Various regions of the brain have been successfully transduced by recombinant adeno-associated virus (rAAV) vectors with no detected toxicity. When using the cytomegalovirus immediate early (CMV) promoter, a gradual decline in the number of transduced cells has been described. In contrast, the use of cellular promoters such as the neuron-specific enolase promoter or hybrid promoters such as the chicken beta-actin/CMV promoter resulted in sustained transgene expression. The cellular tropism of rAAV-mediated gene transfer in the central nervous system (CNS) varies depending on the serotype used. Serotype 2 vectors preferentially transduce neurons whereas rAAV5 and rAAV1 transduce both neurons and glial cells. Recombinant AAV4-mediated gene transfer was inefficient in neurons and glial cells of the striatum (the only structure tested so far) but efficient in ependymal cells. No inflammatory response has been described following rAAV2 administration to the brain. In contrast, antibodies to AAV2 capsid and transgene product were elicited but no reduction of transgene expression was observed and readministration of vector without loss of efficiency was possible from 3 months after the first injection. Based on the success of pioneer work performed with marker genes, various strategies for therapeutic gene delivery were designed. These include enzyme replacement in lysosomal storage diseases, Canavan disease and Parkinson's disease; delivery of neuroprotective factors in Parkinson's disease, Huntington disease, Alzheimer's disease, amyotrophic lateral sclerosis, ischemia and spinal cord injury; as well as modulation of neurotransmission in epilepsy and Parkinson's disease. Several of these strategies have demonstrated promising results in relevant animal models. However, their implementation in the clinics will probably require a tight regulation and a specific targeting of therapeutic gene expression which still demands further developments of the vectors.

Recombinant AAV efficacy has been demonstrated in numerous gene therapy preclinical studies. As this vector is increasingly applied to human clinical trials, it is a priority to evaluate the risks of its use for workers involved in research and clinical trials as well as for the patients and their descendants. At high multiplicity of infection, wild-type AAV integrates into human chromosome 19 in approximately 60% of latently infected cell lines. However, it has been recently demonstrated that only approximately 1 out of 1000 infectious units can integrate. The mechanism of this site-specific integration involves AAV Rep proteins which are absent in vectors. Accordingly, recombinant AAV (rAAV) do not integrate site-specifically. Random integration of vector sequences has been demonstrated in established cell lines but only in some cases and at low frequency in primary cultures and in vivo. In contrast, episomal concatemers predominate.Therefore, the risks of insertional mutagenesis and activation of oncogenes are considered low. Biodistribution studies in non-human primates after intramuscular, intrabronchial, hepatic artery and subretinal administration showed low and transient levels of vector DNA in body fluids and distal organs. Analysis of patients body fluids revealed rAAV sequences in urine, saliva and serum at short-term. Transient shedding into the semen has been observed after delivery to the hepatic artery. However, motile germ cells seemed refractory to rAAV infection even when directly exposed to the viral particles, suggesting that the risk of insertion of new genetic material into the germ line is absent or extremely low. Risks related to viral capsid-induced inflammation also seem to be absent since immune response is restricted to generation of antibodies. In contrast, transgene products can elicit both cellular and humoral immune responses, depending on the nature of the expressed protein and of the route of vector administration. Finally, a correlation between early abortion as well as male infertility and the presence of wt AAV DNA in the genital tract has been suggested. Although no causal relationship has been established, this issue stresses the importance of using rAAV stocks devoid of contaminating replication-competent AAV. This review comprehensively examines virus integration, biodistribution, immune interactions, and other safety concerns regarding the wild-type AAV and recombinant AAV vectors.

Regulated gene delivery systems are usually made of two elements: an inducible promoter and a transactivator. In order to optimize gene delivery and regulation, a single viral vector ensuring adequate stoichiometry of the two elements is required. However, efficient regulation is hampered by interferences between the inducible promoter and (i) the promoter used to express the transactivator and/or (ii) promoter/enhancer elements present in the viral vector backbone. We describe a single AAV vector in which transcription of both the reverse tetracycline transactivator (rtTA) and the transgene is initiated from a bidirectional tetracycline-responsive promoter and terminated at bidirectional SV40 polyadenylation sites flanking both ITRs. Up to 50-fold induction of gene expression in human tumor cell lines and 100-fold in primary cultures of rat Schwann cells was demonstrated. In addition an 80-fold induction in vivo in the rat brain has been obtained. In vitro, the autoregulatory vector exhibits an induced expression level superior to that obtained using the constitutive CMV promoter. Although extinction of the transgene after removal of tetracycline was rapid (less than 3 days), inducibility after addition of tetracycline was slow (about 14 days). This kinetics is suitable for therapeutic gene expression in slowly progressive diseases while allowing rapid switch-off in case of undesirable effects. As compared to previously described autoregulatory tet-repressible (tetOFF) AAV vectors, the tet-inducible (tetON) vector prevents chronic antibiotic administration in the uninduced state.