Investigation and manipulation of adenovirus interactions with host proteins

Adenoviruses are the most commonly used vectors for clinical gene therapy applications, accounting for 24% of all clinical trials to date, the majority of which are based on Ad serotype 5 (Ad5). However, the high prevalence of neutralising antibodies and a range of “off target” interactions result in liver sequestration, hepatic transduction and decreased circulation times. Such interactions include Kupffer cell uptake and binding to blood components such as erythrocytes, platelets, complement and coagulation factors. Recent studies have shown that hepatocyte transduction by Ad5 is mediated by a high-affinity interaction between coagulation factor X (FX) and the Ad5 major capsid protein hexon, with FX bridging the virus to heparan sulphate proteoglycans (HSPGs) on the cell surface. This thesis has focused on gaining a greater understanding of the Ad5:FX pathway and potential strategies for its manipulation. FX, a key component of the blood coagulation system, is a zymogen of a vitamin K-dependent serine protease that is primarily synthesised in the liver and circulates in the bloodstream at 8-10 μg/ml. It is composed of a light chain consisting of a domain rich in γ-carboxylated glutamic acid (Gla) residues, two epidermal growth factor-like domains and a serine protease (SP) heavy chain. The Gla domain of FX binds to the virion by docking in the cup formed by each hexon trimer, whilst the SP domain tethers the Ad5:FX complex to the hepatocyte surface through binding HSPGs. Previously, it was demonstrated that pharmacological blockade of the heparin-binding proexosite (HBPE) in the SP domain prevents FX-mediated cell binding. Here, the specific residues of FX which mediate Ad5 attachment to HSPGs were identified. Employing mutagenesis techniques each of the seven basic residues R93, K96, R125, R165, K169, K236 and R240 that were previously shown to bind heparin, were converted to alanine. This mutated FX was termed “SP mutant”. Stable cell lines were generated to constitutively produce the wild-type and SP mutant rFX protein in the presence of vitamin K. The conditioned media was affinity purified using a FX specific mouse monoclonal antibody 4G3 coupled to sepharose. The rFX proteins were quantified by ELISA, had the predicted molecular weight of 59 kDa and were biologically active, as shown by conversion to FXa in the presence of tissue factor and FVIIa. Surface plasmon resonance (SPR) analysis demonstrated the SP mutations had no effect on FX-specific binding to the Ad5 hexon. However the proexosite mutations ablated FX-mediated Ad5 cell surface binding, internalisation, cytosolic transport and gene transfer as shown by confocal microscopy, qPCR and quantification of transgene expression. Assessing the involvement of rFX with single (R125A) and double (R93A_K96A, R165A_K169A and K236A_R240A) point mutations in the SP domain, indicated the residues exhibit different levels of contribution to Ad5:FX complex binding to HSPGs. The seven SP mutations also inhibited FX-mediated Ad5 binding to mouse liver sections ex vivo. Taken together, this study uncovered that basic residues within the HBPE of FX have a fundamental role in Ad5:FX complex engagement with HSPGs at the surface of target cells. This study contributes to the existing knowledge of the FX-mediated Ad5 transduction pathway. Whilst the classical in vitro CAR-mediated Ad5 infection mechanism has been extensively studied, the post-binding events governing FX-mediated Ad5 intracellular transport and gene expression have not been fully characterised. This study employed a panel of small molecule inhibitors of cellular kinases in vitro to investigate cellular and signalling events occurring during FX-mediated Ad5 infection. Blockade of protein kinase A, p38 mitogen-activated protein kinase and phosphatidylinositol 3-kinase significantly hindered efficient Ad5 intracellular trafficking and colocalisation with the microtubule organising centres (MTOC), as shown by confocal microscopy, indicating their fundamental involvement in the pathway. Screening a library of 80 diverse kinase inhibitors for effects on FX-mediated gene transfer, highlighted the compound ER-27319 had the ability to prevent Ad5 transduction in vitro. Previous work reported that ER-27319 acts by binding to the immunoreceptor tyrosine based activation motif (ITAM) of the FcεRI receptor gamma subunit in mast cells to prevent spleen tyrosine kinase (Syk) activation. Here, this compound had no effect on FX-mediated cell binding but substantially disrupted intracellular transport at 3 h in the absence of toxicity. It was postulated that this effect may be due to ER-27319 binding to a viral or cellular ITAM-containing protein involved in viral trafficking. Sequence analysis of the Ad capsid proteome for ITAM-like motifs ((D/E)-x-x-Y-x-x-(L/I)-(xn=6-8)-Y-x-x-(L/I)) identified two motifs on the hexon. However neither followed that reported for the FcεRI gamma subunit, instead of the conventional 6-8 amino acid residues between the two Y-x-x-I/L, the hexon ITAM-like sequences expressed 17 or 22 amino acids. Alternatively the ITAM-containing cellular proteins, ezrin, radixin and moesin (ERM) were investigated. The ERM family are key regulators of the cell cortex, capable of interacting with both the plasma membrane and filamentous actin. However, in the time frame imposed by this study this hypothesis could not be studied in depth, but warrants further research to investigate whether ERM proteins have a novel role in FX-mediated Ad5 intracellular trafficking. A wide range of approaches have been investigated to detarget Ad5 from the liver. In this thesis, a pharmacological strategy to preclude FX-mediated liver gene transfer was implemented. A high throughput screening platform was developed to identify a novel small molecule(s) to manipulate the Ad5:FX infection pathway. In addition to the value of such an agent in the gene therapy setting, it may also have potential to treat life-threatening disseminated Ad infections in immunocompromised individuals. Using a fluorescence and cell-based in vitro high throughput assay 10,240 small molecules were screened using the Pharmacological Diversity Drug-like Set library. Initial screening identified 288 compounds that reduced FX-mediated Ad5 gene transfer by > 75% without causing toxicity. Upon further analysis, three compounds, T5424837, T5550585 and T5660138 were identified as consistently ablating Ad5 transduction both in the absence and presence of FX and all had IC50 values < 5.5 μM. These compounds did not directly interfere with Ad5 binding to FX, instead they primarily caused a post-binding stage block of the Ad infection pathway and all affected optimal virus trafficking to the MTOC, as demonstrated by SPR, flow cytometry and confocal microscopy. The candidate molecules have common structural features and fall into the “one pharmacophore” model. Focused mini-libraries were generated relating to these molecules and structure-activity relationship analysis was performed. In vitro screening of the analogues revealed novel hits with similar or improved activity, thereby further validating the initial hits and pharmacophore model. Six compounds, T5550585, its analogue T5572402, T5660138, its analogue T5660136, T5424837 and its analogue T5677956 were tested in vivo. 10 μM T5660138 substantially reduced Ad5 liver accumulation 48 h post-injection and, in addition to its closely related analogue T5660136, significantly reduced transgene expression at 48 h post-intravenous administration of a high viral dose (1 x 1011 vp/mouse). Therefore, this study identifies novel small molecule inhibitors of circulating Ad5 infection. Through investigation and manipulation of Ad5 interactions with host proteins the work presented here, increases the understanding of the key in vivo Ad5:FX tropism determining pathway. In summary, in this thesis the mechanism of FX-mediated Ad5 complex binding to hepatocytes was dissected and potent inhibitors of this important Ad5 infectivity pathway both in vitro and in vivo were identified. This data may contribute to the optimisation of Ad vectors for gene therapy applications and potentially the advancement of anti-adenoviral drug development