395. Safety of In Utero Gene Delivery of Lentiviral Vectors

Top of pageAbstract Lentiviral vectors hold great promise for the treatment of chronic diseases. Vectors based on the equine infectious anemia virus EIAV have shown long-term stable correction in models for Parkinson's disease, spinal muscular atrophy, spinal injury, motor neuron disease and diabetes insipidus. They have also been used to create transgenic chickens and pigs where gene expression is stable over several generations. A notable feature in all of these models is the long term expression and the long term safety (see references below). We were concerned to learn of a study where a very early generation EIAV vector (SMART2 with the pONY3.1 packaging system) had been injected in utero in mice and the resulting pups had developed liver tumours (http://www.advisorybodies.doh.gov.uk/genetics/gtac/lentiviruses-1104.pdf). We compared the properties and vector genome structure with other vectors, in particular HIV vectors, which had not had any side effects or tumours reported. In general, transduction efficiencies reported in mice with HIV vectors are lower than with EIAV, possibly due to the known host cell restriction. One hypothesis was therefore that the EIAV vectors had a higher number of integrations per cell providing an increased opportunity for insertional mutagenesis. There is nothing unusual about the integration properties of EIAV vectors. The findings to date show a preference for active regions of the chromosome but no particular bias to a location or to positions relative to a transcription unit. Furthermore some tumours had only one or two integrations. At present we believe that a role for insertional mutagenesis alone seems unlikely. The SMART series of EIAV vectors contain an open reading frame in the WPRE that could in theory express a truncated form of the oncogenic X protein from the woodchuck hepatitis virus. This could potentially act as a weak oncogene (Kingsman et al 2005), particularly in the rapidly proliferating fetal tissues. Certain vector integrations could contribute to the promotion of the tumour and so it remains to be seen whether this truncated protein is either necessary and/or sufficient to cause tumours in this model. We will present data on the integration characteristics of EIAV in cell lines and in transduced livers and liver tumours from the in utero studies. Our data lead us to favour the hypothesis that the X-like protein acts as an oncogene in this model. Clinical EIAV vectors do not express this protein and lack all extraneous viral sequences. Further studies are required to test all of these hypotheses in order to inform the further development of these vectors for in utero gene therapy.