A desired gene therapy protocol would be to precisely deliver a gene of interest to specific cells in vivo using a targeted gene delivery vehicle for administration. As an alternative to pseudotyping with existing envelopes, various modifications have been employed to improve the targeting of vectors to specific cell types and regulate transgene expression.

Cell type-specific lentiviral vectors can be generated through surface engineering, by incorporating cell type–specific ligands or antibodies on a mutated viral envelope.

Several attempts have been made to alter the receptor-recognition attachment function in the envelope glycoprotein, without affecting membrane fusion [64]. Such modifications ablate the native tropism of the vector, so it is no longer able to recognize and bind to its cognate receptor, and redirect its tropism to the corresponding cellular receptors recognized by the ligand or antibody. Targeting systems available at the moment involve, among others, antibody specific binding methods based on (1) molecular events of receptor binding and endocytosis followed by fusogen-induced pH-dependent membrane fusion [65–67] or (2) active membrane fusion [68–70].

In 2006 Yang et al. have proposed an efficient targeting method. The method involves the incorporation of an antibody and a fusogenic protein as two distinct molecules on the lentiviral surface, separating the recognition from the fusion function by providing them as separate molecules. The fusogen is modified so it lacks the ability to bind to cognate receptors, but retains the ability to trigger pH-dependent membrane fusion, thus is binding deficient but fusion competent [65, 67, 71, 72]. The specificity of this LV is solely determined by the antibody chosen to recognize a specific antigen on the surface of the target cells. Recently these targeted LVs have been optimized by using single chain antibodies (scFV) genetically fused to either Sindbis virus or Measles virus (MV) hemagglutinin envelope [69, 73]. Modifications applied to optimize this method, aimed to overcome limitations such as receptor internalization and endosomal release [66], which are necessary when vectors are pseudotyped with certain envelopes.

An additional targeting method involves coating of the viral surface with polymers such as polyethylene glycol, poly-[N-(2-hydroxypropyl) methacrylamide], or biodegradable alginate microparticles [74–76]. All these methods are based on the same principle: ablating native tropism and retargeting through surface display of ligands (peptide, proteins, or antibodies). One disadvantage of the method is that the two-component nature of these bispecific molecules adds complications in vector production and in maintaining stable batch-to-batch homogeneity, as the number of plasmids required for production of these molecules targeted vectors increases (usually 5–6 plasmid co-transfection is required). Other approaches involve the use of a ligand protein or antibody as a bridge to attach the virus to specific cells [67, 71, 77, 78]. However, this approach requires endocytosis for the pH-dependent fusion in addition to the fact, that once the envelope protein is connected to the one end of the bridge molecule, it fuses inefficiently. Additionally to targeting through surface engineering, tissue specific promoters [79, 80] and or tissue-specific regulatory elements [81] have been used and might also optimize the targeting strategy restricting transduction to the tissue of interest. The use of mammalian promoters can result in long-term physiological levels of transgene expression [8]. Glial fibrillary acidic protein (GFAP), synapsin 1 [82–84], neuronal specific enolase (NSE) [85], tyrosine hydroxylase/neurofilament [86], and the prion promoters [87] are some examples of powerful neuronal promoters which have been used in adenoviral and lentiviral systems. Cytomegalovirus (CMV), platelet growth factor-β, and CMV/chicken β-actin promoters have also been tested in AAV and lentiviral vector systems for CNS targeting, where specificity achieved from each can vary between different CNS regions [84].

Alternatively, incorporation of drug-inducible systems (e.g., tetracycline, ecdysone, mifepristone [88, 89]) in lentiviral vectors can better control the levels and timing of gene expression [90]. Such lentiviral systems have been generated and used to mediate regulated gene delivery in Parkinson’s disease [90, 91]. Additionally, it is also possible to achieve cell type specific expression through microRNA (miR) regulation [92, 93]. When shRNA molecules are incorporated into a lentiviral vector, they are able to inhibit mRNA by RNA interference [94]. Moreover, control might be exerted by the inclusion of matrix scaffolding domains which place sequences in transcriptionally favored regions of the nucleus, and insulating elements which can protect promoters from the influences of vector and genomic sequences [95].

As described, transgene expression can be achieved using both integrating and non-integrating LVs. There is, however, a low risk of activation of proto-oncogenes or dedifferentiation following integration, as most transduced neural cells are terminally differentiated [4]. Concerns associated with random integration and insertional oncogenesis can be alleviated with the use of integration deficient lentiviral vectors (IDLV) that are deficient in integrase activity, the enzymatic activity required to catalyze integration into the host cell genome [96]. IDLV can be used for short-term expression of a gene of interest in dividing cells. As far as non-dividing target cells are concerned, specifically brain, liver, or retina, IDLV can induce relatively stable transgene expression [97–99]. Particularly, in short-term studies, it has been demonstrated that use of IDLV resulted in successful transduction of the rat striatum and retrograde transport to the dopaminergic neurons of the substantia nigra [99].

Moreover, IDLV can be used in conjunction with zinc-finger nuclease (ZFNs) hybrid technology to achieve site-specific gene editing/correction or addition at the targeted chromosomal loci [100] in which integration is believed to be safe, which could minimize or overcome the risks associated with random integration. Despite the fact that this technology is promising, requires further optimization in terms of both efficiency and safety. Nonetheless, the necessity for introduction of double-stranded DNA breaks (DSB) can be an issue as there is a potential of off-target effects where DSBs can be introduced in loci different from the desired one. This could result in accumulation of ZNFs to these sites and further contribute to ZFN cytotoxicity [101].

Disorders of the nervous system still remain a major medical challenge, in matters of management and care. It is crucial to investigate and understand the disease mechanisms, identify therapeutic targets, and design and develop strategies for therapy. As already mentioned above, the ability to transfer genes to neuronal cells utilizing viral vectors and specifically lentiviral vectors, allows delivery of therapeutic genes to the nervous system. Neural architecture itself is promising for designing therapy. The morphology of each neuron and the interconnected networks they create, make the nervous system particularly appropriate for vector-mediated gene transfer; thus delivery at one site can lead to action at a distal site, if retrograde transport is possible [4]. Gene therapy approaches include delivery of trophic factors to protect neurons from damage, axogenic or regeneration molecules to restore neural function after injury, siRNA targeted to defective intracellular proteins and genes, or restoring missing proteins by gene replacement. Protocols of gene delivery for neurodegenerative disorders often requires invasive surgery and therefore repeat treatments are impractical; long-term, stable expression at therapeutic levels is essential. The therapeutic potential of viral vector gene therapy has been illustrated in many studies using animal models for human disease. The first use of lentiviral vectors for therapeutic gene expression (Bcl-xL) in the CNS was reported in 1998 [102]. Since then, effective long-term treatment has been reported in many animal models of neurological disorders, such as Parkinson’s disease (PD), Alzheimer’s disease, Huntington’s disease (HD), motor neuron diseases, LSD, and spinal cord injury (reviewed in [4, 43]).

LVs have also been efficiently used to develop transgenic disease models in rats and primates, which is important for behavioral studies. This technology allows over expression of the disease causing transgenes, which is essential for developing clear disease pathology and compared to standard transgenic techniques enables targeting of well-defined brain regions. Models of HD have been developed by using LVs overexpressing parts of the mutated huntingtin gene in the striatum of rats [103–105]. Indeed it was possible to induce pathological changes resembling those of HD. Similar to HD, lentiviral vectors have also been used to create novel animal models for PD by overexpressing α-synuclein in the dopaminergic cells of rats and primates [106–108]. On this front, therapeutic strategies have also been tested beyond the dopamine replacement and neurotrophic support approaches presented before [109]. Localized disease models can also be developed with vectors expressing siRNA using regulatable promoters or by applications of the Cre-lox recombination systems [110, 111].

The research applications of LVs presented above are promising for revealing information critical for our understanding of the nervous system in health and disease and for developing strategies to deliver therapy for neurological disorders. Table 2 summarizes the various gene therapy approaches for neurodegenerative diseases.