Developing an astrocyte-selective AAV-ADAMTS4 gene therapy to promote repair after spinal cord injury

Abstract

Spinal cord injury (SCI) affects an estimated 2.5 million people worldwide yet no effective regenerative therapies are available to victims. Astrocytes play a critical role in the secondary injury pathogenesis of SCI. Reactive astrocytes are key contributors to the deposition of chondroitin sulphate proteoglycans (CSPGs) after injury, which are potent inhibitors to axon regeneration and plasticity. A disintegrin and metalloproteinase with thrombospondin motifs-4 (ADAMTS4) is a human enzyme that catalyses the proteolysis of CSPG protein cores. Infusion of ADAMTS4 into the damaged spinal cord was previously shown to improve functional recovery after SCI; however, this therapy is limited in its enzyme form. Gene therapy is a method that allows for long-term expression of therapeutic molecules. Adeno-associated viral (AAV) vector gene therapy, in particular, has emerged as the vector of choice for safe, robust and long-term transgene expression in the central nervous system (CNS). Since astrocytes play an important role in SCI pathology, the first objective of this thesis involved the creation of an AAV vector capable of producing robust transgene expression selectively in spinal cord astrocytes. An expression cassette containing a destabilised yellow fluorescent reporter (dYFP) reporter transgene under the control of a GFAP promoter was packaged into the astrocyte tropic AAV serotypes: AAV5, AAV9, and AAVRec2. In cell culture and slice culture experiments, AAV5 was superior in terms of transduction efficiency, transgene expression, and astrocyte selectivity. Due to the packaging constraints of AAV vectors, an additional AAV5 vector was created that contained the shorter GFAP variant GfaABC1D, which equalled the qualities observed by the full-sized promoter. In a clinically relevant rodent model of spinal cord contusion injury, AAV5 vectors with either promoter elicited robust, widespread, astrocyte-selective transgene expression without negatively impacting behavioural scores. An expression cassette containing the ADAMTS4 gene under the control of the GfaABC1D promoter was subsequently packaged into an AAV5 vector for astrocyte-mediated degradation of CSPGs in vitro and in vivo. Sustained expression of ADAMTS4 expression was achieved in vivo, which led to widespread degradation of CSPGs. AAV-ADAMTS4 resulted in significantly decreased lesion size, increased neuroplasticity and improved functional recovery after moderate contusive SCI. Whilst histologically, the effects of AAV-ADAMTS4 appeared to be large, improvements to motor functions were only modest. The last objective included hindlimb-specific exercise rehabilitation as a method to drive and strengthen the formation of beneficial connections whilst pruning aberrant connections. Indeed, the combination of hindlimb rehabilitation with AAV-ADAMTS led to enhanced functional recovery after SCI. Thus, widespread and long-term degradation of CSPGs through AAV-ADAMTS4 gene therapy, from a single administration, represents a promising candidate for clinical translation.