from Harlan Sprague Dawley (Indianapolis, IN).

Generation of PEGylated virus (Ad5-PEG). Our previous work used succinimidyl propionate (SPA) 5 kd PEG from Nektar Therapeutics for PEG shielding of adenovirus.⁹ As Nektar unfortunately ceased selling their reagents, we used another succinimide-activated 5 kd PEG, Sunbright ME-050-HS, from NOF (Tokyo, Japan) which yielded essentially the same results as those of SPA-5 kd PEG.¹¹ For this study, larger 20 and 35 kd succinimide-activated PEGs were purchased from JenKem Technology, Beijing, China (m-SCM-20K and m-SCM-35K). The PEG and N-hydroxysuccinimide ends of these reagents are identical. They differ only in the linker molecule between these ends (Supplementary Figure S1). SPA-PEG has a -COO-(CH₂)₂- linker, ME-050-HS has a -COO-(CH₂)₅- linker, and the mSCM reagents have -COO-(CH₂)- linkers. These linkages are predicted to have similar stabilities. In the context of the additional 5,000, 20,000, and 35,000 d PEG appendages, the 16, 32, and 80 d linkers on these PEGs do not contribute substantially to the size of the modification to the virus. All PEGylation was performed under the same conditions with the same virus, which was stored and reacted in 0.5 mol/l sucrose potassium phosphate-buffered saline (10 mmol/l K₂PO₄, 150 mmol/l NaCl, 1 mmol/l MgCl₂, 5% wt/vol sucrose, pH 8). The mock PEGylated Ad (Ad) group was treated exactly as the PEGylated groups, but no PEG was added. The PEGylation reactions were carried out at room temperature for 1 hour, and the excess of nonreactive PEG molecules were then removed using gel filtration chromatography (Sephadex G100; Amersham Biosciences, Piscataway, NJ). The concentration of virus was determined using real-time PCR of hexon in Ad5 as described in ref. 9, to ensure that PEG did not skew absorbance readings.

Quantitation of free amines. Free amines were determined using the CBQCA Protein Quantitation Kit (Molecular Probes, Eugene, OR) according to the manufacturer's instructions. In brief, modified viruses were diluted in 135 l of reaction buffer (Dulbecco's phosphate-buffered saline, pH 9.3) to a final concentration of 1 10¹² v.p./ml in a 96-well black plate (3603; Corning, Corning, NY). Five microliters of 20 mmol/l KCN and 10 l of 5 mmol/l CBQCA were added, and the plate was covered with aluminum foil and incubated at room temperature with shaking for 1 hour. Fluorescence emission was detected at 550 nm with excitation at 465 nm using a Beckman Coulter DTX 880 Multimode Detector (Beckman Coulter, Fullerton, CA). The percentage of free amines was determined relative to a standard curve of unmodified virus and the inverse was expressed as the percentage of PEG-conjugated free amines.

In vitro transduction. In vitro activity was assessed in both A549 and 293 cells. Cells were grown in 96-well black plates (3603; Corning). Virus was added to the plate at 1 10⁴ v.p./cell and incubated for 20 hours at 37 °C. After incubating, 25 l of 5 reporter lysis buffer (Promega, Madison, WI) was added and the plate was subjected to one freeze-thaw cycle at -80 °C. The plate was thawed at room temperature and 50 l of luciferase assay reagent was added (Promega, Madison, WI). Luminescence was determined using the Beckman Coulter DTX 880 Multimode Detector (Beckman Coulter). The percent transduction was determined as compared to unmodified virus.

Particle sizing. Particle sizing was performed using the 90Plus/BI-MAS Multi Angle Particle Sizer (Brookhaven Instruments, Holtsville, NY). Viruses were diluted in 10 mmol/l KNO₃ to a final concentration of 1 10¹¹ v.p./ml. Viruses were measured using three 3-minute runs. Particle sizes are shown as the mean of three runs and SE. Polydispersity was calculated and is proportional to the variance of the intensity weighted diffusion coefficient distribution. Molecular weights were calculated using the Mark–Houwink–Sakadura equation.

Effect of blood factors on Ad transduction of liver cells in vitro. Mock PEGylated Ad, 5 kd PEG-Ad, and 35 kd PEG-Ad were incubated with PC, FVII, FIX, or FX (Haematologic Technologies, Essex Junction, VT) for 10 minutes at 37 °C in the Dulbecco's modified Eagle's medium without serum and were transferred onto HepG2 cell monolayers (grown in a 96-well plate) at multiplicity of infection of 3,000 v.p./cell. Blood factors were used at physiological concentration (10 g/ml). After 3 hours of adsorption, the medium was replaced with fresh medium containing 10% fetal bovine serum. At 24 hours postinfection, D-luciferin (Molecular Imaging Products, Bend, OR) was added to the cells at 1 g/ml and luminescence was measured using DTX 880 Multimode Detector (Beckman Coulter).

In vivo effect of different PEGs on luciferase expression from IV administration. The different modified Ad samples mock PEGylated Ad; Ad-5 kd, Ad-20 kd, and Ad-35 kd were diluted to a concentration of 5 10¹² v.p./ml. in 0.5 mol/l sucrose potassium phosphate-buffered saline and injected IV on day 0 at a multiplicity of infection of 5 10¹⁰ v.p. and a volume of 100 l/mouse.

In vivo warfarin-mediated depletion of blood factors. Mice were injected subcutaneously with 133 g of warfarin (Sigma-Aldrich, St. Louis, MO) in 100 l of peanut oil on day 3 preinjection and day 1 preinjection. These warfarinized mice were injected with the different modified Ad samples as described earlier.

In vivo effect of different PEGs on luciferase expression from IP administration. Mock PEGylated Ad, and Ad-5 kd were injected IP into groups of four female outbred ICR mice. Approximately 110¹¹ v.p. were injected in a total volume of 100 l.

Luciferase imaging of IV injected mice. Molecular light imaging of luciferase in vivo was accomplished using a Lumazone imaging system (Roper Scientific Photometrics, Pleasanton, CA). Female BALB/c mice 8–12 weeks old were each injected IV with Ad expressing firefly luciferase at a dose of 5 10¹¹ v.p./mouse. This concentration was determined using real-time PCR as described earlier. At 1, 3, and 7 days postinjection, mice were anesthetized with isoflurane, injected IP with D-luciferin at a concentration of 20 mg/ml in phosphate-buffered saline in a volume of 200 l, and the mice were immediately placed into the Lumazone Imager (Photometrics, Pleasanton, CA) and images were captured. All images were taken with a 1-minute exposure and 1 1 binning using no filters and no photo-multiplication. Data analysis was performed on each image using background-subtracted mean intensities detected by the Lumazone Imaging Software at each time point and graphed using Prism Graphing Software (GraphPad Software, San Diego, CA).

Luciferase imaging of IP-injected mice. In vivo imaging was performed as described for the IV injected mice. Images were acquired on days 1, 2, 3, 8, 12, and 24 days postinjection.

Ad5 genome quantification in liver of IV injected mice. Quantitative PCR was performed as described in ref. 11 with the following modifications. Mice were killed at day 7 following imaging analysis. The major lobes of the livers were immediately removed from each mouse and flash-frozen in liquid nitrogen. DNA purification was performed using Qiagen DNeasy Blood and Tissue Kit (Qiagen, Valencia, CA). DNA concentration was quantified by Abs₂₆₀. DNA was diluted to 20 ng/l to normalize the input quantity of total DNA per reaction to 100 ng/reaction. Quantitative PCR was performed as follows. Primers that amplify a 150-base pair region of Hexon were used. A standard curve was set up as instructed in ABI document 4371090 (Applied Biosystems, Foster City, CA); using plasmid DNA containing the Ad5 hexon gene, tenfold dilutions were made from 3 10² to 3 10⁹ copies of hexon per 5 l. Two master mixes were setup, one for the standards and one for the unknown liver samples. The standard master mix was made from 5 l of DNA purified from an uninjected mouse liver at a concentration of 20 ng/ l, 5 l of H₂o, 5 l of 3 mol/l hexon 5' primer (ACAAGCGAGTGGTGGGACTC), 5 l of 3 mol/l hexon 3' primer (GCATTGCGGTGGTGGTTAA), and 25 l of Qiagen QuantiTect SYBR Green PCR Master Mix (Qiagen, Valencia, CA) per reaction. The uninjected mouse DNA was used to reduce the effect of nonspecific amplification in the unknown samples. The unknown sample master mix was made from 10 l of H₂0, 5 l of 3 mol/l hexon 5' primer, 5 l of 3 mol/l hexon 3' primer, and 25 l of Qiagen QuantiTect SYBR Green PCR Master Mix (Qiagen, Valencia, CA) per reaction. A 96-well plate was set up using 45 l of the standard or unknown sample master mix and 5 l of the standard or unknown sample, respectively. Each sample was run in triplicate. Real-time PCR was performed and analyzed with an ABI 7900HT (Applied Biosystems, Foster City, CA) in the absolute quantification mode. A dissociation curve was created to test the specificity of the PCR. The genome quantity was determined using ABI analysis software.

Ad5 genome quantification in liver of IP-injected mice. Three mice per group of outbred ICR mice were injected with 1 10¹¹ v.p. of mock PEGylated Ad and Ad-5 kd. At day 3 postinjection, the mice were killed, and as described earlier, quantitative PCR was performed on their livers to quantify the number of Ad genomes per 500 ng of total liver DNA.

References

REFERENCES

1. Shayakhmetov, DM, Gaggar, A, Ni, S, Li, ZY and Lieber, A (2005). Adenovirus binding to blood factors results in liver cell infection and hepatotoxicity. J Virol 79: 7478–7491. | Article | PubMed | ISI | ChemPort |
2. Parker, AL, Waddington, SN, Nicol, CG, Shayakhmetov, DM, Buckley, SM, Denby, L et al. (2006). Multiple vitamin K-dependent coagulation zymogens promote adenovirus-mediated gene delivery to hepatocytes. Blood 108: 2554–2561. | Article | PubMed | ISI | ChemPort |
3. Waddington, SN, McVey, JH, Bhella, D, Parker, AL, Barker, K, Atoda, H et al. (2008). Adenovirus serotype 5 hexon mediates liver gene transfer. Cell 132: 397–409. | Article | PubMed | ChemPort |
4. Kalyuzhniy, O, Di Paolo, NC, Silvestry, M, Hofherr, SE, Barry, MA, Stewart, PL et al. (2008). Adenovirus serotype 5 hexon is critical for virus infection of hepatocytes in vivo. Proc Natl Acad Sci USA 105: 5483–5488. | Article | PubMed |
5. Akiyama, M, Thorne, S, Kirn, D, Roelvink, PW, Einfeld, DA, King, CR et al. (2004). Ablating CAR and integrin binding in adenovirus vectors reduces nontarget organ transduction and permits sustained bloodstream persistence following intraperitoneal administration. Mol Ther 9: 218–230. | Article | PubMed | ISI | ChemPort |
6. O'Riordan, CR, Lachapelle, A, Delgado, C, Parkes, V, Wadsworth, SC, Smith, AE et al. (1999). PEGylation of adenovirus with retention of infectivity and protection from neutralizing antibody in vitro and in vivo. Hum Gene Ther 10: 1349–1358. | Article | PubMed | ISI | ChemPort |
7. Croyle, MA, Chirmule, N, Zhang, Y and Wilson, JM (2001). "Stealth" adenoviruses blunt cell-mediated and humoral immune responses against the virus and allow for significant gene expression upon readministration in the lung. J Virol 75: 4792–4801. | Article | PubMed | ISI | ChemPort |
8. Croyle, MA, Chirmule, N, Zhang, Y and Wilson, JM (2002). PEGylation of E1-deleted adenovirus vectors allows significant gene expression on readministration to liver. Hum Gene Ther 13: 1887–1900. | Article | PubMed | ISI | ChemPort |
9. Mok, H, Palmer, DJ, ng, P and Barry, MA (2005). Evaluation of polyethylene glycol modification of first-generation and helper-dependent adenoviral vectors to reduce innate immune responses. Mol Ther 11: 66–79. | Article | PubMed | ISI | ChemPort |
10. Croyle, MA, Le, HT, Linse, KD, Cerullo, V, Toietta, G, Beaudet, A et al. (2005). PEGylated helper-dependent adenoviral vectors: highly efficient vectors with an enhanced safety profile. Gene Ther 12: 579–587. | Article | PubMed | ISI | ChemPort |
11. Hofherr, SE, Mok, S, Gushiken, FC, Lopez, JA and Barry, MA (2007). Polyethylene glycol modification of adenovirus reduces platelet activation, endothelial cell activation, and thrombocytopenia. Hum Gene Ther 18: 837–848. | Article | PubMed | ChemPort |
12. Ogawara, K, Rots, MG, Kok, RJ, Moorlag, HE, Van Loenen, AM, Meijer, DK et al. (2004). A novel strategy to modify adenovirus tropism and enhance transgene delivery to activated vascular endothelial cells in vitro and in vivo. Hum Gene Ther 15: 433–443. | Article | PubMed | ISI | ChemPort |
13. Wortmann, A, Vöhringer, S, Engler, T, Corjon, S, Schirmbeck, R, Reimann, J et al. (2007). Fully detargeted polyethylene glycol-coated adenovirus vectors are potent genetic vaccines and escape from pre-existing anti-adenovirus antibodies. Mol Ther 16: 154–162. | Article | PubMed | ChemPort |
14. Foreman, PK, Wainwright, MJ, Alicke, B, Kovesdi, I, Wickham, TJ, Smith, JG et al. (1998). Adenovirus-mediated transduction of intestinal cells in vivo. Hum Gene Ther 9: 1313–1321. | Article | PubMed | ChemPort |
15. Braet, F and Wisse, E (2002). Structural and functional aspects of liver sinusoidal endothelial cell fenestrae: a review. Comp Hepatol 1: 1. | Article | PubMed |
16. Green, NK, Herbert, CW, Hale, SJ, Hale, AB, Mautner, V, Harkins, R et al. (2004). Extended plasma circulation time and decreased toxicity of polymer-coated adenovirus. Gene Ther 11: 1256–1263. | Article | PubMed | ISI | ChemPort |

Acknowledgments

We thank Mary Barry and Jenica Schwegel for their helpful technical assistance. This work was supported by grants to M.A.B. from the Muscular Dystrophy Association, the Propionic Acidemia Foundation, and the Organic Acidemia Association. The authors have no competing financial interests.