The protocol described here enables researchers to specifically modify adenovirus capsids at selected sites by simple chemistry. Shielded adenovirus vectors particles and retargeted gene transfer vectors can be generated, and vector host interactions can be studied.
Adenovirus vectors are potent tools for genetic vaccination and oncolytic virotherapy. However, they are prone to multiple undesired vector-host interactions, especially after in vivo delivery. It is a consensus that the limitations imposed by undesired vector-host interactions can only be overcome if defined modifications of the vector surface are performed. These modifications include shielding of the particles from unwanted interactions and targeting by the introduction of new ligands. The goal of the protocol presented here is to enable the reader to generate shielded and, if desired, retargeted human adenovirus gene transfer vectors or oncolytic viruses. The protocol will enable researchers to modify the surface of adenovirus vector capsids by specific chemical attachment of synthetic polymers, carbohydrates, lipids, or other biological or chemical moieties. It describes the cutting-edge technology of combined genetic and chemical capsid modifications, which have been shown to facilitate the understanding and overcoming of barriers for in vivo delivery of adenovirus vectors. A detailed and commented description of the crucial steps for performing specific chemical reactions with biologically active viruses or virus-derived vectors is provided. The technology described in the protocol is based on the genetic introduction of (naturally absent) cysteine residues into solvent-exposed loops of adenovirus-derived vectors. These cysteine residues provide a specific chemical reactivity that can, after production of the vectors to high titers, be exploited for highly specific and efficient covalent chemical coupling of molecules from a wide variety of substance classes to the vector particles. Importantly, this protocol can easily be adapted to perform a broad variety of different (non-thiol-based) chemical modifications of adenovirus vector capsids. Finally, it is likely that non-enveloped virus-based gene transfer vectors other than adenovirus can be modified from the basis of this protocol.
Adenoviruses (Ad), members of the family Adenoviridae, are non-enveloped DNA viruses of which more than 70 types so far have been identified (http://hadvwg.gmu.edu). Depending on hemaglutination properties, genome structure, and sequencing results, the 70 Ad types can be divided into seven species (human adenoviruses A to G)1,2. The human Ad genome is 38 kb in size and encapsulated by an icosahedral nucleocapsid3. Due to their abundance, the capsid protein hexon, penton base, and fiber are all referred to as major capsid proteins. The most abundant and largest capsid protein hexon forms 20 capsid facets, each consisting of 12 hexon homotrimers4,5. Penton, located on each icosahedral edge (vertex), consists of pentamers of a penton base and represents the base for the vertex spike that is built of glycosylated fiber trimers5,6. The native Ad cell entry is basically composed of two major steps. First, the fiber knob binds to the primary receptor. In Ad types from species A, C, E, and F, this is the coxsackie and adenovirus receptor (CAR). This interaction brings the virion into spatial proximity of the cell surface, thereby facilitating interactions between cellular integrins and RGD-motif in the penton base and consequently inducing cellular responses. Second, changes in the cytoskeleton lead to internalization of the virion and transport to the endosome7. Upon partial disassembly in the endosome, the virion is released to the cytoplasm und ultimately travels to the nucleus for replication.
While Ad can be delivered locally (e.g., for genetic vaccination), systemic delivery through the bloodstream as required for onco-virotherapy faces several barriers. While circulating in the bloodstream, injected virions encounter the defense system of the host's immune system, leading to fast neutralization of the virus-based vectors and rendering Ad-based vectors extremely inefficient in systemic applications. Furthermore, the natural hepatotropism of Ad interferes with systemic delivery and must be resolved to redirect Ad to its new target cells.
Germline-encoded natural IgM antibodies of the innate immune system rapidly recognize and bind highly repetitive structures on the surface of the virion8,9. These immune complexes then activate the classical and non-classical pathways of the complement system, leading to fast complement-mediated neutralization of a large portion of the virions8. A second pathway resulting in major removal of Ad virions is mediated by macrophages10 and associated with acute toxic and haemodynamic side effects11,12. In the case of Ad in particular, Kupffer cells residing in the liver bind to and phagocytically take up the Ad virions via specific scavenger receptors, thereby eliminating them from the blood13,14,15. Specific scavenger receptors have also been identified on liver sinusoidal endothelial cells (LSE cells)16,and LSE cells also seem to contribute to vector elimination17, but to what extent still needs clarification. Furthermore, some Ad types and their derived vectors are efficiently sequestered by human erythrocytes18 to which they bind via CAR or the complement receptor CR119. Of note, this sequestering mechanism cannot be studied in the mouse model system as in contrast to human erythrocytes, mouse erythrocytes do not express CAR.
Specific anti-Ad antibodies generated by the adaptive immune system after exposure to Ad either due to previous infections with Ad or after the first delivery in systemic applications raise a further barrier to effective use of Ad vectors, and they should be evaded in efficient systemic delivery.
Finally, the strong hepatotropism of some Ad types (including Ad5) severely hinders application of Ad in systemic therapy. This tropism resulting in hepatocyte transduction is due to the high affinity of the Ad virion to blood coagulation factor X (FX), mediated by the interaction of FX with the Ad hexon protein20,21,22. FX bridges the virion to heparin sulfate glycans (HSPGs) on the surface of hepatocytes20,23,24,25. A crucial factor for this interaction seems to be the specific extent of N- and O-sulfation of the HSPGs in liver cells24, which is distinct from HSPGs on other cell types. In addition to this FX-mediated pathway, recent studies suggest further pathways not yet identified that result in Ad transduction of hepatocytes26,27,28.
Recently, it has been shown that FX is not only involved in hepatocyte transduction of Ad, but also by binding, the virion shields the virus particle against neutralization by the complement system26. Reduction of hepatocyte transduction by preventing FX binding, therefore, would create the unwanted side effect of increasing Ad neutralization via the innate immune system.
A profound knowledge of the complex interactions between vectors and host organisms is therefore necessary to develop more efficient vectors for systemic applications that circumvent the obstacles imposed by the host's organism.
One strategy that has been originally used for therapeutic proteins has been adapted for Ad vectors to at least partially overcome the above described barriers. Antigenicity and immunogenicity of therapeutic protein compounds could be reduced by coupling to polyethylene glycol (PEG)29,30. Hence, the covalent coupling of polymers such as PEG or poly[N-(2-hydroxypropyl)methacrylamid] (pHPMA) to the capsid surface shields the vector from unwanted vector-host interactions. Commonly, polymer coupling targets ε-amine groups from lysine side residues that are randomly distributed on the capsid surface. Vector particles in solution are, due to the hydrophilic nature of the attached polymers, surrounded by a stable water shell that reduces the risk of immune cell recognition or enzymatic degradation. Moreover, PEGylated Ad vectors were shown to evade neutralization by anti-hexon antibodies in vitro and in pre-immunized mice in vivo31. In contrast to genetic capsid modifications, chemical coupling of polymers is performed after production and purification, allowing not only for the use of conventional producer cells and production of high titer vector stocks, but also for simultaneous modification of thousands of amino acids on the capsid surface. However, amine-directed shielding occurs randomly throughout the whole capsid surface, resulting in high heterogeneities and not allowing for modification of specific capsomers. Furthermore, the large polymer moieties required for beneficial effects impair virus bioactivity32.
To overcome these limitations, Kreppel et al.33 introduced a geneti-chemical concept for vector re- and de-targeting. Cysteines were genetically introduced into the virus capsid at solvent-exposed positions like fiber HI-loop33, protein IX34, and hexon35,36. Although not naturally-occurring, cysteine-bearing Ad vectors can be produced at high titers in normal producer cells. Importantly, insertion of cysteines in certain capsomers and in different positions within a single capsomer allows for highly specific modifications of thiol group-reactive moieties. This geneti-chemical approach has been shown to overcome numerous obstacles in Ad vector design. The combination of amine-based PEGylation for detargeting and thiol-based coupling of transferrin to the fiber knob HI-loop has been proven to successfully retarget modified Ad vectors to CAR-deficient cells33. Since hexon is involved in most undesired interactions (neutralizing antibodies, blood coagulation factor FX), thiol-based modification strategies were also applied to hexon. Coupling small PEG moieties to HVR5 of hexon prevented Ad vector particles to transduce SKOV-3 cells in the presence of FX, whereas large PEG moieties increased hepatocyte transduction14,35. Ad vector particles carrying mutations in the fiber knob inhibiting CAR binding and in HVR7 inhibiting binding of FX (and bearing inserted cysteines in HVR1 for position-specific PEGylation) were shown to evade antibody- and complement-mediated neutralization, as well as scavenger receptor-mediated uptake without loss of infectivity. Interestingly, despite a lack of the natural FX shield, PEGylation again improved transduction of hepatocytes as a function of PEG size36. However, it was shown that covalent shielding does have an impact on intracellular trafficking processes. Prill et al. compared irreversible versus bioresponsive shields based on pHPMA and demonstrated that neither the mode of shielding nor co-polymer charge had an impact on cell entry but did affect particle trafficking to the nucleus. Employing a bioresponsive shield with positively charged pHPMA co-polymers allowed for particle trafficking to the nucleus, maintaining the high transduction efficiencies of Ad vectors in vitro and in vivo37.
In summary, these data indicate that, even under the assumption that all vector-host-interactions were known and considered, excessive capsid surface modifications are necessary to overcome the hurdles associated with systemic vector delivery.
Here we provide a protocol to perform site-specific chemical modifications of adenovirus vector capsids for shielding and/or retargeting of adenovirus vector particles and adenovirus-based oncolytic viruses. The concept of this technology is outlined in Figure 1. It allows the shielding of certain capsid regions from unwanted interactions by covalent attachment of synthetic polymers. At the same, it also provides a means to attach ligands and combine shielding and targeting. Using simple chemistry, experimenters will be able to covalently modify adenovirus vector surface with a wide variety of molecules including peptides/proteins, carbohydrates, lipids, and other small molecules. Furthermore, the protocol provides a general concept for the chemical modification of biologically active virus-derived vectors under maintenance of their biological integrity and activity.
NOTE: In the following, a protocol for geneti-chemical PEGylation of an Ad vector is described to detail. To enable specific coupling of the PEG moiety, an Ad5 vector was beforehand genetically modified by introducing a cysteine residue into the hexon protein at the hypervariable loop 5 as described in a previous publication36, and a maleimide-activated PEG compound is used as coupling compound.
1. Preparation of Buffers for Vector Purification by CsCL Step Gradients
2. Coupling Moieties: Storage and Preparation
NOTE: Moeities used for coupling to cysteines need to bear thiol-reactive groups. Maleimide-activated compounds will form stable thioether bonds with the genetically introduced cysteines. Alternatively, ortho-pyridyldisulfide (OPSS)-activated compounds can be used, which form bioresponsive disulfide bridges between the vector particles and coupling moiety. Lyophilized malPEG-750 as well as most other coupling reagents are sensitive to hydrolysis and should be stored dry in the form of lyophilized powders at -80 °C.
3. Amplification, Purification and Chemical Modification of Ad Vectors:
4. Verification of Coupling by SDS-PAGE:
NOTE: If compounds with sufficiently high molecular weights are used for coupling to Ad virions, coupling can be verified by polyacrylamide gel electrophoresis (SDS-PAGE). Successful coupling should then result in a shift of the protein band corresponding to the modified Ad virion protein, compared to the protein in the unmodified Ad virion (see Figure 4).
Figure 2 shows examples of the cytopathic effect (CPE) on 293 (HEK 293) cells that indicates successful vector production. Cells should show morphology (Figure 2C) 40-48 hours after inoculation with the virus vector. The right timepoint for harvesting is crucial for not losing virus particles by cell lysis and preventing oxidation of the genetically introduced thiol groups. If vector particles are released into the medium by cell lysis, the genetically introduced thiols will almost immediately become oxidized, and it will become difficult to purify and chemically modify them.
Figure 3 shows exemplary CsCl bands. The purpose of discontinuous CsCl banding before chemical modification is to remove cellular debris from the virus particles. The modification itself can then take place in CsCl. The second banding after chemical modification serves to remove an excess of (unreacted) coupling moiety. Of note, a successful modification of the virus cannot be seen in the gradient and requires further molecular analysis, ideally by SDS-PAGE.
Figure 4 shows typical examples of successful capsid modifications. Most moieties coupled to specific capsomeres will alter the running behavior of respective capsomeres in SDS-PAGE. By direct comparison with an unmodified control, the success of coupling can be verified, and densitometric analysis can help determine overall coupling efficiency (the percent of shifted and unshifted bands for the modified capsomere).
Figure 1: Combined genetic and chemical capsid modifications of adenovirus vector capsids enable covalent attachment of shielding polymers and targeting ligands. Cysteine residues are genetically introduced at selected, solvent-exposed capsid loops of adenovirus vector capsids to equip the vector particles with new chemical reactivity. The vectors can be produced using conventional producer cells and protocols. After production, the genetically introduced cysteine residues are specifically and covalently modified with thiol-reactive coupling moieties (ligands, shielding polymers, carbohydrates, small molecules, fluorescent dyes, etc.). For coupling, maleimide-activated compounds (which form stable thioether bonds with the vector particle surface) or pyridyldisulfide derivatives (which form bioresponsive disulfide bridges) can be employed. After coupling, the vectors are purified by discontinuous CsCl banding to remove unreacted excess coupling moieties. PEG, polyethylene glycol (shielding polymer); L, ligands (derived from a broad variety of substance classes); -SH, thiol functionality of the genetically introduced cysteine residues. Using this technology, a defined number of molecules can be covalently coupled to the vector capsids, and a precise selection of the coupling site is feasible. Please click here to view a larger version of this figure.
Figure 2: Cytopathic effect during vector production. Conventional producer cells (here, 293-HEK cells) can be used for production of the genetically modified vectors prior to chemical modification. The appearance of CPE indicates the best timepoint to harvest the cells. (A) Cells two hours after infection, (B) cells 24 hours after infection, and (C) cells 40 hours after infection prior to harvesting. Scale bars = 20 µm. Please click here to view a larger version of this figure.
Figure 3: Exemplary results of CsCl gradients. CsCl gradients before (A, B) and after (C) chemical modification. (A) An adenovirus vector is banded by discontinuous CsCl density gradient centrifugation. The upper phase appears green since the vector shown bears an hCMV-promoter-driven expression cassette for EGFP. The vector virion is banded as a single, white band in the lower third of the tube. The two smaller bands (above) stem from incomplete particles, which should not be collected. (B) A cysteine-bearing EGFP-expressing Ad vector before chemical modification. (C) The same vector after modification. The band in the lower third of the tube will be collected and purified by a desalting column. The pen mark in B and C labels the border of the two densities and allows for identifying the location of the vector band that will be collected. Please click here to view a larger version of this figure.
Figure 4: SiIver-stained SDS-PAGE to assess coupling efficiency. MW marker in lane 1. A vector with cysteines introduced into the capsomeres penton base, and hexon was left unmodified (lane 2) or modified with maleimide-activated PEG (polyethylene glycol) with a molecular weight of 5 kDa (lane 3). Both hexon and penton base exhibit a shift during SDS-PAGE, indicating successful modification of > 90% of the monomers per capsid. A vector with a cysteine residue in hexon was left unmodified (lane 4), modified with 5 kDa PEG (lane 5), or with 20 kDa PEG (lane 6). It should be noted that the larger shift of the hexon band is due to the higher molecular weight of the 20 kDa PEG, compared to the 5 kDa PEG (lane 5). The gel demonstrates specificity and efficiency of the coupling. Please click here to view a larger version of this figure.
The efficiency by which the genetically introduced cysteines can be chemically modified is typically 80-99%, and certain variables influence this efficiency. First, it is paramount that the genetically introduced cysteines do not undergo premature oxidation. While being well-protected in the reducing environment of the producer cells, it is mandatory to provide a non-oxidative environment after releasing vector particles from the producer cells and during chemical modification. To this end, reducing reagents can be used at concentrations from 0.1-10 mM, and it is necessary to use reducing reagents that do not contain thiol groups, which would readily react with maleimide-activated compounds. Second, when using maleimide-activated compounds, pH during the coupling reaction should not exceed 7.35, since a pH of greater than 7.4 can decrease the specificity of the reaction. Third, in any case, amine-free buffers (e.g., HEPES, PBS) should be used for the reaction. Of note, cysteine-bearing vector particles can be purified by double CsCl banding as described, then stored in HEPES/10% glycerol prior to chemical modification. In this case, 0.1-1 mM TCEP should be used a reducing reagent, or preferably, the vectors should be stored in an argon atmosphere. Argon-filled plastic jars with lids can be used for this purpose. Additionally, modification efficiency is influenced by the hydrodynamic diameter of the compound coupled to the capsids and the site to which it is coupled. Many capsomeres that can serve as targets for the specific geneti-chemical modification are present in the capsid as trimers (fiber, hexon) or pentamers (penton base). Therefore, depending on the location of the genetically introduced cysteine, a moiety molecule coupled to one monomer might sterically hinder the coupling of another molecule to the neighboring monomer. This should be considered when selecting ideal sites for geneti-chemical capsid modifications.
To facilitate troubleshooting, a few problems may arise when following the protocol described. First, low vector yields (below 10,000 vector particles per producer cell) may result from suboptimal infection of producer cells. Ideally, 100-300 MOI should be used for the infection of producer cells. Please note that both too-high and too-low vector amounts for the inoculum generate suboptimal yields. It is ideal to passage the producer cells the day before infection. Second, smeary bands in the CsCl gradients can appear. The most frequent reason for smeary appearance of bands in the CsCl gradients is a pH of the CsCl solutions that is too acidic. The reducing reagent TCEP is very acidic, and care must be taken to adjust pH before loading the vector onto the gradient, since adenovirus vectors readily disintegrate in an acidic environment. Third, low coupling efficiencies (< 80% of coupling) are the most frequent result of impure vector preparations and/or premature oxidation of thiol groups on the vector surface. Impurities can be detected by SDS-PAGE and silver staining. In the case that significant impurities occur, vector production should be restarted. In addition, low coupling efficiencies can be caused by free non-vector thiols in the reaction. Therefore, it is advised not to use ß-mercaptoethanol or dithiothreitol as reducing reagents, since both contain free thiol groups. Of note, to select solvent-accessible sites to introduce cysteine residues that can readily be modified, crystal structures should be used; otherwise, the cysteines may not be accessible to the coupling partner. Premature oxidation of thiols on the vector surface can be avoided by the stringent use of argon and/or TCEP.
Finally, incorrect stoichiometric calculations can also lead to low coupling efficiencies. The following example is meant to illustrate the generic formula presented above and facilitate the stoichiometric calculations. In the example, one cysteine is introduced into the hexon capsomere. The titer of vector particles after the first CsCl step gradient purification is determined as 1.5 x 1010 vector particles per µL, and as coupling moiety, malPEG-750 is used. Each vector capsid contains 720 hexon proteins (240 trimers), so the titer of cysteines per µL therefore sums to 720 x 1.5 x 1010 = 1.08 x 1013 cysteines/µL. In order to modify 1.5 mL of vector particles, a total of 1.62 x 1016 cysteines must be reacted with the coupling moiety. To reach efficient coupling, a 50x molar excess of the coupling compound to the total of cysteine residues is recommended: 1.62 x 1016 x 50 = 8.1 x 1017 molecules of malPEG-750 is needed to efficiently couple to the 1.5 x 1016 cysteines of 1.5 mL of vector solution purified by the first CsCl step gradient. MalPEG-750 exhibits a molecular weight of 750 Da; hence, 6.022 x 1023 molecules of malPEG-750 weigh 750 g. Therefore, for efficient coupling of the 1.62 x 1016 cysteine residues in 1.5 mL of vector solution with a 50x excess of malPEG-750, 8.1 x 1017 molecules of malPEG-750 = (750 x 8.1 x 1017) / (6,022 x 1023) = 1 x 10-3 g = 1 mg malPEG-750 are required.
The presented method complements and exceeds the method of vector PEGylation. Conventional, amine-directed vector PEGylation does not allow for site-specific attachment of shielding or targeting moieties, whereas the geneti-chemical coupling technology presented here can be used to precisely attach moieties to adenovirus vector surfaces at defined positions and in defined copy numbers. Therefore, it is suitable for both shielding and targeting of adenovirus vector particles.
Of note, this protocol can also be used for a conventional amine-directed PEGylation of Ad vectors mentioned above. In that case, the reducing reagent TCEP can be omitted from the buffers. For stoichiometric calculations, the number of accessible amine groups per particle can be considered to be 18,000, and typical molar excesses of amine-reactive coupling moieties are 20-100 fold.
The authors have nothing to disclose.
Vector purification and chemical modification | |||
Argon gas | Air liquide | local gas dealer | |
Liquid Nitrogen | Air liquide | local gas dealer | |
500 mL centrifuge tubes | Corning | 431123 | |
Stericup Express Plus 0.22 µm | Millipore | SCGPU02RE | |
Tris(2-carboxyethyl) phosphine (TCEP) | Sigma-Aldrich | C4706-10g | |
2 mL (3mL) Norm Ject (syringes) | Henke Sass Wolf | 4020.000V0 | |
Fine-Ject needles for single use (yellow 0.9 x 40 mm) | Henke Sass Wolf | 4710009040 | |
Caesium chloride 99.999% Ultra Quality | Roth | 8627.1 | |
Silica gel beads | Applichem | A4569.2500 | |
Methoxypolyethylene glycol maleimide – 750 (PEG mal-750) | Iris Biotech | store in silica gel beads at -80 °C | |
13.2 mL Ultra Clear Ultracentrifuge Tubes | Beckman Coulter | 344059 | only open in hood |
PD-10 size exclusion chromatography column | GE Healthcare | 17-0851-01 | store at 4 °C |
Hepes | AppliChem | A1069.1000 | |
SDS Ultrapure | AppliChem | A1112,0500 | |
Glycerol | AppliChem | A1123.1000 | |
Name | Company | Catalog Number | Comments |
Material for cell-culture | |||
DPBS | PAN Biotech | P04-36500 | |
DMEM | PAN Biotech | P04-03590 | |
Trypsin/EDTA | PAN Biotech | P10-0231SP | |
FBS Good | PAN Biotech | P40-37500 | |
Penicillin/Streptomycin | PAN Biotech | P06-07100 | |
Biosphere Filter Tips (various sizes) | Sarstedt | ||
Serological Pipettes (various sizes) | Sarstedt | ||
reaction tubes (various sizes) | Sarstedt | ||
TC plates 15cm | Sarstedt | 83.3903 | |
Name | Company | Catalog Number | Comments |
Material for silver staining protocol | |||
Methanol | J.T.Baker | 8045 | |
Ethanol absolute | AppliChem | 1613,2500PE | |
Acetic Acid | AppliChem | A0820,2500PE | |
Formaldehyde 37% | AppliChem | A0877,0250 | |
Ethanol absolute | AppliChem | A1613,2500PE | |
Sodium thiosulfate | AppliChem | 1,418,791,210 | |
Silver nitrate | AppliChem | A3944.0025 | |
Sodium carbonate | AppliChem | A3900,0500 | |
Name | Company | Catalog Number | Comments |
Special Lab Equipment | |||
Desiccator | Nalgene | 5311-0250 | |
Megafuge 40 | Heraeus | ||
Roter for Megafuge TX750 + Adapter andLlids for 500 mL tubes | Heraeus | ||
Water bath | Conventional | ||
Ultracentrifuge e.g. Optima XPN-80 | Beckman Coulter | ||
suitable Ultrazentrifuge Rotor e.g. SW41 | Beckman Coulter | ||
pH -Meter | Conventional | ||
Stand with clamps | Conventional | ||
Goose neck lamp | Conventional | ||
Over-head rotor | Conventional | ||
Thermal Block | Conventional | ||
Photometer (OD 260) | Conventional |