Gene therapy has been around for nearly a decade as a modern therapeutics [1]. However, gene therapy for disorders of the nervous system still remains a major medical challenge. Gene therapy for the brain diseases is particularly challenging due to the post-mitotic nature of neuronal cells and the restricted accessibility of the brain itself. Lentiviral vectors and adeno-associated viral vectors (AAV) are increasingly regarded as the two most useful gene therapy vectors for the CNS. Viral vectors based on lentiviruses are particularly attractive vehicles, routinely used in developing gene-based therapies to treat neurological diseases [2]. Both vectors have been successfully used to express various foreign genes in a variety of brain regions and neuronal cell types [3,4,5,6].

There are several factors that should be concerned in using viral vectors, such as efficiency of gene delivery, targeting of expression, duration of expression, and safety [1]. By evaluating a variety of cell lines, it has been revealed that virus efficiency is cell-type dependent, a fact that has significant implications for the preparation of standard virus stocks [7]. The efficiency of virus is different by cell type or region [7]. In order to identify the property of each cell types, many researchers have investigated the tropism of viruses. Moreover, improving the efficiency of viral gene delivery has become increasingly important to increase the success of clinical trials [8].

Many researchers have further advanced the tropism of AAV through pseudotyping, variant serotypes and hybrid capsid from variant serotypes, which also alter viral tropism [9]. AAV serotype 1 through 6 were tested in the substantia nigra of the African green monkey brain [9]. In that study, the authors found that AAV serotype 5 displayed the most efficient transduction of neurons in this brain region. Recently, AAV-DJ vector was generated as one of the recombinogenic hybrid vector from gene shuffling of eight serotypes and has been widely used across a broad range of cell types because it has a higher transduction efficiency both in vitro and in vivo [10,11]. Even though AAV serotype 5 displayed the most efficient transduction of neurons in the African green monkey brain region [9], AAV-DJ has not been tested in the primate. Therefore, in the present study we tested the AAV-DJ in putamen of cynomolgus monkey for the first time. Furthermore, comparison of AAV and lentivirus or analysis of tropism of lentivirus especially in the monkey brain has not been conducted yet. Monkeys are becoming extremely valuable in the field of neuroscientific research. Therefore, in the current study, we set out to investigate the cellular tropism of lentivirus and AAV in cynomolgus monkey.

Another important issue in using virus for brain research is to establish physiological levels of virally delivered gene expression and therefore ways of regulating the expression [1]. Cell body and processes of astrocytes are usually hypertrophied at the sites of injury or surround the lesions [12,13,14]. The hypertrophied morphological change is generally called reactiveness of astrocytes. For example, a previous study examined the harmful effect of astrocyte-specific overexpression of GFP by viral infection of AAV-GFAP (Glial fibril acidic protein, marker for astrocyte)-GFP in mouse brain [15]. The authors found a profound neuroinflammation with infected astrocytes turned from normal astrocytes into reactive astrocytes as evidenced by presence of morphological changes and increased GFAP and vimentin level. The consequence of the appearance of reactive astrocytes by AAV was that in nearby neurons the amplitude of inhibitory synaptic activities was decreased. Therefore, it is important to find a virus type and expression level that do not cause reactivity in astrocytes. The degree of reactivity in astrocytes by viral infection in nonhuman primate is unknown. In this study, we examined which viral vector causes more reactivity in astrocytes in the brain of cynomolgus monkeys.

To investigate the cellular tropism, lentivirus and AAV were injected into putamen of 6~7 year-old cynomolgus monkeys with Hamilton syringe with 22 gauge needle (Fig. 1). The animals survived for 30 days in good health with postoperative clinical evaluation showing no evidence of behavioral abnormalities, loss of appetite, weight loss, or other abnormalities. 1 month after injection, monkeys were sacrificed and fixed in 4% paraformaldehyde solution. For sectioning frozen tissues, brains were incubated for 3 days in sucrose solution for dehydration (Fig.1a). Tissues were sliced with thickness of 50 µm, and stained with antibodies against GFAP, or NeuN (neuronal nuclei, marker for neuron) and DAPI. Colocalization with GFAP or NeuN positive cells and GFP expressing cells were counted (Fig. 1b).

To identify which cells are infected with lentivirus, 25 µl Lenti-GFP diluted with saline was injected into putamen of cynomolgus monkey unilaterally. Brain sections were stained with GFAP and NeuN antibodies and cells with positive staining were counted (Fig. 2a and 2b). Transduction efficiency was measured by counting GFP expressing cells with DAPI. Then we counted how many GFP positive cells were colocalized with GFAP or NeuN. We found that 60% (57/99) of GFP expressing cells were co-localized with GFAP, whereas only 10% (11/99) of GFP expressing cells were NeuN positive (n=99: total number of cells) (Fig. 2c). These results indicated that lentivirus had tropism for GFAP positive astrocytes in putamen of cynomolgus monkey

To identify which cells were infected with AAV, 25 µl AAV-GFP diluted with saline was injected into putamen of cynomolgus monkey unilaterally. As it is described already, AAV-DJ has been widely used for the vehicle of the gene delivery because it has a higher transduction efficiency [11]. Therefore, we packaged with AAV-DJ. After being sacrificed, tissues were stained with GFAP and NeuN antibodies and counted (Fig. 3a and 3b). As shown in Fig. 3a, we found that AAV-DJ infected numerous cells. We counted GFP positive cells with DAPI and assessed how many cells are co-localized with GFAP and NeuN. We found that 80% (168/206) of GFP expressing cells were co-localized with GFAP, whereas only 7% (13/206) of GFP positive cells were NeuN positive (n=206: total number of cells) (Fig. 3c). These results indicated that AAV has tropism toward neurons in putamen of cynomolgus monkey brain.

Astrocytes nearby injury site or lesion are morphologically transformed and hypertrophied. It is possible that these hypertrophied reactive astrocytes may not function normally. To assess the degree of reactiveness of infected astrocytes, we performed the Sholl analysis [18] from the infected and uninfected GFAP positive cells (Fig. 4a and 4b). The results of the analysis showed that both the sum of intersections and ramification index, which are indicators of branching of processes, slightly but not significantly increased in lentivirus-infected group (Fig. 4c and 4d), whereas both sum of intersections and ramification index increased significantly in AAV-infected group (Fig. 4e and 4f), compared to corresponding uninfected group. These results demonstrate that the reactiveness induced by lentivirus is less severe than that by AAV. In other words, the astrocytes infected with lentivirus are closer to normal condition than those with AAV.

From the results of current study, we conclude that lentivirus is well-suited for infecting astrocytes whereas AAV can be more appropriate for infecting neurons in cynomolgus monkey brain. In addition, we have demonstrated that AAV-DJ is an effective tool for gene delivery to the monkey brain. This conclusion is based on three observations: 1) AAV-DJ vector showed the high transduction effectivity even in the cynomolgus monkey brain. 2) We found a preferred cellular tropism of lentivirus for astrocytes and AAV for neurons, 3) infected astrocytes showed less hypertrophy by lentivirus than AAV. The reactivity of astrocytes by virus infection is not an uncommon phenomenon. However, in this study we found that infection of lentivirus causes minimal reactiveness in astrocytes. This is an important factor to consider as neuroinflammation is directly linked to reactivity of glial cells including astrocytes. Therefore, to minimize neuroinflammation during virus-mediated gene-delivery, lentivirus should be considered in monkey brain. The origin of the cellular tropism of certain viruses is not clear. Future studies are needed to determine the molecular mechanism of cellular tropism of various viruses and the cause of reactivity by viral infection in astrocytes.