Gene transfer directly into neurons, using specific virus vectors, has potential for developing gene therapy treatments for specific neurological diseases and for studying neuronal physiology. However, due to the heterogeneous cellular composition of the brain, neuronal subtype-specific expression is required for many potential uses of neural gene transfer. The two predominate approaches are to target gene transfer to a specific cell type using a modified vector particle or to use a cell type-specific promoter to control expression [1–6]. A higher level of cell type-specific expression may be achieved by using these two complementary approaches together.
Helper virus-free HSV-1 vectors are attractive; they efficiently transduce neurons, have a large capacity, and cause minimal cytotoxicity [7–9]. The HSV-1 particle is composed of four components: i) The ~152 kb double stranded DNA genome is contained within ii) an icosahedral protein capsid, that is surrounded by iii) the tegument, a layer of proteins, and enclosed within iv) the envelope, a lipid bilayer containing 10 virally-encoded glycoproteins . HSV-1 infection occurrs in two stages . The initial binding to the cell surface is mediated by specific domains on glycoprotein C (gC) and gB that represent binding sites for glycosaminoglycans, principally heparin sulfate, present on cell surface proteoglycans [12–15]. Entry requires the subsequent binding of gD to a specific receptor. gD receptors include nectin-1 or nectin-2 of the immunoglobulin superfamily; the herpesvirus entry mediator (HVEM), which is a member of the tumor necrosis receptor family; and sites in heparin sulfate produced by specific isoforms of 3-O-sulfotransferases . Entry occurs by fusion of the cell membrane and the viral envelope, and requires gB, gD, gH, and gL.
Targeted gene transfer with HSV-1 vectors was achieved by modifying gC to remove the heparin binding domain, and addition a binding site for a specific cell surface protein to either gC or gD. The initial study  reported a recombinant virus containing a chimeric gC – erythropoietin (EPO) that supported enhanced binding to cells that contained EPO receptors. Analogous designs used gD – IL13, or gD – urokinase plaminogen activator, or gD – single-chain anti-EGF receptor antibody chimeric proteins to target infection to tumor cells containing the cognate receptor [17–19]. Another study  used a HSV-1 plasmid (amplicon) vector expressing gC – His tag, and packaging with a helper virus deleted in gC; this strategy targeted infection to a cell line containing a pseudo-His-tag receptor. All of these reports [16–20] used specific HSV-1 viruses that grow productively and kill infected cells, confounding potential gene transfer studies.
We reported targeted gene transfer to a specific type of neuron in the brain . Helper virus-free packaging was performed using pHSVlac  and either gC – glial cell line-derived neurotrophic factor (GDNF) or gC – brain-derived neurotrophic factor (BDNF) chimeric protein. Delivery of either vector stock into the midbrain supported 2.2 to 5.0-fold targeted gene transfer to nigrostriatal neurons , which contain both the GDNF receptor α-1 (GFRα-1 [23, 24]) and the high-affinity BDNF receptor, TrkB[25, 26]. Specifically, 75 to 80% of the transduced cells were nigrostriatal neurons, compared to 15–30% using untargeted gene transfer; and for the untargeted expression, 5 to 10% was in each of GABAergic neurons and glial fibrillary acidic protein-containing cells . At times shortly after gene transfer, the HSV-1 immediate early (IE) 4/5 promoter in pHSVlac  supports expression in most cell types, which enabled an accurate assessment of the types of transduced cells. However, pHSVlac does not support significant levels of long-term expression in most neural cell types, and these rats were sacrificed at 4 days after gene transfer.
Specific promoters support long-term expression in neurons, or specific types of neurons, from HSV-1 vectors. To obtain long-term, neuronal-specific expression, we constructed a chimeric promoter containing a neurofilament heavy gene (NF-H) promoter, an upstream enhancer from the tyrosine hydroxylase (TH) promoter, and a β-globin insulator (INS-TH-NFH promoter ). Vectors containing the INS-TH-NFH promoter supported expression for 7, 8, or 14 months [27–29], and at 6 months, ~11,400 striatal neurons contained recombinant gene products (with 3 injection sites for gene transfer, ). Alternatively, vectors containing the TH promoter supported expression for at least 2 months in midbrain dopaminergic neurons, including nigrostriatal neurons (helper virus system [6, 30]; helper virus-free system ). Of note, 40 to 60% of the transduced cells were nigrostriatal neurons [6, 31], compared to 5% using pHSVlac . Targeted gene transfer or use of the TH promoter each supported only partial nigrostriatal-specific expression, raising the possibility that further improvements in nigrostriatal-specific expression might be obtained by combining these complementary approaches.
Both to improve nigrostriatal neuron-specific expression, and to establish that targeted gene transfer can be followed by long-term expression, we performed targeted gene transfer using vectors that contain either the TH or INS-TH-NFH promoter. The combination of targeted gene transfer and a neuronal-specific promoter improved nigrostriatal neuron-specific expression to 83 to 93%. Of note, the levels of expression at 1 month were similar to those previously observed using untargeted gene transfer with each promoter.