Search

The clinical efficacy of adenovirus serotype 5-based vectors is limited by complex interactions of the vector with host blood components. These interactions trigger sequestration of systemically administered vector particles mainly by the reticuloendothelial system. In order to generate retargeted Ad vectors suitable for i.v. delivery it is mandatory to understand and manipulate these non-target interactions.In mice, it was shown that blood coagulation factor X bound to hexon reduces sequestration by shielding the virus from natural antibodies and macrophage uptake, but also mediates hepatocyte transduction. To generate Ad vectors deficient for hepatocyte transduction but shielded from natural antibodies, we employed a combination of point mutations ablating (i) CAR binding (Ad-ΔCAR), (ii) integrin binding, (iii) FX binding (Ad-ΔFX), and rendered the vectors amenable for position-specific PEGylation at hexon HVR1 (Ad-HVR1) to compensate the lack of a FX-mediated shielding. Surface plasmon resonance analysis confirmed complete ablation of FX binding by ΔFX mutation in combination with PEGylation. Upon i.v. delivery, unPEGylated Ad-HVR1-ΔCAR-ΔFX did not transduce hepatocytes in BALB/c mice, whereas robust transduction levels were observed in antibody-deficient JHD mice. In agreement with the literature, this suggested that the ΔFX vector particles became susceptible to natural antibodies. However and surprisingly, PEGylated Ad-HVR1-ΔCAR-ΔFX showed a strong FX-independent hepatocyte transduction in both BALB/c and JHD mice compared to unPEGylated Ad-HVR1-ΔCAR-ΔFX and wildtype capsid vectors. Preliminary results suggested that integrins were involved in this FX-independent transduction. Further, PEGylated Ad-HVR1-ΔCAR-ΔFX vectors displayed 20-fold increased blood persistence without being associated to blood cells in BALB/c mice.To analyze vector sequestration by macrophages more closely we performed in vitro assays measuring the uptake of vector particles by Raw264.7 cells in the presence of murine plasma. We found enhanced uptake of unPEGylated Ad-HVR1-ΔFX vectors compared to wildtype capsid vectors. Using antibody-deficient and heated plasma of Ad naive mice confirmed a natural antibody- and complement-dependent uptake mechanism. This uptake was completely prevented by PEGylation of the vector particles, suggesting that position-specific PEGylation of hexon HVR1 interfered with natural antibody binding. Moreover, using hirudinized human whole blood we could show that a complement-dependent natural antibody-mediated binding of unPEGylated Ad-HVR1-ΔCAR-ΔFX vectors to erythrocytes could be prevented by PEGylation of hexon HVR1 with large PEG moieties.Thus, we conclude that a PEG-mediated evasion from sequestration by macrophages by shielding from natural antibodies or by a yet unknown mechanism maintained high vector concentrations in blood followed by enhanced hepatocyte transduction.

Adenovirus-based vectors are promising tools for genetic vaccination. However, several obstacles have to be overcome prior to a routine clinical application of adenovirus-based vectors as efficacious vectored vaccines. The linear trisaccharide epitope αGal (alpha-Gal) with the carbohydrate sequence galactose-α-1,3-galactosyl-β-1,4-N-acetylglucosamine has been described as a potent adjuvant for recombinant or attenuated vaccines. Humans and α-1,3-galactosyltransferase knockout mice do not express this epitope. Upon exposure of α-1,3-galactosyltransferase-deficient organisms to αGal in the environment, large amounts of circulating anti-Gal antibodies are produced consistently. Immunocomplexes formed between recombinant αGal-decorated vaccines and anti-Gal antibodies exhibit superior immunogenicity. We studied the effects of the trisaccharide epitope on CD8 T cell responses that are directed specifically to vector-encoded transgenic antigens. For that, covalently αGal-decorated adenovirus vectors were delivered to anti-Gal α-1,3-galactosyltransferase knockout mice. We generated replication-defective, E1-deleted adenovirus type 5 vectors that were decorated with αGal at the hexon hypervariable regions 1 or 5, at fiber knob, or at penton base. Surprisingly, none of the adenovirus immunocomplexes being formed from αGal-decorated adenovirus vectors and anti-Gal immunoglobulins improved the frequencies of CD8 T cell responses against the transgenic antigen ovalbumin. Humoral immunity directed to the adenovirus vector was neither increased. However, our data indicated that decoration of Ad vectors with the αGal epitope is a powerful tool to analyze the fate of adenovirus immunocomplexes in vivo.

Capsid surface shielding of adenovirus vectors with synthetic polymers is an emerging technology to reduce unwanted interactions of the vector particles with cellular and non-cellular host components. While it has been shown that attachment of shielding polymers allows prevention of undesired interactions, it has become evident that a shield which is covalently attached to the vector surface can negatively affect gene transfer efficiency. Reasons are not only a limited receptor-binding ability of the shielded vectors but also a disturbance of intracellular trafficking processes, the latter depending on the interaction of the vector surface with the cellular transport machinery. A solution might be the development of bioresponsive shields that are stably maintained outside the host cell but released upon cell entry to allow for efficient gene delivery to the nucleus. Here we provide a systematic comparison of irreversible versus bioresponsive shields based on synthetic N-(2-hydroxypropyl)methacrylamide (HPMA) copolymers. In addition, the chemical strategy used for generation of the shield allowed for a traceless bioresponsive shielding, i.e., polymers could be released from the vector particles without leaving residual linker residues. Our data demonstrated that only a bioresponsive shield maintained the high gene transfer efficiency of adenovirus vectors both in vitro and in vivo. As an example for bioresponsive HPMA copolymer release, we analyzed the in vivo gene transfer in the liver. We demonstrated that both the copolymer's charge and the mode of shielding (irreversible versus traceless bioresponsive) profoundly affected liver gene transfer and that traceless bioresponsive shielding with positively charged HPMA copolymers mediated FX independent transduction of hepatocytes. In addition, we demonstrated that shielding with HPMA copolymers can mediate a prolonged blood circulation of vector particles in mice. Our results have significant implications for the future design of polymer-shielded Ad and provide a deeper insight into the interaction of shielded adenovirus vector particles with the host after systemic delivery.

Vectors based on Adenovirus type 5 (Ad5) are well suited for the purpose of genetic vaccination due to their high immunogenicity. However, it is known that immune complexes formed upon contact with vector-specific antibodies (Abs) can impede gene transfer and influence vector immune profiles. To model, analyze, and understand the effects of immune complex formation on in vivo biodistribution and vector immunogenicity, we established a novel model based on chemical tagging of Ad capsomers with a carbohydrate epitope. Of note, this model allows to direct antibodies to specific capsid regions and thus to generate highly defined Ad immune complexes independent of species and MHC restrictions.As an artificial immunogenic epitope, we chemically coupled alpha-Gal (Galactose-a1,3-Galactose-b1,4-N-Acetylglucosamine) to specific sites of Ad capsomers. We generated ΔE1 Ad5 vectors being alpha-Galylated at hexon HVR1, hexon HVR5, penton or fiber, and an unmodified control vector.We demonstrated that on account of the alpha-Gal epitope, anti-Gal IgG or IgM bound to selected tagged capsomers. Thus, we analyzed biodistribution and vector-induced immune responses in alpha-1,3-Galactosyltransferase knockout (a1,3-GT KO) mice, which harbor large amounts of anti-Gal Abs.First, to compare the effect of epitope tagging on vector biodistribution, we analyzed hepatic and splenic transgene expression after i.v. injection of alpha-Galylated vectors into anti-Gal-positive a1,3-GT KO mice. We observed that transgene expression from fiber-modified vectors was mildly decreased, while expression from hexon-modified vectors decreased drastically, in a way resembling neutralization. These data suggested that, in fact, our model is suitable to precisely analyze the role of capsomer-specific Ab binding on in vivo vector biodistribution and activity.Furthermore, we analyzed the effect of immune complexation on the induction of vector- and transgene product-directed immune responses after repetitive i.m. delivery of alpha-Galylated vectors into anti-Gal-positive a1,3-GT KO mice. Here, we found that the Ad immune complexes induced transgene product-directed immune responses, even upon repeated vector administration. In particular, the extents of hepatic and splenic transgene expression, and transgene-directed immune responses were inversely correlated. For hexon-modified vectors, in spite of weakest transgene expression, transgene-directed cellular immune responses were strongest.In summary, we showed that alpha-Galylation of Ad capsomers markedly influenced specific immune responses directed to the vector as well as the transgene product. Our concept shall be used as an in vivo model to characterize whether directing or diverting immune complexation to or from specific capsomer sites can improve future strategies for vaccination regimens using Ad vectors.