Directed Evolution of New Viruses for Therapeutic Gene Delivery

Strong basic and translational efforts in the gene therapy field have culminated in successes in an increasing number of human trials, including for Leber’s congenital amaurosis, severe combined immunodeficiencies, hemophila, and X-linked adrenoleukodystrophy.  These studies have established that certain viral vectors are capable of efficient, safe, and therapeutic gene delivery to numerous target cells and tissues.  Vectors based on adeno-associated virus and lentivirus thus have considerable promise; however, numerous challenges can limit the applicability of these biological nanoparticles, such as anti-vector humoral and cellular immunity, low transduction of some therapeutically relevant cells in vitro or in vivo, difficulty in overcoming cellular and physical barriers within complex tissue structures, and an inability for targeted delivery to specific cells.  These challenges are not surprising, as nature did not evolve viruses for use as human therapeutics.  Rational design has made progress in engineering viral variants to address several such shortcomings; however, in many situations there is insufficient mechanistic knowledge of underlying virus structure-function relationships to effectively empower rational design with the capacity to improve a vector. As an alternative, directed evolution has been emerging as a strategy to engineer novel viral variants that meet specific biomedical needs.

We have been developing directed evolution and library selection approaches to address a number of problems with adeno-associated viral (AAV) and retro/lentiviral vectors for a decade.  Directed evolution – which emulates the natural evolution process – is the iterative genetic diversification of components of a viral genome and functional selection for desired properties.  In the case of adeno-associated virus, genetic diversification has included the random diversification of peptide sequences at defined locations in the viral capsid structure, random point mutagenesis of the cap gene, and recombination of cap genes from a number of parental serotypes to create random chimeras.  By iterative selection and further diversification of such libraries, the receptor specificity, tropism, biodistribution, and antigenicity of the viral particles can be fundamentally shifted.  As one example, it would be desirable to develop non-invasive means to deliver genes to the retina; however, natural AAV variants can only transduce cell types involved in retinal degeneration (photoreceptors and retinal pigment epithelium) with an invasive subretinal injection, which in many cases can further damage an already degenerating tissue.  We evolved AAV for the capacity to mediate high efficiency gene delivery to photoreceptors upon a simple injection into the vitreous humor of the eye, and this ability to traverse a formidable tissue barrier may have implications for treating retinal degenerations that cause blindness.  We have also recently applied directed evolution to create vector variants that evade human neutralizing antibodies or mediate gene delivery and gene targeting in stem cells, attractive targets for basic and applied studies.

In summary, natural variants of viruses are suitable for a number of therapeutic applications; however, in many situations these “off the shelf” products of nature do not meet urgent biomedical needs.  Continuing the process of evolution, but changing the trajectory towards functions and properties that are useful in medical rather than natural settings, is therefore a promising approach to create designer viral nanoparticles to meet specific human therapeutic needs.