A new strategy to achieve targeted adenoviral gene delivery in vivo: mechanistic and therapeutic insights

Gene therapy holds tremendous promise for the treatment of a wide range of diseases that lack a suitable standard drug therapy. To deliver a gene as a therapeutic entity, a 'vector' needs to be used - a biological system through which the gene can be delivered to the correct cells in the body to provide the therapeutic end point. The choice of 'gene delivery vector' for each application is critical in determining the success of the approach and these can generally be divided into viral and non-viral vectors. Non-viral vectors are generally of lower efficiency for delivery of genes into cells as they lack specific mechanisms to interact with the cell surface, although recent advances have improved their efficiency. Viral vectors, on the other hand, have evolved to infect cells through specific interactions at the cell surface and usually have a much higher efficiency rates for gene delivery. Of course, viral vectors have additional safety issues that require careful attention and pre-clinical work to minimise the potential side effects that may occur in response to virus delivery. In fact, adenovirus p53 is now licensed as a gene therapy product and additional results in clinical trials look very encouraging for future gene therapy applications. However, viral delivery has also been problematic, as evidenced by the development of leukaemia in a percentage of children being treated for immunodeficiency disease on a clinical trial in France. Further, adenovirus-mediated gene delivery (at high dose via the bloodstream) resulted in the tragic death of Jesse Gelsinger. The adenovirus vector family holds tremendous potential for gene therapy for many diseases including cancer, cardiovascular disease and liver-directed gene therapy but we need to increase our understanding of virus:host cell interactions to be able to use this knowledge to improve gene delivery and safety. In relation to adenoviruses they can be used effectively for local gene delivery (where direct access to the target tissue is achievable) or via the bloodstream. When the virus is injected into the blood it immediately contacts blood plasma proteins, cells and the vascular endothelium, however it has a tremendous affinity for transducing the liver, in the relatively complete absence of transduction of other organs in the body. The mechanism(s) by which the virus targets the liver has been something of a mystery until recently. In a recent publication we showed that the virus interacts with blood coagulation factors (coagulation factors FVII, FIX, FX and protein C that all have a very similar structure) and that the complex formed then targets receptors on liver cells called heparan sulphate proteoglycans and LDL receptor related protein (LRP). The aim of the current proposal is to understand in much more detail about this interaction and how it impacts on the viruses ability to bind to liver cells and the effects on toxicity (for example, induction of cytokines). We will 'map' the exact locations on the virus and on the coagulation factors that lead to the binding (we have already narrowed this down considerably). This will be important in the subsequent design of adenovirus serotype 5 vectors that are devoid of this coagulation factor binding - we believe such engineering towards future medical applications will have a significant impact on gene delivery mediated by adenovirus with the hope that we can engineer adenoviruses that target other tissues and organs. We will address this potential by specific assessment of targeting adenovirus to pulmonary and cardiac endothelium in vitro and in vivo in the presence of blockade of adenovirus:coagulation factor binding. If successful, this novel and achievable approach would have signficant appeal and utility for in vivo gene delivery in the future.

The utility of adenoviral vectors based on the commonly used serotype 5 for in vivo gene delivery and therapy via the intravascular route is limited by substantial liver targeting and compromised by resulting immune responses to the virus and toxicity at high doses. In previous work funded by the BBSRC (BB/C000048/1) we have shown the ability of genetic engineering approaches to achieve targeted gene delivery of adenovirus vectors in vitro through a combination of capsid mutations to ablate virus binding to native receptors (such as CAR) and retargeting through incorporation of ligands into fiber structure (in the exposed HI loop). We have been very successful in developing in vivo phage display to identify peptide ligands that target individual tissues following intravenous injection of phage libraries. Insertion of these peptides into the HI loop of the virus allows retargeting in vitro but not in vivo. This lack of in vitro-to-in vivo translation of adenovirus type 5 targeting was perplexing - we therefore investigated why. In a substantial series of experiments we have identifed the elusive and critical pathway that dictates adenovirus liver targeting. This involves direct binding of the adenovirus capsid to coagulation factors with a gla-EGF-EGF-SP domain structure (i.e. Factors VII, FIX, FX and PC). Factor X is the predominate coagulation factor involved in this process and binding results in delivery of the complex to heparan sulphate proteoglycans in the liver. Here, we will detail the binding sites important for this interaction on both FX and adenovirus. We will use our in vitro and in vivo model systems to assess the ability to retarget adenovirus to pulmonary and cardiac endothelium using antibody and peptide-targeting approaches. This is therefore a novel approach to targeting adenovirus 5 and is a logical continuation to my previous grant application.