Adenoviral Vectors Coated with PAMAM Dendrimer Conjugates Allow

Jan 2, 2013 - Oncolytic virotherapy for urological cancers. Zahid Delwar , Kaixin Zhang , Paul S. Rennie , William Jia. Nature Reviews Urology 2016 13...
2 downloads 0 Views 509KB Size
Download PDF
Article pubs.acs.org/molecularpharmaceutics

Adenoviral Vectors Coated with PAMAM Dendrimer Conjugates Allow CAR Independent Virus Uptake and Targeting to the EGF Receptor Alexandra Vetter,† Kulpreet S. Virdi,‡ Sigrid Espenlaub,§ Wolfgang Rödl,† Ernst Wagner,†,∥ Per S. Holm,⊥ Christina Scheu,‡,∥ Florian Kreppel,§ Christine Spitzweg,*,# and Manfred Ogris*,†,∥ †

Center for System Based Drug Research, Department of Pharmacy, LMU Munich, Germany Department of Chemistry, Physical Chemistry, LMU Munich, Germany § Department of Gene Therapy, University of Ulm, Germany ⊥ Institute of Experimental Oncology and Therapy Research, TU Munich, Germany ∥ Center for NanoScience (CeNS), LMU Munich, Germany # Department of Internal Medicine II, LMU Munich, Germany ‡

S Supporting Information *

ABSTRACT: Adenovirus type 5 (Ad) is an efficient gene vector with high gene transduction potential, but its efficiency depends on its native cell receptors coxsackieand adenovirus receptor (CAR) for cell attachment and αvβ3/5 integrins for internalization. To enable transduction of CAR negative cancer cell lines, we have coated the negatively charged Ad by noncovalent charge interaction with cationic PAMAM (polyamidoamine) dendrimers. The specificity for tumor cell infection was increased by targeting the coated Ad to the epidermal growth factor receptor using the peptide ligand GE11, which was coupled to the PAMAM dendrimer via a 2 kDa PEG spacer. Particles were examined by measuring surface charge and size, the degree of coating was determined by transmission electron microscopy. The net positive charge of PAMAM coated Ad enhanced cellular binding and uptake leading to increased transduction efficiency, especially in low to medium CAR expressing cancer cell lines using enhanced green fluorescent protein or luciferase as transgene. While PAMAM coated Ad allowed for efficient internalization, coating with linear polyethylenimine induced excessive particle aggregation, elevated cellular toxicity and lowered transduction efficiency. PAMAM coating of Ad enabled successful transduction of cells in vitro even in the presence of neutralizing antibodies. Taken together, this study clearly proves noncovalent, charge-based coating of Ad vectors with ligandequipped dendrimers as a viable strategy for efficient transduction of cells otherwise refractory to Ad infection. KEYWORDS: adenovirus, PAMAM dendrimer, EGFR targeting, GE11 peptide, CAR motive in the penton base protein with the cell surface αvβ3 and αvβ5 integrins, leading to internalization.7 Another interaction, which is especially important, when considering application of the adenovirus via the bloodstream, is the interaction between coagulation factor X and the hexon protein, which mediates hepatocyte infection.8 In our study we focused on CAR, which is stated to be the key player for efficient cell infections and is often downregulated on tumor cells, correlating with the aggressiveness of the tumor and making infection a major bottleneck to overcome for effective gene therapy.9 By increasing the multiplicity of infection (MOI) and the contact time between the Ad and the cell surface, the problem of insufficient transduction efficiency can be addressed, but at the same time this is elevating the risk of toxicity and inflammatory

1. INTRODUCTION Viral gene transfer vectors like the adenovirus type 5 (Ad) have the advantage of a high transduction efficiency in many cell types. They can infect dividing as well as nondividing and differentiated cells without integrating their DNA into the host genome therefore exhibiting a limited mutagenesis rate.1 Additionally, the transgenic protein can be expressed at high levels of up to 35% of the total cellular protein. Still, the in vivo application of adenoviral vectors is facing significant hurdles, for example due to neutralization of type 5 adenovirus by preexisting antibodies, leading to a strong reduction of target cell infection.2−4 Yet another problem to solve is that the transduction efficiency is limited by the level of viral receptors on target cells. Ad is infecting its target cells through at least four different interactions: the fiber knob protein of the Ad capsid is binding to the coxsackie- and adenovirus receptor (CAR) for cell attachment,5 as is the KKTK (Lys-Lys-Thr-Lys) motive of the fiber shaft to heparan sulfate proteoglycans (HSPG),6 followed by interaction of the RGD (Arg-Gly-Asp) © XXXX American Chemical Society

Received: July 9, 2012 Revised: November 7, 2012 Accepted: January 2, 2013

A

dx.doi.org/10.1021/mp300366f | Mol. Pharmaceutics XXXX, XXX, XXX−XXX

Molecular Pharmaceutics

Article

EGF (mEGF) was purchased from Peprotech (Hamburg, Germany). Amine terminated PAMAM dendrimers (generations G2, G3 and G5, DAB core, molecular weight 3,284 Da, 6,937 Da and 28,854 Da, respectively) were purchased from Dendritic Nanotechnologies (Mount Pleasant, MI 48858, USA), Branched PEI (BPEI, 25 kDa) was purchased from Sigma-Aldrich and used after gel filtration (Sephadex G25 superfine, HR 10/30) as a 1 mg/mL stock solution in HBS. HEPES (N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid) from Biomol GmbH (Hamburg, Germany), H-L-cysteine from IRIS Biotech GmbH (Marktredwitz, Germany) and MacroPrep High S from BioRad GmbH (Munich, Germany). Dialysis was performed with Spectra/Por membranes (molecular mass cutoff 1, 3.5, or 10 kDa) from Spectrum Laboratories Inc. (Breda, Netherlands). 2.2. Cell Culture. Human cancer cell lines were maintained as follows. HuH7 hepatoma (JCRB0403; Japanese Cancer Research Resources Bank, Tokyo, Japan) was cultivated in Ham’s F12/DMEM (1:1), SKOV-3 ovarian carcinoma (HTB77; American Type Culture Collection [ATCC], Manassas, VA), HepG2 hepatoma (HB-8065, ATCC) and DU145 prostate carcinoma cells (ACC 261; German Collection of Microorganisms and Cell Culture [DSMZ], Braunschweig, Germany) in RPMI 1640, A549 lung carcinoma (CCL-185, ATCC) and U87MG glioblastoma (HTB-14, ATCC) in DMEM. All cell culture media were supplemented with 10% fetal bovine serum, 100 units/mL penicillin, 100 μg/mL streptomycin and 1% L-alanine-L-glutamine (200 mM). Medium for U87MG cells was additionally supplemented with 1% NEAA (nonessential amino acids), and cell culture flasks were coated with 0.1 mg/mL collagen A before use. All cells were passaged twice a week and cultured at 37 °C in 5% CO2. 2.3. Adenoviral Vectors. Ad-Luc (1 × 1010 VP/μL (viral particles per microliter)), 5 × 108 PFU (plaque-forming units) was purchased from Vector Biolabs (Philadelphia, PA 19104, USA) and amplified in HEK 293 cells. The virus was purified by two consecutive cesium chloride gradient centrifugations and desalted by size-exclusion chromatography (disposable PD10 desalting columns, GE Healthcare, Freiburg, Germany). Viral titer was determined by AdEasy Viral Titer Kit (Stratagene, Waldbronn, Germany) and viral particle number by absorption measurement at 260 nm. Ad-Alexa Fluor488 (Ad-Alexa488, 1 × 108 VP/μL, 5 × 106 PFU), enhanced green fluorescent protein encoding Ad (Ad-EGFP, 5 × 108 VP/μL, 1 × 108 PFU) and enhanced green fluorescent and luciferase fusion protein encoding Ad (Ad-EGFPLuc 1 × 109 VP/μL, 6 × 10 7 PFU) are Ad5-based E1-deleted first-generation vectors, which were amplified in the E1-transcomplementing N52E6 cell line.25 The Ad-EGFP and Ad-EGFPLuc carry an hCMVpromoter controlled EGFP expression cassette. Vectors were purified by one CsCl density step gradient followed by one subsequent continuous CsCl density gradient and were desalted by PD-10 size exclusion columns (GE Healthcare). The titers for infectious and physical particles were determined by a DNA-based slot blot assay26 and, in the case of the titer for physical particle, confirmed by measuring the optical density at 260 nm. For random labeling, 5 × 1011 AdEmpty (no transgene) vector particles were incubated with a 20-fold excess of amine reactive Alexa Fluor488-5TFP (Invitrogen, Darmstadt, Germany) over the 18,000 amino groups present at the capsid surface. After an overnight incubation at ambient temperature, the reaction mixture was filled up to a final

responses in patients. As an alternative to viral vectors, synthetic gene delivery systems have evolved, like cationic lipids or polymers,10 which are able to condensate DNA forming a positively charged complex that can enter the cell and escape the endosome via “a proton sponge effect”.11 Such vectors have the advantage that no preformed antibodies exist, which could prevent a successful infection. Furthermore, modification with hydrophilic polymers, like polyethyleneglycol (PEG), enables circulation of drugs and gene vectors in the bloodstream, protects from the immune system and allows accumulation in tumor tissue.12 Synthetic delivery systems can also be designed in a modular way, which allows the incorporation of cell targeting and internalizing ligands considerably improving their efficiency and specificity.13 Still, there are major hurdles for nonviral gene delivery vectors to overcome, like poor nuclear import and the absence of a transcriptional enhancer, making them less effective compared to their viral counterparts.14 Hence, a combination of both would be beneficial to achieve a high transgene expression level also in cells not accessible by Ad and at the same time avoid inactivation by neutralizing antibodies. Due to its net negative surface charge, Ad capsids can be coated with (poly)cationic molecules by noncovalent charge interaction. Fasbender et al. were the first to demonstrate the increase in binding and subsequent transgene expression of Ad noncovalently coated with cationic lipids on differentiated airway epithelia in vitro as well as nasal epithelium of mice in vivo.2 Other researchers confirmed these findings by using cationic lipids, (proteo)liposomes or cationic polymers like poly L -lysine or polyethylenimine.2,15−20 As certain polycations, like linear polyethylenimine (LPEI), the gold standard of nonviral gene delivery systems, suffer from undeniable toxicity due to crosslinking of erythrocytes in the lung,21 there is a search for welltolerated and efficient Ad coating reagents. Although less efficient than LPEI for nonviral gene delivery, PAMAM dendrimers are well-known drug carriers and exhibit a considerably lower toxicity when applied systemically.22 PAMAM dendrimers have the advantage over LPEI to have a defined structure and a defined molecular weight due to the synthesis procedure. In our study we designed a hybrid system by coating of Ad with cationic PAMAM dendrimers that maximizes viral infection ability and gene delivery efficiency but at the same time reduces toxicity compared to standard coating agents used so far. Having established a novel system with PAMAM dendrimers as coating agents, we could show a clear transduction benefit for low to medium CAR level cancer cell lines. By the conjugation of the peptide ligand GE11 (CYHWYGYTPQNVI),23 targeting the epidermal growth factor receptor (EGFR), onto PAMAM dendrimers, it was possible to further increase specificity for EGFR overexpressing tumor cells.

2. EXPERIMENTAL METHODS 2.1. Materials. DMSO (dimethyl sulfoxide purissimum), EtOH (ethanol absolutum), DTT (DL-dithiothreitol) and TNBS (2,4,6-trinitrobenzenesulfonic acid solution) were obtained from Sigma-Aldrich GmbH (Munich, Germany). Water was used as purified, deionized water. NHS-PEG-OPSS (ω-2-pyridyldithio polyethylene glycol α-succinimidylester, 2 kDa) was synthesized by Rapp Polymere GmbH (Tübingen, Germany), peptide GE11 (CYHWYGYTPQNVI-OH, TFA salt, >95% purity) by Biosyntan GmbH (Berlin, Germany), and LPEI (22 kDa) as recently described.24 Recombinant murine B

dx.doi.org/10.1021/mp300366f | Mol. Pharmaceutics XXXX, XXX, XXX−XXX

Molecular Pharmaceutics

Article

volume of 2,500 μL with PBS. The labeled vector particles were then purified from excess dye by PD10 column purification (GE Healthcare). 2.4. Conjugate Syntheses. Synthesis of PAMAM G2PEG-OPSS was carried out in principle as described recently for LPEI-PEG-OPSS27 with some modifications. In brief, 1.4 μmol of PAMAM G2 (MW = 3,284 Da) in EtOH was reacted with 4.2 μmol of NHS-PEG-OPSS (2 kDa) dissolved in DMSO for 3 h under agitation at 37 °C. Thereafter, 2 M HEPES pH 7.4, 3 M NaCl and water were added to give a final concentration of 20 mM HEPES and 0.6 M NaCl and the pH was adjusted to 7.4 using hydrochloric acid. The reaction mixture was loaded on a cation-exchange column (Macro-Prep High S; 10/10; BioRad, Munich, Germany) and fractioned with a salt gradient from 0.6 to 3 M NaCl in 20 mM HEPES, pH 7.4. The product eluted between 2 and 2.6 M NaCl and was dialyzed overnight at 4 °C against HBS (20 mM HEPES pH 7.4, 150 mM NaCl). Dialysis was performed with Spectra/Por membrane (molecular mass cutoff 1 kDa). The PAMAM G2 content of the conjugate was determined by TNBS assay,28 the amount of dithiopyridine after reduction of an aliquot with dithiothreitol (DTT) followed by absorption measurement of released pyridine-2-thione (ε343 = 8,080 M−1 × mol−1; ref 29). It resulted in a molar ratio of PAMAM G2/PDP of 1/0.6 M/M. For PAMAM G2-PEG-GE11 synthesis, 1.98 μmol of GE11 in 75 μL of 30% acetonitrile, 70% H2O and 0.1% TFA (trifluoroacetic acid) and 0.79 μmol of PAMAM G2-PEGOPSS in 3.6 mL of HBS were mixed and incubated at ambient temperature. Reaction was finished after all PDP was released. The release was controlled during the reaction at 343 nm and recalculated to the used PDP. Purification and analysis was carried out as described above with the difference of using a salt gradient from 0.6 to 3 M NaCl in 20 mM HEPES including 10% acetonitrile, pH 7.4. The amount of GE11 was calculated via the extinction coefficient at 280 nm (ε = 9,970 M−1 × mol−1). The molar ratio of PAMAM G2 to GE11 was 1 to 0.75. PAMAM G2-PEG-Cys was synthesized by mixing one equivalent of PAMAM G2-PEG-OPSS with four equivalents of cysteine at ambient temperature. Purification was carried out on a gel-filtration column (Sephadex G-25; HR10/30 column; 20 mM HEPES, pH 7.4). Cyanine 5 (Cy5) labeling of PAMAM G5 was performed with amine-reactive FluoroLink Cy5 monofunctional dye (GE Healthcare, Freiburg, Germany) in principle as described.30 In brief, 2 mg of PAMAM G5 in 1 mL of HBS was mixed with 69 nmol of Cy5 dissolved in 40 μL of water free DMSO. After 2 h at ambient temperature the sample was purified by gel filtration to remove unreacted dye. The molar ratio of PAMAM G5/Cy5 was 1/0.3. 2.5. Coating of Ad with PAMAM Dendrimer or LPEI and BPEI Polymer. The complexes of Ad and PAMAM, LPEI or BPEI were formed by diluting indicated amounts of Ad in Opti-MEM (Invitrogen) and polycation in HEPES buffered glucose (HBG, pH 7.4). Viral particles were added to the polycation solution, immediately mixed by gentle aspiration with the pipet tip and allowed to incubate for 30 min at ambient temperature (20−25 °C) before further use. 2.6. Zeta Potential. Ad-Luc (5 × 1010 VP/mL in PBS/ HBG, pH 7.4) and Ad/PAMAM G5 complexes (formed as described above, but with Ad diluted in PBS instead of OptiMEM to reduce background signal) were analyzed on a Zetasizer Nano ZS (Malvern Instruments, Malvern Hills, U.K.) at 25 °C in triplicate.

2.7. Transmission Electron Microscopy (TEM). The experiment was performed with a method similar to Chen et al.31 with some modifications. A drop of 20 μL containing either 1 × 1010 VP of Ad-Luc in water or Ad complexed with 950 molecules of PAMAM G5 per Ad in PBS, which equals 4 molecules per each hexon protein, was put on a sheet of parafilm. A plasma cleaned carbon coated 200 mesh copper grid (Plano GmbH, Wetzlar, Germany) was placed on top of the drop (filmed side down) for 5 min and excessive liquid drained off using filter paper. Subsequently the grid was incubated with 20 μL of a 1% phosphotungstic acid solution (PTA) (Science Services, Munich, Germany) for 3 min and drained as before. All solvents were filtered with a 0.02 μm spin column filter (Whatman, Dassel, Germany) before use. Air-dried grids were analyzed within 48 h by a JEOL 2011 TEM (JEOL Ltd.Tokyo, Japan) fitted with a lanthanum hexaboride thermal emitter and operating at an accelerating voltage of 200 kV and at low dose. 2.8. Estimation of the Receptor Level. 600,000 cells in 100 μL of FACS buffer (10% FCS in PBS) were incubated with either mouse anti-CAR (1:200 dilution, Millipore, Schwalbach, Germany, clone RmcB), mouse anti-EGFR antibody (1:200 dilution, Dako, Glostrup, Denmark, clone H11) or mouse IgG control antibody (1:100 dilution, Dako) for 1.5 h on ice, washed twice with FACS buffer and thereafter stained with Alexa488-labeled goat anti-mouse antibody (1:400 dilution, Invitrogen) for 1 h on ice, washed, counterstained with DAPI (1 μg/mL) to exclude dead cells and analyzed on a Cyan ADP flow cytometer (Dako) as described;27 4 × 104 events (living cells) were collected per sample. Percentage of CAR and EGFR positive cells was determined compared to control IgG stained cells. Mean fluorescence intensity (MFI) corresponds to the arithmetic mean of the living cell population. 2.9. Cellular Uptake and Colocalization Studies of Coated Ad Particles. For flow cytometry analyses, 5 × 104 cells were seeded in 24 well plates 24 h prior transduction. At the day of transduction cells reached a confluence of 70−80%. Medium was removed, and 250 μL of Opti-MEM containing Ad-Alexa488 only (10,000 VP/cell, MOI 500) or complexed with various amounts of PAMAM G5-Cy5 dendrimer was incubated with cells for 2 h at 37 °C and 5% CO2, thereafter washed twice with PBS, harvested with trypsin/EDTA, fixed with 2% PFA for 30 min at ambient temperature and analyzed by flow cytometry as described above using a FACS Canto II flow cytometer (Becton-Dickinson, Heidelberg, Germany). For laser scanning microscopy (LSM) studies, 2 × 104 cells were seeded in collagen A (0.1 mg/mL) coated chamberslides, 24 h prior to transduction. PAMAM or LPEI coated Ad was incubated with cells for 2 h at 37 °C; thereafter cells were washed three times with PBS and fixed with 4% paraformaldehyde for 30 min. Nuclei were stained using DAPI (1 μg/mL). Samples were analyzed with a confocal laser scanning microscope (LSM 510 META; Carl Zeiss, Jena, Germany) as described in ref 27. 2.10. Transduction Studies and Cell Viability Assay. Cells were seeded at a density of 1 × 104 cells/well in 96 well plates 24 h prior to transduction reaching a confluence of 70− 80% at the day of transduction. The medium was removed, and 60 μL of serum free Opti-MEM (or 2% and 10% FCS in DMEM for stability studies, Figure 5) containing Ad-Luc or Ad-EGFPLuc (MOI 100) alone or complexed with various amounts of dendrimer or polymer was incubated with cells for 2 h at 37 °C (or 45 min respectively for EGFR-targeting experiments). To show the specificity of PAMAM G2-PEGC

dx.doi.org/10.1021/mp300366f | Mol. Pharmaceutics XXXX, XXX, XXX−XXX

Molecular Pharmaceutics

Article

GE11, cells were preincubated with 10 μg/well mEGF for 45 min at 37 °C. DNA/PAMAM polyplexes were also formed within 30 min at RT and incubated with cells for 2 h at 37 °C accordingly. The infection solution was replaced by fresh serum containing medium, and after 48 h cells were lysed using cell lysis buffer (Promega, Mannheim, Germany). Luciferase activity was quantified on a luminometer (Centro LB 960 instrument, Berthold Technologies, Bad Wildbad, Germany).22 Two nanograms of recombinant luciferase (Promega) corresponds to 107 relative light units (RLU). The influence on cell metabolism of the Ad/PAMAM, Ad/LPEI and Ad/BPEI complexes was evaluated using a MTT colorimetric assay. Absorbance was measured at 590 nm with background correction at 630 nm by a microplate reader (Tecan, Grödig, Austria). 2.11. Neutralization Assay. 5 × 104 A549 cells were seeded into 24 well plates 24 h prior to transduction. On the day of transduction cells reached a confluence of 70−80%. Complexes of Ad-EGFP and dendrimer were preformed at ambient temperature for 30 min. Thereafter, 50 μL of Ad or Ad/PAMAM G5 containing 2.5 × 107 VP were mixed with 10 μL of diluted human IgG containing anti Ad antibodies (Privigen, CSL Behring, Marburg, Germany) and incubated for a further 30 min; final IgG concentration was in the range of 0.08−3.3 mg/mL. Cells were transduced with a MOI of 100 under serum free conditions in a total volume of 250 μL in Opti-MEM. After a 2 h incubation, 1 mL of fresh serum containing medium was added and the medium replaced after 24 h. After total 48 h, cells were analyzed for EGFP expression by flow cytometry collecting 2 × 104 gated events per sample. 2.12. Targeting to the EGF Receptor. 5 × 104 cells seeded in 24 well plates 24 h prior to transduction were transduced with Ad-EGFP alone or coated with PAMAM G2PEG-GE11 or PAMAM G2-PEG-Cys at a MOI of 100 under serum free conditions in a total volume of 250 μL. After 45 min incubation at 37 °C the solution was replaced by 1 mL of fresh serum containing medium. After 48 h, cells were harvested, and 2 × 104 gated cells were analyzed for EGFP expression using flow cytometry. For LSM imaging 2 × 104 cells seeded into collagen A coated chamberslides were treated accordingly and after 48 h fixed with 4% PFA for 30 min before evaluation by laser scanning microscopy as described earlier.27

Figure 1. Biophysical evaluation of coated adenovirus. (A) Transmission electron micrographs of PTA stained Ad (top row) and Ad/ PAMAM G5 complexes (bottom row); scale bars: 100 nm. (B) Zeta potential of uncoated Ad versus Ad/PAMAM G5 complexes measured by photon correlation spectroscopy using 5 × 1010 VP in 1 mL of solvent plus the indicated amounts of dendrimer. n = 3 ± SD.

Uncoated Ad capsids display the characteristic hexon structure as well as the icosahedral shape of the viral particle. In contrast, PAMAM G5 coated capsids reveal a rough surface structure, indicating complete coating of the Ad particles. In general, single coated particles were observed, but also smaller numbers of dimers or moderate numbers of aggregates, similar to uncoated Ad, occurred. Single Ad particles doubled in size after coating: capsid diameters of 70−80 nm were measured for uncoated Ad and 140−160 nm for PAMAM coated Ad (Figure 1A and Figure S1 in the Supporting Information). No excessive aggregation of capsids after coating was seen. For measuring the surface charge, 5 × 1010 VP in PBS were coated with indicated amounts of PAMAM G5 diluted in HBG. Uncoated Ad exhibited a negative surface charge with a zeta potential corresponding to −17.4 ± 1.6 mV. In sharp contrast, a zeta potential corresponding to +3.4 ± 0.7 mV was measured when Ad was coated by 20 μg of PAMAM G5 (Figure 1B). 3.2. Influence of PAMAM G5 and LPEI Coating on AdLuc Mediated Transduction and Metabolic Activity of Low CAR or High CAR Cell Lines. For in vitro transduction studies, two cell lines with different CAR expression levels were treated with PAMAM G5 coated Ad: the human glioblastoma cell line U87MG expressing low CAR levels (16% CAR positive, MFI 5), and the human hepatocellular carcinoma cell line HuH7, where all cells are positive for CAR (100%, MFI 135). Cells were transduced with a MOI of 100 (2 × 107 VP/ well of a 96 well plate) with uncoated Ad or Ad coated with PAMAM G5 ranging from 2.5 ng to 1,000 ng per well (Figure 2). On U87MG, uncoated Ad induced only very low luciferase expression levels, whereas with PAMAM coating an up to 36fold increase in luciferase expression was observed at the optimum amount of 10 ng of PAMAM G5 (Figure 2A). Although HuH7 were well transducable with uncoated Ad,

3. RESULTS 3.1. Biophysical Characterization of Ad/PAMAM Complexes. PAMAM dendrimers with terminal primary amino groups were used for the coating of the net negatively charged Ad capsid by virtue of electrostatic interaction. To optimize the coating process, PAMAM and Ad were diluted separately in different buffers, including the low ionic buffer HEPES buffered glucose (HBG), PBS and Opti-MEM. Best transduction results were obtained with Ad diluted in OptiMEM and PAMAM in HBG (data not shown). Hence, all further studies were conducted in this buffer system with the exception of zeta potential measurements, where PBS was used instead of Opti-MEM for Ad dilution, and TEM, where water was used for Ad dilution and PBS for the Ad/PAMAM complex to reduce background signal and background staining. The structure of uncoated Ad and Ad/PAMAM complexes was analyzed by TEM using phosphotungstic acid (PTA) stained samples (Figure 1A). D

dx.doi.org/10.1021/mp300366f | Mol. Pharmaceutics XXXX, XXX, XXX−XXX

Molecular Pharmaceutics

Article

Figure 2. Influence of PAMAM G5 and LPEI coating on transduction efficiency and metabolic activity. 10,000 U87MG glioblastoma (A, C) and HuH7 hepatoma (B, D) cells were transduced in 96 well plates with 2 × 107 Ad-Luc particles/well (MOI 100) either alone (striped bars) or complexed with the indicated amount of PAMAM G5 (gray bars) or LPEI (black bars). Luciferase expression level was evaluated 48 h after transduction for U87MG (A) and HuH7 (B), expressed as relative light units per well (RLU). After 48 h also an MTT assay was performed for U87MG (C) and HuH7 (D). Untreated cells served as control (white bars) with 100% viability. Mean values (n = 3) plus SD are shown.

PAMAM G5 used for coating), when stored at 4 °C. Nevertheless, after storage at room temperature and especially at 37 °C, most of the beneficial effect on transduction due to coating was lost. 3.3. PAMAM Mediated Gene Delivery: A Direct Comparison of Adenovirus and Plasmid. As PAMAMs can be used for both plasmid delivery22 and Ad (see above), we conducted a side-by-side comparison of both systems using the same polycations (Figure S4 in the Supporting Information) and in each case a CMV promoter driven EGFPLuc transgene expression cassette. U87MG and HuH7 cells were transduced with either an optimal amount of Ad (MOI 100, corresponding to 1 × 106 infectious and 1.5 × 107 total VP) or plasmid at an optimal amount of 400 ng/well (corresponding to approximately 5.7 × 1010 plasmid molecules). On the low CAR cell line U87MG, optimal amounts of plasmid condensed with optimal amounts of the most effective PAMAM (400 ng of plasmid condensed with 800 ng of PAMAM G5) could still only reach luciferase levels about 5-fold below Ad transduced cells; compared to coated Ad the level was 102-fold lower. On the CAR positive HuH7 cells, plasmid/PAMAM complexes were 55-fold less efficient compared to the naked Ad and 83fold less efficient than the coated Ad. When taking into account that a 57,000-fold excess of DNA molecules over infectious Ad particles was used, Ad coated with PAMAM dendrimer leads to 5.8 × 106 for U87MG and 4.7 × 106 for HuH7 higher transgene expression levels than DNA/PAMAM polyplexes. 3.4. Cellular Uptake of Fluorescently Labeled Ad and Ad/Polycation Complexes. To correlate improved transduction efficiency with higher uptake levels, we utilized an Ad fluorescently labeled by covalent attachment of amine reactive Alexa-Fluor488-5TFP and coated it with Cy5 labeled PAMAM G5. U87MG and HuH7 were infected with 10,000 particles per cell of Ad-Alexa488 alone or complexed with 2.5, 10 or 100 ng

lower doses of PAMAM (2.5 ng) could further increase luciferase expression by approximately 30%, whereas with higher PAMAM doses the luciferase activity decreased below the level achieved with uncoated Ad (Figure 2B). In addition, we evaluated LPEI (22 kDa), a linear polycation and the “gold standard” in polymer mediated nonviral gene therapy. On U87MG cells, LPEI elevated Ad mediated luciferase expression only 6.5-fold at maximum; on HuH7 cells, all LPEI concentrations tested negatively affected Ad activity. A MTT assay revealed that cells treated with PAMAM G5 maintained high cell viability even when 1,000 ng/well of PAMAM G5 were applied (Figure 2). In contrast, LPEI induced toxicity from 50 ng per well onward leading to only 3% cell viability remaining, when incubated with 1,000 ng of LPEI per well. As a further control, we used 25 kDa branched PEI (BPEI) for Ad coating, which exhibits a molecular weight comparable to that of PAMAM G5, but is randomly branched (Figure S2 in the Supporting Information). When compared to PAMAM G5, the increase in transduction efficiency was far less pronounced on both cell lines tested (HuH7 and U87MG), even less than with LPEI. The reduction in metabolic activity (MTT test) was comparable to that for LPEI, showing toxicity from 100 ng per well onward retaining only 4% cell viability when 1,000 ng of BPEI per well was applied. Usually we prepare coated Ad particles immediately prior to use. As for certain applications preparation of larger amounts “on stock” would be advantageous, we also studied the transduction efficiency of coated Ad after 24 h of storage under different conditions. After coating, particles were stored at 4 °C, room temperature or 37 °C for 24 h, and thereafter HuH7 or U87MG cells transduced and transduction efficiency compared with freshly coated Ad (Figure S3 in the Supporting Information). Transduction efficiency of coated Ad was only slightly lower or unaffected (depending on the amount of E

dx.doi.org/10.1021/mp300366f | Mol. Pharmaceutics XXXX, XXX, XXX−XXX

Molecular Pharmaceutics

Article

Figure 3. Influence of PAMAM coating on cellular uptake. U87MG (low CAR) and HuH7 (high CAR) were transduced with Ad capsids fluorescently labeled with amine reactive Alexa Fluor488-5TFP, either alone or coated with PAMAM G5 (MOI 500, 10,000 VP/cell). After a 2 h incubation period at 37 °C cells were analyzed by either flow cytometry (A, B) or LSM (C−H). (A, B) U87MG (A) or HuH7 cells (B) were incubated with Ad-Alexa488 alone or coated with 2.5, 10 or 100 ng of PAMAM G5. Cells without virus served as control; 20,000 gated events per sample are shown. C−H: U87MG (C-E) and HuH7 (F−H) cells seeded in chamber slides were incubated with Ad-Alexa488 alone (C, F), coated with 10 (D, G) or 100 ng (E, H) of PAMAM G5 and fixed, and the nucleic acid was counterstained with DAPI and analyzed by LSM; for most cells central sections of cells are shown; in panel G some cytoplasmic virus signal (green) appears above the nucleus; Alexa488 signal in green, DAPI signal in blue, scale bar represents 20 μm.

of PAMAM G5. Cells were incubated with Ad for 2 h at 37 °C, thereafter fixed and cellular association and internalization analyzed by flow cytometry and LSM (Figure 3). Total cellular association of uncoated Ad was higher on HuH7 when compared to U87MG cells. After virus coating with 10 ng of PAMAM G5, a 6-fold increase in MFI was seen on U87MG by flow cytometry (Figure 3A). A further increase in PAMAM concentration to 100 ng did not lead to further increased signals, whereas coating with 2.5 ng led to slightly lower signal shifts. On HuH7 cells, the already high cellular association of Ad was increased approximately 3-fold using 10− 100 ng of PAMAM G5 (Figure 2B). While, on U87MG, the increase in Ad binding clearly correlates with the observed transduction benefit in luciferase expression, on HuH7, high amounts of PAMAM G5 (10 ng and 100 ng), which further increase Ad binding to HuH7, led to decreased luciferase expression (Figure 2B). To further clarify these on the first sight contradicting results, we studied intracellular uptake and distribution of Ad and Ad/PAMAM G5 (Figure 3C−H): On the low CAR expressing U87MG cells low uptake of uncoated

Ad (Figure 3C) was observed, whereas with PAMAM coated Ad highly effective internalization occurred, especially when 10 ng of PAMAM was used and Ad/PAMAM particles occurred small and evenly distributed (Figure 3D). Ad/PAMAM G5 complexes were found located in close proximity to the nucleus after 2 h of incubation at 37 °C, indicating that the intracellular trafficking was not grossly negatively influenced. HuH7 on the other hand exhibited, due to their high CAR level, already a significant uptake of uncoated Ad (Figure 3F), while Ad complexed with PAMAM G5 tended to form aggregates at the cell membrane. The aggregation process is already observed with low amounts of PAMAM, i.e., 10 ng (Figure 3G), which is even more pronounced when 100 ng of PAMAM was used (Figure 3H). Apparently, the increase in total cellular association on HuH7 cells was not followed by an increase in transduction efficiency due to a hampered internalization/ delivery process by excessive PAMAM coating. In order to shed light on the reasons for the low transduction performance of LPEI coated Ad, we also studied the role of LPEI on particle internalization and intracellular distribution (Figure S5 in the F

dx.doi.org/10.1021/mp300366f | Mol. Pharmaceutics XXXX, XXX, XXX−XXX

Molecular Pharmaceutics

Article

Figure 4. Colocalization of Alexa488 labeled Ad and Cy5 labeled PAMAM G5 dendrimer. U87MG and HuH7 cells seeded in chamber slides were transduced with Ad-Alexa488 coated with PAMAM G5-Cy5 (MOI 500, 10,000 virus particles/cell). After a 2 h incubation period at 37 °C cells were analyzed by LSM. Central sections of U87MG (A−C) or HuH7 (D−F) incubated with Ad-Alexa488 coated with 100 ng of PAMAM G5-Cy5 are shown. (A, D) Alexa488 signal (green) (B, E) PAMAM G5-Cy5 signal (red). (C, F) Overlay; DAPI signal in blue, colocalization Alexa488 signal and Cy5 signal in yellow.

observed for the Cy5 signal, which is more prominent for the HuH7 when compared to U87MG. We also investigated how the presence of serum influences the transduction efficiency of Ad/PAMAM complexes (Figure 5). Depending on the FCS concentration used, the absolute

Supporting Information). When comparing the uptake of Ad/ LPEI and Ad/PAMAM G5 by LSM it became obvious why PAMAM G5 resulted in the higher transgene expression on both cell lines, U87MG (low CAR) and HuH7 (high CAR): coating with LPEI led to excessive particle aggregation and reduced uptake resulting in only few Ad particles reaching the nuclear membrane after 2 h compared to PAMAM G5 coated Ad. As stability is a major issue for noncovalent Ad coating, U87MG and HuH7 cells were incubated with Ad-Alexa488/ PAMAM G5-Cy5 complexes for 2 h at 37 °C to find out if Ad and PAMAM remain colocalized in the transduction medium and during the cell binding process (Figure 4 and Figure S6 in the Supporting Information). Almost all Ad positive spots exhibit a correlating red signal from the Cy5 labeled PAMAM, which lights up as yellow in Figure 4C,F. In addition, numerous PAMAM signals are found especially on HuH7 cells (Figure 4E,F), which do not correlate with Ad fluorescence. This could also be confirmed using the “histo and profile tool” within the Zeiss 510 META software, where all Alexa488 signals were accompanied by a corresponding more or less prominent PAMAM G5-Cy5 signal, but there were also PAMAM G5-Cy5 peaks without Alexa488 signal (Figure S6 in the Supporting Information). This demonstrates that PAMAM is present in excess, when using 100 ng of PAMAM G5, where the free dendrimer in solution is subsequently internalized by the cells. Colocalization is also observed near the nucleus, indicating that the complexes travel together to the nuclear membrane. In addition, we calculated colocalization coefficients for the signals: in both cell lines the overlap coefficient after Manders32 ranged between 0.7 and 0.8, indicating a high degree of correlation. To obtain a simultaneous, semiquantitative view on cellular Ad and PAMAM association, density blots are shown in Figure S7 in the Supporting Information. For both HuH7 and U87MG there is already a positive effect of 2.5 ng of PAMAM G5 on cellular Ad association, without PAMAM G5-Cy5 being detectable. At a PAMAM concentration of 100 ng, where the dendrimer is also detectable by LSM, there is a clear shift

Figure 5. U87MG cells were transduced with Ad-Luc only or coated with 10 ng, 100 ng or 250 ng of PAMAM G5 in OptiMEM (white bars) or DMEM supplemented with 2% (gray bars) or 10% FCS (black bars), and the luciferase level was quantified after 48 h. Mean values (n = 5) ± SD.

level of luciferase expression is reduced both for naked Ad (3.5fold with 10% FCS) and coated Ad (between 1.6- and 2.2-fold), but the beneficial effect of PAMAM coating remains also in the presence of 10% FCS. 3.5. Correlation of CAR Level and Transduction Improvement by PAMAM Coating. On five human cancer cell lines with increasing CAR expression levels (low CAR level in SKOV-3 ovarian carcinoma (MFI 3, 0.2% positive) and U87MG glioblastoma (MFI 5, 16%), medium CAR level in A549 lung carcinoma (MFI 33, 98%), high CAR level in DU145 prostate carcinoma (MFI 123.3, 100%) and HuH7 hepatocellular carcinoma (MFI 135, 100%) the transduction efficiency of uncoated Ad was compared with PAMAM G5 coated Ad. Here a clear inverse correlation between the CAR expression level (Figure 6A) and the benefit gained through transduction with Ad/PAMAM G5 complexes compared to G

dx.doi.org/10.1021/mp300366f | Mol. Pharmaceutics XXXX, XXX, XXX−XXX

Molecular Pharmaceutics

Article

3.3 mg/mL (1:30 dilution), 100 ng of PAMAM G5 helps protect the Ad from being inhibited by neutralizing antibodies, therefore still being able to infect 47% of the cells on average (Figure 7). In contrast, uncoated Ad is rapidly inactivated by the IgG exposure reaching baseline level of 0.5% positive cells when incubated with 0.8 mg/mL of IgG (1:120 dilution).

Figure 7. Influence of PAMAM G5 coating on the effect of antiadenovirus neutralizing antibodies on transduction. A549 cells seeded in 24 well plates were transduced with Ad-EGFP alone (triangles) or coated with 25 ng (squares) or 100 ng (diamonds) of PAMAM G5 at a MOI of 100. A range of IgG concentrations (Privigen) were incubated with uncoated Ad-EGFP or Ad-EGFP/ PAMAM G5 complexes for 30 min before application to A549 cells. After 48 h, EGFP expression was analyzed by flow cytometry. Representative data out of 3 independent experiments are shown.

3.7. Targeting of Ad to the EGF Receptor. Having a system at hand that increases uptake in CAR positive, and to a much stronger extent in CAR negative cell lines, the aim was to increase specificity of uptake and therefore transgene expression for tumor cells expressing distinct receptors on their surface. EGFR, upregulated on over 60% of solid tumors and a common target of antitumor therapies, was chosen as target receptor. As EGF, the natural ligand for the EGFR, has a mitogenic potential, we used GE11, a peptide ligand known to bind the EGFR without receptor activation.23 The peptide was coupled to PAMAM G2 dendrimer via a 2 kDa PEG spacer, and a PAMAM G2-PEG-Cys conjugate with a distal cystein was used as control for Ad coating (Figure 8). On HepG2 cells, which express only low levels of EGFR (MFI 7, 29% EGFR positive, Figure S9 in the Supporting Information), no significant difference in transgene expression was observed between PAMAM-PEG-Cys and PAMAM-PEG-GE11 coated Ad-EGFP, neither by flow cytometry (Figure 8A) nor by fluorescence microscopy (Figure 8C,D). In contrast, the GE11 conjugate was able to induce a significant, up to 2.3-fold higher EGFP expression level on the highly EGFR positive lung carcinoma cell line A549 (MFI 213, 100% EGFR positive) when compared to PAMAM-PEG-Cys coating (Figure 8B, flow cytometry; E and F, laser scanning microscopy), especially with lower amounts of conjugate used. To specify the GE11 retargeting effect, a preincubation of cells with 10 μg/well mEGF prior to transduction was performed (Figure S10 in the Supporting Information). This led to a significant downregulation of the EGFR on the cell surface and reduced transgene expression using the PAMAM G2-PEG-GE11 coated Ad, whereas PAMAM G2-PEG-Cys coated Ad was unaffected.

Figure 6. Correlation between CAR level and gain in transduction efficiency obtained with PAMAM G5. (A) Indicated cell lines where stained for CAR and analyzed by flow cytometry. Control staining was performed with IgG control antibody. (B) Cell lines were transduced in 96 well plates at MOI 100 with Ad-Luc alone or coated with an optimized amount of PAMAM G5 (SKOV-3, 100 ng/well; U87MG, 10 ng; A549, 50 ng; DU145, 50 ng; HuH7, 5 ng), and the luciferase activity was quantified after 48 h. Relative transfection efficiency is expressed as ratio between luciferase activity obtained with PAMAM G5 coated Ad-Luc and uncoated Ad-Luc. Mean values (n = 3) ± SD.

uncoated Ad can be observed (Figure 6B). Meanwhile we did not observe any correlation between αvβ3 integrin expression levels and improved transduction efficiency due to PAMAM coating (data not shown). SKOV-3, which are virtually CAR negative, showed the highest gain in transduction efficiency (46-fold) with 100 ng of PAMAM G5 used for coating. For A549 still a 10-fold increase could be found, with an optimum of 50 ng of PAMAM G5, whereas HuH7, which show the highest CAR level, revealed only a 1.3-fold increase with a maximum at 5 ng of PAMAM G5 used. The transgene expression values for all coating ratios of LPEI and PAMAM G2, G3 and G5 are shown in Figure S8 in the Supporting Information. 3.6. Dendrimer Coating Prevents Inhibition by Neutralizing Antibodies. In order to study the ability of the PAMAM G5 dendrimer to protect the Ad from neutralizing antibodies, uncoated Ad-EGFP or preformed Ad-EGFP/ PAMAM G5 complexes were incubated for 30 min with increasing concentrations of human IgG (Privigen, obtained from pooled plasma from 60,000 donors, stock 100 mg/mL) before transducing A549 cells. FACS analysis for EGFP 48 h after transduction revealed that, at IgG concentrations reaching

4. DISCUSSION Chemical modification of adenoviral capsids with polymers or lipids has been described by several groups to be beneficial for H

dx.doi.org/10.1021/mp300366f | Mol. Pharmaceutics XXXX, XXX, XXX−XXX

Molecular Pharmaceutics

Article

Figure 8. Targeting of Ad to EGFR using PAMAM G2-PEG-GE11 for coating. HepG2 (A, C, D) or A549 (B, E, F) cells seeded in 24 well plates (A, B) or chamberslides (C−F) were transduced at a MOI of 100 with Ad-EGFP alone or coated with the indicated amount of PAMAM G2-PEG-GE11 or PAMAM G2-PEG-Cys. EGFP expression was analyzed 48 h after transduction. (A, B) MFI of transduced cells (striped bars, Ad-EGFP; light gray bars, PAMAM G2-PEG-Cys coated Ad; dark gray bars, PAMAM G2-PEG-GE11 coated Ad. (C−F) Laser scanning microscopy of cells transduced with Ad-EGFP coated with 25 ng of PAMAM G2-PEG-Cys (C, E) or PAMAM G2-PEG-GE11 (D, F). * p < 0.05, ** p < 0.01, PAMAM G2-PEGCys versus PAMAM G2-PEG-GE11, t test unpaired.

One aspect could be a difference in cellular toxicity: We could show that PAMAM G5 dendrimer is virtually nontoxic up to a concentration of 20 μg/mL used in vitro.22 When comparing PAMAM G5 and LPEI polyplexes after systemic administration in vivo, PAMAM G5 is well tolerated up to a dose of almost 7 mg/kg, whereas with LPEI polyplexes22 and BPEI35 polyplexes liver toxicity already occurs at 2 mg/kg. Here we observed a similar trend, where LPEI and BPEI coated Ad leads to pronounced toxicity at higher concentrations, whereas PAMAM is well tolerated up to 1,000 ng/well (Figure 2C,D, Figure S4 in the Supporting Information). With a polycation transfection reagent at hand, which enhances both viral and plasmid mediated gene delivery, we conducted a side-by-side comparison of both systems (Figure S4 in the Supporting Information). With optimized plasmid content of 400 ng and PAMAM G5/plasmid w/w ratio of 2/1, the luciferase activity measured was at least 5-fold below the activity achieved with EGFPLuc encoding, uncoated Ad on a low CAR level cell line. When calculating the number of DNA molecules used per well, optimal nonviral transfection still requires a 3,700-fold higher number of gene copies when compared to the total Ad particles and a 5,700-fold higher number in relation to the infectious Ad particles. Although the uptake route of both systems (PAMAM polyplex and PAMAM coated Ad) is similar, this study confirms the great potential of Ad, once inside the cell, in overcoming intracellular hurdles, like endosomal release and nuclear transport but especially efficient transcription. According to mathematical models, the major advantage of Ad lies in postdelivery events, as already a small number of intranuclear Ad genomes is able to induce high transgene expression levels, i.e., >10,000-fold more efficient then plasmid DNA.36 In our study we even found a maximum benefit, depending on the cell line used, of 3 × 105 to 3.7 × 105 taking the whole Ad particles into account and 4.7 × 106 to 5.8 × 106 when referring only to the infectious particles comparing DNA/PAMAM polyplexes with Ad/PAMAM in matters of transgene expression.

the reduction of immunogenic response, enhancement of transduction efficiency and targeting to receptors others than CAR.2−4 Here we utilized the cationic molecules PAMAM dendrimers and linear polyethylenimine to coat the Ad capsid by virtue of electrostatic interaction to make low CAR level cell lines accessible to Ad infection. Targeting to the EGFR is achieved with an EGFR binding peptide covalently coupled to PAMAM. This represents a vital alternative to genetic targeting approaches, where ligands are usually presented at the distal end of the fiber.33 When measuring the zeta potential of Ad coated with increasing amounts of PAMAM G5, we observed a strong increase after coating leading to positively charged particles (Figure 1). Structural studies with TEM revealed that mostly single capsids and some moderate aggregates exist, which are completely coated with PAMAM G5 above a ratio of 200 PAMAM G5 molecules per capsid (data not shown). Early studies conducted with, e.g., cationic liposomes as Ad coat rather showed large aggregates of several capsids.20 Nonetheless, for further in vivo studies, e.g., systemic tumor targeting, this would be less beneficial. First cell transductions were carried out on the low CAR level cell line U87MG and the high level CAR cell line HuH7 to determine optimal amounts of PAMAM G5 needed for Ad coating (Figure 2). In addition, LPEI was utilized for Ad coating, as this polycation is currently one of the most efficient transfection reagents on the market for nonviral gene delivery.34 With PAMAM G5, already low concentrations boosted transgene expression on U87MG cells, but only marginally increased transduction efficiency on HuH7. In sharp contrast LPEI only moderately increased transduction efficiency on U87MG cells, whereas on HuH7 cells LPEI even led to a strong decrease. Of interest, the beneficial effect was apparently not solely due to the branched (PAMAM) versus linear (LPEI) structure of the polycation: with 25 kDa of BPEI, which bears a branching point on average on one-third of the nitrogen atoms, the beneficiary effect of coating was even lower than with LPEI (Figure S2 in the Supporting Information). I

dx.doi.org/10.1021/mp300366f | Mol. Pharmaceutics XXXX, XXX, XXX−XXX

Molecular Pharmaceutics

Article

polycation coated Ad particles can be retargeted to this pathway, which also leads to efficient transduction. Polycation induced toxicity usually increases with its molecular weight,22 hence we also screened lower generation PAMAMs (G2 and G3) for their ability to improve Ad transduction (Figure S8 in the Supporting Information). Although on most cell lines the highest increase was found for PAMAM G5, also PAMAM G3 and G2 led to considerable improvements; with the latter higher concentrations were necessary to obtain optimal improvement of transduction efficiency. Former studies by Szoka et al. demonstrated that only PAMAM G > 4 were able to transfect cells with plasmid;39 apparently in the case of Ad complexes the increase in cellular binding, followed by enhanced cellular uptake, appears to be efficient for all dendrimers tested, whereas endosomal escape and nuclear import of transferred DNA is achieved by Ad itself. After intravascular administration, most Ad subtypes interact with several plasma proteins. This interaction determines the virus’s fate: binding of coagulation factor X directs the Ad to the liver leading to hepatocyte transduction,8 complement and scavenger receptor binding causes Ad internalization and inactivation by Kupffer cells,40 while the high prevalence of Ad 5 neutralizing antibodies in the human population leads to inactivation of Ad.41 As repeated application cycles of gene vectors are usually necessary for efficient treatment of disease, the high immunogenic potential of systemically administered Ad, which will rapidly induce the expression of neutralizing antibodies, represents a serious problem. To test how efficiently the PAMAM coat can prevent antibody binding, we performed a neutralization assay by preincubating uncoated Ad-EGFP and Ad-EGFP/PAMAM G5 complexes with increasing concentrations of plasma pooled from 60,000 mostly seropositive donors (Privigen) and thereafter incubated the samples with A549 cells. While uncoated Ad was rapidly inactivated, PAMAM G5 coated Ad could still infect 60% of the cells at the same IgG concentration (Figure 7). Antibodies able to neutralize Ad 5 vectors bind epitopes either on the hexon or fiber proteins.42 Here, the PAMAM coating seems to efficiently protect Ad capsid proteins from this interaction. To make sure that the effect seen is due to masking of the Ad surface by PAMAM and that anti-Ad antibodies are not scavenged by free PAMAM dendrimer, we performed an indirect ELISA. Virus particles absorbed to a plate were incubated with Privigen dilutions preincubated for 30 min at RT with 100 ng of PAMAM G5 or HBG buffer as control before adding them to immobilized Ad. For both PAMAM and HBG preincubation we observed similar absorption values indicating that Ad antibodies were able to bind to the Ad surface regardless of PAMAM present in the solvent (data not shown). The EGF receptor (EGFR) is upregulated in >60% of all solid tumors and a valid surface marker for targeting of anticancer drugs. Numerous attempts with adenoviral vectors have proven its applicability for targeted gene delivery (see refs 43−45). In principle, Ad targeting to the EGFR is possible with a nonmitogenic ligand like cetuximab, an anti EGFR chimeric human/mouse antibody, which led to successful internalization into EGFR positive cells.46 Here, we coupled the short peptide GE11, which has been identified by phage display technique,23 covalently to the PAMAM dendrimers prior to Ad coating. In contrast to the natural ligand EGF, this peptide avoids activation of the EGFR after binding, which could otherwise induce mitogenic activity.23,27 On the EGFR overexpressing cell line A549, GE11 enabled a significant, 2.3-fold higher transgene

Additionally we had a look at the aggregation behavior of the Ad/polycation complexes, as it is known for polyplexes consisting of LPEI and plasmid that they can rapidly aggregate on the cell surface, when added to the cell culture media.37 When conducting LSM studies, LPEI/Ad complexes were found to form large aggregates remaining mostly on the cell surface, whereas PAMAM/Ad complexes remained small and were well internalized (Figure S5 in the Supporting Information). This effect, besides the toxicity observed, appears to be the reason for the lower transduction efficiency, as these big particles were poorly internalized by the cells. When analyzing cellular Ad association semiquantitatively by flow cytometry, the strong increase in transduction efficiency observed for Ad/PAMAM G5 on U87MG in Figure 2A was also reflected by elevated levels of cellular association (Figure 3A); accordingly, the increase on HuH7 was less pronounced, as the uptake rate for uncoated Ad was already high (Figure 3B). This clearly points to a correlation of CAR level and Ad binding. As the binding capacity of negatively charged Ad capsids for cationic molecules is limited, an excess of cationic charge might result in free PAMAM molecules apart from virus bound ones, which can compete for cellular uptake. This has already been shown for nonviral transfections, where free polycations can be of advantage for the endosomal release of plasmid DNA delivered, but at the same time also compete for cellular uptake.30 Apparently, with Ad similar effects in terms of competition occur. As for transduction, also cellular Ad/ PAMAM G5 association was reaching a maximum at 10 ng of PAMAM used. While Ad association could not be further boosted using higher amounts than 10 ng, the fluorescent signal for labeled PAMAM G5 further increased at a dose of 100 ng, indicating the uptake of free dendrimer (Figure 4). LSM studies could also reveal that low amounts of PAMAM G5 enable internalization of small coated Ad particles, whereas with 100 ng there was already a tendency toward larger particles (Figure 3). We conclude from these data that, depending on the cell line, excess of polycation can either induce toxic effects (in the case of LPEI), compete with coated Ad for cell uptake or lead to accumulation on the cell surface (PAMAM and LPEI on HuH7). As coated Ad is also found in the perinuclear region of cells (Figure 4 and Figure S6 in the Supporting Information), we deduce that Ad/polycation complexes are transported together once inside the cell, while the coat does not negatively affect transport toward the nucleus. The beneficial effect of PAMAM coating on transduction efficiency is also retained in the presence of 10% serum during the transduction process, illustrating integrity of the complexes in the presence of negatively charged albumin (Figure 5). Apparently, PAMAM coating of Ad can compensate for lack of CAR mediated binding. To ensure that this uptake route also results in successful transduction, the gain in transduction efficiency obtained with PAMAM G5 coating was correlated to the CAR level on different cell lines with varying CAR expression levels (Figure 6). We could see a clear inverse correlation of the CAR level and the increase in transduction efficiency gained: on SKOV-3 and U87MG, which express low CAR levels, transduction was increased 38- to 46-fold, A549 cells with medium CAR level showed a 10-fold increase, whereas on DU145 and HuH7 exhibiting high CAR levels the increase was marginal. For nonviral polyplexes it has been postulated that cellular binding and internalization occurs via heparan sulfate proteoglycans (HSPGs).38 Apparently, also J

dx.doi.org/10.1021/mp300366f | Mol. Pharmaceutics XXXX, XXX, XXX−XXX

Molecular Pharmaceutics

Article

hybrid system is in principle suitable for treatment of metastasized disease. Furthermore, the applied Ad dose can be reduced due to the gain on transduction efficiency by targeting, which reduces toxicity and immune response of the host. Our ongoing studies imply that this system is suitable for targeting human xenografts in mice after systemic vector injection.51 We used the theranostic gene for the human sodium iodide symporter (hNIS) as a reporter and therapy gene using either replication competent or incompetent Ad5 vector coated with PAMAM G5 as described in the manuscript here. Under optimized conditions, PAMAM coating of intravenously injected Ad prevented hNIS expression in the liver, but at the same time increased tumoral transgene expression in a subcutaneously implanted HuH7 hepatoma tumor. Apparently, the noncovalent coating exhibits sufficient stability to induce re- and detargeting effects also in vivo after intravenous injection.

expression level per cell, when compared to the control conjugate PAMAM G2-PEG-Cys lacking the ligand (Figure 8). On HepG2, which exhibit only low EGFR levels, no significance could be seen. To prove the specificity of the GE11 mediated retargeting effect, we preincubated A549 cells with mEGF, which led to a profound internalization of the EGFR as shown by FACS analysis (Figure S10 in the Supporting Information). Only with PAMAM-PEG-GE11 transgene expression was reduced after mEGF preincubtion, whereas PAMAM-PEG-Cys was unaffected. This clearly demonstrates the selectivity of the system for EGFR. We also used PAMAM G5-PEG-GE11 for coating, although this conjugate was less efficient in terms of targeting, when compared to G2 (data not shown). The affinity of GE11 to EGFR has been reported to be approximately 10−20-fold lower, when compared to EGF.23 Hence, a high number of GE11 ligands is necessary to achieve a cooperative binding effect of several GE11 molecules in close vicinity to cluster EGFR.47 Due to the lower molecular weight of PAMAM G2 (3,284 Da) and a higher coupling rate (PAMAM G2:GE11 of 1:0.75) when compared to PAMAM G5-PEG-GE11 (MW PAMAM G5: 28,854 Da, coupling rate PAMAM G5:GE11 of 1:0.47), the GE11 density is in theory 14 times higher with the G2 derivate explaining its superiority. Very recently we could demonstrate that GE11 peptide per se as single molecule has no measurable binding affinity for the EGFR, but when several peptides, covalently attached to LPEI, are incorporated into a particle with nucleic acid, EGFR binding is specific and measurable.47 Coating of Ad with PAMAM G2-PEG-GE11 led to a significant, but less pronounced, targeting effect when compared to other EGFR retargeting studies performed using the natural ligand EGF covalently coupled either via poly[N-2hydroxypropyl)methacrylamide] (HPMA)46 or via biotin− streptavidin interaction.43 Morrison et al. demonstrated that HPMA coupled EGF is still able to activate the EGFR leading to its internalization and initiation of downstream signaling.44 From our own work with nonviral gene carriers we learned that this leads to an overall boost in particle uptake due to generally increased surface internalization and macropinocytosis induced by EGFR activation.27 Still, GE11 can lead to similar effective uptake rates but after prolonged incubation times, with specific binding to the EGFR and subsequent actin mediated internalization. 48 The absence of EGFR activation by PAMAM G2-PEG-GE11, which would otherwise lead to macropinocytosis and mitotic activity, but also the fact that noncovalent PAMAM coating does not cause an ablation of viral infection as shown for HPMA coated virus explain this lower increment in transduction efficiency by EGFR targeting. Nevertheless GE11 is a promising and safe candidate for adenovirus retargeting studies to cancer cells, and has been proven to specifically enhance transgene expression in EGFR overexpressing tumors after systemic injection of gene carriers.49 Taken together, this study presents a novel strategy for noncovalent coating of adenoviral capsids with low generation PAMAM dendrimers. With covalent linkage of polymers to the Ad capsid a hampered endosomal release and an interference with intracellular trafficking is often seen (F. Kreppel, unpublished observations); this was not the case using PAMAM dendrimers in this charged base coating strategy, where efficient transduction of cancer cells with low or absent CAR expression is obtained. As low CAR and high EGFR level correlate with the aggressiveness and state of a tumor,50 this

■

ASSOCIATED CONTENT

S Supporting Information *

Additional figures as discussed in the text. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*C. Spitzweg: Department of Internal Medicine II, LMU, Marchioninistrasse 15, 81377 Munich, Germany; phone, 49-897095-0; fax, 49-89-7095-8887; e-mail, Christine.Spitzweg@ med.uni-muenchen.de. M.O.: Department of Pharmacy, Center for Drug Research, Pharmaceutical Biotechnology, Butenandtstrasse 5-13, 81377 Munich, Germany; phone, 49-89-2180-0; fax, 49-89-2180-77791; e-mail, [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This study was supported by Grant SFB 824 (Sonderforschungsbereich 824) from the Deutsche Forschungsgemeinschaft (Bonn, Germany) to C. Spitzweg and M.O., grant SPP1230 (Schwerpunktprogramm 1230) from the Deutsche Forschungsgemeinschaft to F.K., a grant from the Center for Nanoscience (CeNS) to M.O. and C. Scheu, and a grant from the Wilhelm-Sander-Stiftung (2008.037.1) to C. Spitzweg. C. Scheu and E.W. received financial support from the Nanosystems Initiative Munich (NIM), and K.S.V. would like to thank Elitenetzwerk Bayern for financial support.

■

REFERENCES

(1) McConnell, M. J.; Imperiale, M. J. Biology of adenovirus and its use as a vector for gene therapy. Hum. Gene Ther. 2004, 15 (11), 1022−33. (2) Fasbender, A.; Zabner, J.; Chill¢n, M.; Moninger, T. O.; Puga, A. P.; Davidson, B. L.; Welsh, M. J. Complexes of adenovirus with polycationic polymers and cationic lipids increase the efficiency of gene transfer in vitro and in vivo. J. Biol. Chem. 1997, 272 (10), 6479− 89. (3) Kreppel, F.; Kochanek, S. Modification of adenovirus gene transfer vectors with synthetic polymers: a scientific review and technical guide. Mol. Ther. 2008, 16 (1), 16−29. (4) Fisher, K. D.; Seymour, L. W. HPMA copolymers for masking and retargeting of therapeutic viruses. Adv. Drug Delivery Rev. 2010, 62 (2), 240−5.

K

dx.doi.org/10.1021/mp300366f | Mol. Pharmaceutics XXXX, XXX, XXX−XXX

Molecular Pharmaceutics

Article

(5) Bergelson, J. M.; Cunningham, J. A.; Droguett, G.; Kurt-Jones, E. A.; Krithivas, A.; Hong, J. S.; Horwitz, M. S.; Crowell, R. L.; Finberg, R. W. Isolation of a common receptor for Coxsackie B viruses and adenoviruses 2 and 5. Science 1997, 275 (5304), 1320−3. (6) Dechecchi, M. C.; Tamanini, A.; Bonizzato, A.; Cabrini, G. Heparan sulfate glycosaminoglycans are involved in adenovirus type 5 and 2-host cell interactions. Virology 2000, 268 (2), 382−90. (7) Wickham, T. J.; Mathias, P.; Cheresh, D. A.; Nemerow, G. R. Integrins alpha v beta 3 and alpha v beta 5 promote adenovirus internalization but not virus attachment. Cell 1993, 73 (2), 309−19. (8) Waddington, S. N.; McVey, J. H.; Bhella, D.; Parker, A. L.; Barker, K.; Atoda, H.; Pink, R.; Buckley, S. M.; Greig, J. A.; Denby, L.; Custers, J.; Morita, T.; Francischetti, I. M.; Monteiro, R. Q.; Barouch, D. H.; van Rooijen, N.; Napoli, C.; Havenga, M. J.; Nicklin, S. A.; Baker, A. H. Adenovirus serotype 5 hexon mediates liver gene transfer. Cell 2008, 132 (3), 397−409. (9) Freimuth, P. A human cell line selected for resistance to adenovirus infection has reduced levels of the virus receptor. J. Virol. 1996, 70 (6), 4081−5. (10) Kaneda, Y. Update on non-viral delivery methods for cancer therapy: possibilities of a drug delivery system with anticancer activities beyond delivery as a new therapeutic tool. Expert Opin. Drug Delivery 2010, 7 (9), 1079−93. (11) Sonawane, N. D.; Szoka, F. C., Jr.; Verkman, A. S. Chloride Accumulation and Swelling in Endosomes Enhances DNA Transfer by Polyamine-DNA Polyplexes. J. Biol. Chem. 2003, 278 (45), 44826−31. (12) Fishburn, C. S. The pharmacology of PEGylation: balancing PD with PK to generate novel therapeutics. J. Pharm. Sci. 2008, 97 (10), 4167−83. (13) Ogris, M.; Wagner, E. To be targeted: Is the magic bullet concept a viable option for synthetic nucleic acid therapeutics? Hum. Gene Ther. 2011, 22 (7), 799−807. (14) Kamiya, H.; Tsuchiya, H.; Yamazaki, J.; Harashima, H. Intracellular trafficking and transgene expression of viral and nonviral gene vectors. Adv. Drug Delivery Rev. 2001, 52 (3), 153−64. (15) Arcasoy, S. M.; Latoche, J. D.; Gondor, M.; Pitt, B. R.; Pilewski, J. M. Polycations increase the efficiency of adenovirus-mediated gene transfer to epithelial and endothelial cells in vitro. Gene Ther. 1997, 4 (1), 32−38. (16) Han, J.; Zhao, D.; Zhong, Z.; Zhang, Z.; Gong, T.; Sun, X. Combination of adenovirus and cross-linked low molecular weight PEI improves efficiency of gene transduction. Nanotechnology 2010, 21 (10), 105106. (17) Kasman, L. M.; Barua, S.; Lu, P.; Rege, K.; Voelkel-Johnson, C. Polymer-enhanced adenoviral transduction of CAR-negative bladder cancer cells. Mol. Pharmaceutics 2009, 6 (5), 1612−9. (18) Meunier-Durmort, C.; Grimal, H.; Sachs, L. M.; Demeneix, B. A.; Forest, C. Adenovirus Enhancement Of Polyethylenimine Mediated Transfer Of Regulated Genes In Differentiated Cells. Gene Ther. 1997, 4 (8), 808−14. (19) Park, K. H.; Yun, C. O.; Kwon, O. J.; Kim, C. H.; Kim, J. R.; Cho, K. H. Enhanced delivery of adenovirus, using proteoliposomes containing wildtype or V156K apolipoprotein A-I and dimyristoylphosphatidylcholine. Hum. Gene Ther. 2010, 21 (5), 579−87. (20) Qiu, C.; De Young, M. B.; Finn, A.; Dichek, D. A. Cationic liposomes enhance adenovirus entry via a pathway independent of the fiber receptor and alpha(v)-integrins. Hum. Gene Ther. 1998, 9 (4), 507−20. (21) Chollet, P.; Favrot, M. C.; Hurbin, A.; Coll, J. L. Side-effects of a systemic injection of linear polyethylenimine-DNA complexes. J. Gene Med. 2002, 4 (1), 84−91. (22) Navarro, G.; Maiwald, G.; Haase, R.; Rogach, A. L.; Wagner, E.; de Ilarduya, C. T.; Ogris, M. Low generation PAMAM dendrimer and CpG free plasmids allow targeted and extended transgene expression in tumors after systemic delivery. J. Controlled Release 2010, 146 (1), 99−105. (23) Li, Z.; Zhao, R.; Wu, X.; Sun, Y.; Yao, M.; Li, J.; Xu, Y.; Gu, J. Identification and characterization of a novel peptide ligand of

epidermal growth factor receptor for targeted delivery of therapeutics. FASEB J. 2005, 19 (14), 1978−85. (24) Schaffert, D.; Kiss, M.; Rodl, W.; Shir, A.; Levitzki, A.; Ogris, M.; Wagner, E. Poly(I:C)-Mediated Tumor Growth Suppression in EGFReceptor Overexpressing Tumors Using EGF-Polyethylene GlycolLinear Polyethylenimine as Carrier. Pharm. Res. 2011, 28 (4), 731−41. (25) Schiedner, G.; Hertel, S.; Kochanek, S. Efficient transformation of primary human amniocytes by E1 functions of Ad5: generation of new cell lines for adenoviral vector production. Hum. Gene Ther. 2000, 11 (15), 2105−16. (26) Kreppel, F.; Biermann, V.; Kochanek, S.; Schiedner, G. A DNAbased method to assay total and infectious particle contents and helper virus contamination in high-capacity adenoviral vector preparations. Hum. Gene Ther. 2002, 13 (10), 1151−6. (27) Schafer, A.; Pahnke, A.; Schaffert, D.; van Weerden, W. M.; de Ridder, C. M.; Rodl, W.; Vetter, A.; Spitzweg, C.; Kraaij, R.; Wagner, E.; Ogris, M. Disconnecting the Yin and Yang Relation of Epidermal Growth Factor Receptor (EGFR)-Mediated Delivery: A Fully Synthetic, EGFR-Targeted Gene Transfer System Avoiding Receptor Activation. Hum. Gene Ther. 2011, 22 (12), 1463−73. (28) Snyder, S. L.; Sobocinski, P. Z. An improved 2,4,6trinitrobenzenesulfonic acid method for the determination of amines. Anal. Biochem. 1975, 64 (1), 284−8. (29) Carlsson, J.; Drevin, H.; Axen, R. Protein thiolation and reversible protein-protein conjugation. N-Succinimidyl 3-(2pyridyldithio)propionate, a new heterobifunctional reagent 90. Biochem. J. 1978, 173, 723−37. (30) Boeckle, S.; von Gersdorff, K.; van der Piepen, S.; Culmsee, C.; Wagner, E.; Ogris, M. Purification of polyethylenimine polyplexes highlights the role of free polycations in gene transfer. J. Gene Med. 2004, 6 (10), 1102−11. (31) Chen, H. Comparative observation of the recombinant adenoassociated virus 2 using transmission electron microscopy and atomic force microscopy. Microsc. Microanal. 2007, 13 (5), 384−9. (32) Manders, E. M. M.; Verbeek, F. J.; Aten, J. A. Measurement of colocalization of objects in dual-color confocal images. J. Microsc. (Oxford, U.K.) 1993, 169, 375−82. (33) Alemany, R. Designing adenoviral vectors for tumor-specific targeting. Methods Mol. Biol. 2009, 542, 57−74. (34) Vicennati, P.; Giuliano, A.; Ortaggi, G.; Masotti, A. Polyethylenimine in medicinal chemistry. Curr. Med. Chem. 2008, 15 (27), 2826−39. (35) Russ, V.; Elfberg, H.; Thoma, C.; Kloeckner, J.; Ogris, M.; Wagner, E. Novel degradable oligoethylenimine acrylate ester-based pseudodendrimers for in vitro and in vivo gene transfer. Gene Ther. 2008, 15 (1), 18−29. (36) Varga, C. M.; Tedford, N. C.; Thomas, M.; Klibanov, A. M.; Griffith, L. G.; Lauffenburger, D. A. Quantitative comparison of polyethylenimine formulations and adenoviral vectors in terms of intracellular gene delivery processes. Gene Ther. 2005, 12 (13), 1023− 32. (37) Wightman, L.; Kircheis, R.; Rossler, V.; Carotta, S.; Ruzicka, R.; Kursa, M.; Wagner, E. Different behavior of branched and linear polyethylenimine for gene delivery in vitro and in vivo. J. Gene Med. 2001, 3 (4), 362−72. (38) Kopatz, I.; Remy, J. S.; Behr, J. P. A model for non-viral gene delivery: through syndecan adhesion molecules and powered by actin. J. Gene Med. 2004, 6 (7), 769−76. (39) Haensler, J.; Szoka, F. C., Jr. Polyamidoamine cascade polymers mediate efficient transfection of cells in culture. Bioconjugate Chem. 1993, 4 (5), 372−9. (40) Xu, Z.; Tian, J.; Smith, J. S.; Byrnes, A. P. Clearance of adenovirus by Kupffer cells is mediated by scavenger receptors, natural antibodies, and complement. J. Virol. 2008, 82 (23), 11705−13. (41) Chirmule, N.; Propert, K.; Magosin, S.; Qian, Y.; Qian, R.; Wilson, J. Immune responses to adenovirus and adeno-associated virus in humans. Gene Ther. 1999, 6 (9), 1574−83. (42) Bradley, R. R.; Lynch, D. M.; Iampietro, M. J.; Borducchi, E. N.; Barouch, D. H. Adenovirus serotype 5 neutralizing antibodies target L

dx.doi.org/10.1021/mp300366f | Mol. Pharmaceutics XXXX, XXX, XXX−XXX

Molecular Pharmaceutics

Article

both hexon and fiber following vaccination and natural infection. J. Virol. 2012, 86 (1), 625−9. (43) Pereboeva, L.; Komarova, S.; Roth, J.; Ponnazhagan, S.; Curiel, D. T. Targeting EGFR with metabolically biotinylated fiber-mosaic adenovirus. Gene Ther. 2007, 14 (8), 627−37. (44) Morrison, J.; Briggs, S. S.; Green, N.; Fisher, K.; Subr, V.; Ulbrich, K.; Kehoe, S.; Seymour, L. W. Virotherapy of ovarian cancer with polymer-cloaked adenovirus retargeted to the epidermal growth factor receptor. Mol. Ther. 2008, 16 (2), 244−51. (45) Harvey, T. J.; Burdon, D.; Steele, L.; Ingram, N.; Hall, G. D.; Selby, P. J.; Vile, R. G.; Cooper, P. A.; Shnyder, S. D.; Chester, J. D. Retargeted adenoviral cancer gene therapy for tumour cells overexpressing epidermal growth factor receptor or urokinase-type plasminogen activator receptor. Gene Ther. 2010, 17 (8), 1000−10. (46) Morrison, J.; Briggs, S. S.; Green, N. K.; Thoma, C.; Fisher, K. D.; Kehoe, S.; Seymour, L. W. Cetuximab retargeting of adenovirus via the epidermal growth factor receptor for treatment of intraperitoneal ovarian cancer. Hum. Gene Ther. 2009, 20 (3), 239−51. (47) Abourbeh, G.; Shir, A.; Mishani, E.; Ogris, M.; Rodl, W.; Wagner, E.; Levitzki, A. PolyIC GE11 polyplex inhibits EGFRoverexpressing tumors. IUBMB Life 2012, 64 (4), 324−30. (48) Mickler, F. M.; Mockl, L.; Ruthardt, N.; Ogris, M.; Wagner, E.; Brauchle, C. Tuning Nanoparticle Uptake: Live-Cell Imaging Reveals Two Distinct Endocytosis Mechanisms Mediated by Natural and Artificial EGFR Targeting Ligand. Nano Lett. 2012, 12 (7), 3417−23. (49) Klutz, K.; Schaffert, D.; Willhauck, M. J.; Grunwald, G. K.; Haase, R.; Wunderlich, N.; Zach, C.; Gildehaus, F. J.; SenekowitschSchmidtke, R.; Goke, B.; Wagner, E.; Ogris, M.; Spitzweg, C. Epidermal Growth Factor Receptor-targeted (131)I-therapy of Liver Cancer Following Systemic Delivery of the Sodium Iodide Symporter Gene. Mol. Ther. 2011, 19 (4), 676−85. (50) Matsumoto, K.; Shariat, S. F.; Ayala, G. E.; Rauen, K. A.; Lerner, S. P. Loss of coxsackie and adenovirus receptor expression is associated with features of aggressive bladder cancer. Urology 2005, 66 (2), 441− 6. (51) Grunwald, G. K.; Vetter, A.; Klutz, K.; Willhauck, M.; Schwenk, N.; Senekowitsch-Schmidtke, R.; Schwaiger, M.; Zach, C.; Wagner, E.; Goke, B.; Holm, P. S.; Ogris, M.; Spitzweg, C. Non-Covalent Dendrimer Coating of Adenoviral Vectors for Systemic Sodium Iodide Symporter (NIS)-Mediated Radiovirotherapy of Liver Cancer. Not published.

M

dx.doi.org/10.1021/mp300366f | Mol. Pharmaceutics XXXX, XXX, XXX−XXX