Eighteen C57BL/6 male mice (6–8 weeks old) were purchased from Harlan Laboratories (Barcelona, Spain). Animal handling was conducted in accordance with the European Council Directive 2010/63/UE, as well as in agreement with the “Policy on the Use of Animals in Neuroscience Research” issued by the Society for Neuroscience. The experimental design was approved by the Ethical Committee for Animal Testing of the University of Navarra (protocol Ref: 102-16). Animal handling was conducted in accordance with the European Council Directive 2010/63/UE, as well as in agreement with the “Policy on the Use of Animals in Neuroscience Research” issued by the Society for Neuroscience. The experimental design was approved by the Ethical Committee for Animal Testing of the University of Navarra (protocol Ref: 102-16). Anesthesia was induced by intraperitoneal injection of ketamine (100 mg/kg) and xylazine (10 mg/kg). The coordinates for targeting the striatum were 0.5 mm rostral, 2 mm lateral and 3.5 mm ventral from the bregma (Paxinos et al., 2001). All animals received two pressure injections: one of 2 μl of PBS/5% sucrose containing AAV vector on the left side (4 × 10⁹ vp), and a second of vehicle alone on the right side of the striatum. Injections were performed using a Hamilton syringe driven by a syringe pump at a flow rate of 0.2 μl/min. Following the injection, the needle was left in place for 2 min prior to being slowly retracted to avoid vector leakage from the injection tract. After surgery, animals were kept under constant monitoring with ad libitum access to food and water.

Human embryonic kidney fibroblast (HEK-293) cells, were purchased from the ATCC and were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% (v/v) heat-inactivated fetal bovine serum (FBS), penicillin (100 μg/ml), and streptomycin (100 U/ml) (all supplements were from Invitrogen, Pisley, Scotland, UK). Cells were maintained at 37°C in a humidified atmosphere of 5% CO₂.

cDNA encoding enhanced green fluorescent protein (eGFP) was isolated from the vector pBSKII-CMV-EGFP and inserted into the multiple cloning site (MCS) of an rAAV2 plasmid, which contained AAV2 inverted terminal repeats (ITR), to obtain rAAV2-eGFP. Upstream of the eGFP coding sequence different promoters were inserted: a constitutive hybrid promoter composed of the CMV immediate-early enhancer fused to chicken ß-actin promoter (CAG pr; Niwa et al., 1991), two reduced versions of the human glial fibrillar acidic protein (GFAP) promoter (hGFAP pr, 587 bp, containing the A, B, C₁, and D elements) (Lee et al., 2008) and hGFAPΔD (512 bp), in which the D sequence of was removed. This sequence was previously shown to play an important role in the functionality of the promoter (Besnard et al., 1991). Furthermore, using the structure of the human gfaABC₁D promoter and the sequence of the murine GFAP promoter, a reduced version of the murine gfaABC₁D promoter was constructed (581 bp, mGFAP pr). The proximal promoter of murine BM88 (88 bp; Papadodima et al., 2005) and the minimal promoter driving neuron-specific expression of the ß2 subunit of the nicotine acetylcholine receptor (CHNRB2 pr, 177 bp; Bessis et al., 1995), were used for neuron-specific targeting. Moreover, the human growth hormone (hGH) poly A signal and the ß-globin intron were cloned into the plasmid (Figure 1). Minipreps and maxipreps were prepared using commercial kits according to the manufacturer's instructions (Macherey-Nagel, Düren, Germany). In order to study the functionality of the constructs, HEK-293T cells were transfected with plasmid DNA using Lipofectamine 2000 reagent (Invitrogen, ThermoFisher Scientific, Waltham, MA, USA). Transfections were performed according to the manufacturer's protocols. 1–2 μg of plasmid was transfected, depending on the size of the culture plate used (6- or 12-wells). Expression was analyzed 24–48 h post-transfection (hpt) under a microscope equipped with epifluorescent illumination (Nikon Eclipse 800).

Recombinant single-stranded AAV8 vectors were purified from HEK-293T cells that had been co-transfected using linear polyethylenimine 25 kDa (Polysciences, Warrington, PA, USA) with two different plasmids: a plasmid containing ITR-flanked transgene constructs and a plasmid containing the adenoviral helper genes plus AAV2 rep and AAV8 cap (named pDP8.ape, Plasmid Factory, Bielefeld, Germany) as described (Durocher et al., 2002). Seventy-two hpt the supernatant was collected and treated with polyethylene glycol solution (PEG8000, 8% v/v final concentration) for 48–72 h at 4°C. Supernatant was then centrifuged at 3000 rpm for 15 min. Pellet containing particles from the supernatant was resuspended in lysis buffer and kept at −80°C. Cells containing AAV particles were collected and treated with lysis buffer (50 mM Tris-Cl, 150 mM NaCl, 2 mM MgCl₂, 0.1% Triton X-100) and kept at −80°C. Three cycles of freezing and thawing were applied to both supernatant and cell lysate. Viral particles obtained from cell supernatant and lysate were purified by ultracentrifugation in an iodioxanol gradient according to the method of Zolotukhin et al. (1999). The viral batches were then concentrated further by passage through centricon tubes (YM-100; Millipore, Bedford, MA). All vector stocks were kept at −80°C until used.

AAV vector titers (viral particles (vp)/ml) were determined by quantitative PCR for viral genome copies extracted from DNAase-treated viral particles (High Pure Viral Nucleic Acid Kit, Roche). The primers used in the q-PCR were Forward-eGFP: 5′-GTCCGCCCTGAGCAAACA-3′ and Reverse-eGFP: 5′-TCCAGCAGGACCATGTGATC-3′. Vector titers obtained ranged from 2 × 10¹² to 9 × 10¹² vp/ml.

Mice were sacrificed 3 weeks post-surgery by transcardiac perfusion with saline Ringer solution followed by 4% paraformaldehyde in 0.1 M phosphate buffer (PB). Brains were dissected and stored for 48 h in a cryopreservation solution containing 10% glycerin and 2% dimethylsulphoxide (DMSO) in 0.125 M PB, pH 7.4 at 4°C. Frozen serial coronal sections (40 μm thickness) were obtained using a sliding microtome and collected in cryopreservation solution in series of 10 adjacent sections.

Free-floating sections were rinsed with Tris buffer pH 7.4 (TBS) and then incubated in a blocking solution containing 1% cold fish gelatin (Sigma), 1% bovine serum albumin (BSA), and 0.05% Triton X-100 in TBS for 1 h; sections were then incubated overnight at room temperature (RT), with the appropriate primary antibodies diluted in blocking solution.

The following primary antibodies were used in double immunofluorescent stains: (1) rabbit anti-GFAP (1:400, Dako, Glostrup, Denmark; catalog number Z0334). (2) Mouse anti-GFAP 1:400, AbD Serotec, Killington, UK; catalog number 4650-0309). (3) Mouse anti-neuronal nuclear antigen (NeuN; 1:500, Millipore, Darmstadt, Germany; catalog number MAB 377). (4) Goat anti-olig2 (1:200, R&D systems, Minneapolis, MN; catalog number AF2418). (5) Rabbit anti-Iba1 (1:500, Wako, Neuss, Germany; catalog number 019-19741). After rinsing with TBS, sections were incubated with the appropriate fluorescent secondary antibodies diluted as before for 1 h. The following secondary antibodies were used (all purchased from Molecular Probes and used 1:200): Alexa Fluor® 633 donkey anti-rabbit IgG (#A21070), Alexa Fluor® 633 donkey anti-mouse IgG (#A21050); Alexa Fluor® 546 donkey anti-mouse IgG (#A10036); Alexa Fluor® 633 donkey anti-goat IgG (#A21080); Alexa Fluor® 555 donkey anti-rabbit IgG (#A31572), Alexa Fluor® 546 goat anti-rabbit (#A11010) Alexa Fluor® 546 goat anti-mouse (#A11003). Finally, sections were rinsed in PBS and mounted on SuperFrost Ultra Plus slides, dried at RT and coverslipped with Depex (VWR International). As negative control and to verify the specificity of the secondary antibodies, the same immunohistochemistry procedure was performed omitting the primary antibodies. No staining was observed. Furthermore, all antibodies used here were used in other publications (see Eng et al., 2000; Talbott et al., 2007; Gil-Perotin et al., 2009; Lalancette-Hebert et al., 2012; Seto et al., 2014; Haberl et al., 2015). Sections were inspected under a confocal laser-scanning microscope (LSM 800; Zeiss, Jena, Germany). To ensure appropriate visualization of the labeled elements and to avoid false positive results, the emission from the argon laser at 488 nm was filtered through a band pass filter of 505–530 nm and color-coded in green. The emission following excitation from the helium laser at 543 nm was filtered through a band pass filter of 560–615 nm and color-coded in red. A long-pass filter of 650 nm was used to visualize the emission from the helium laser at 633 nm and color-coded in pale blue.

As the main goal of this study was to determine the specificity of the chosen promoters in the context of AAV8-mediated gene delivery to the striatum, we focused our analysis on the transduced area only. The numbers of eGFP-positive cells infected with each vector were determined on images of eight random areas within the transduced striatum regions (i.e., containing at least one eGFP⁺ cell) per mouse using a 40x objective and ImageJ software. Percentages were calculated based on the total number of transduced cells (number of eGFP⁺NeuN⁺/total NeunN⁺, eGFP⁺GFAP⁺/total GFAP⁺, or eGFP⁺Olig2⁺/total Olig2⁺, respectively).

The results were expressed as mean ± standard deviation (SD). Statistical analyses were performed using the software GraphPadPrism. To test for difference in transduction efficacy, a non-parametric one-way ANOVA with Tukey post-test was applied, except for Figure 5 where we used Chi square analysis. All tests were considered significant if p < 0.05.

A total of six recombinant AAV genomes carrying an eGFP reporter gene were constructed (Figure 1; for a more detailed description of the vectors see Section Materials and Methods). In brief, five constructs carried CNS cell-specific promoters and one a ubiquitous promoter, CAG pr, which was used as control. Three promoters targeting astrocytes were tested: hGFAP pr, hGFAPΔD pr, and mGFAP pr, as well as two neuronal ones, BM88 pr and CHNRB2 pr. The expression of eGFP, driven by the different constructs, was first analyzed in vitro by plasmid transfection of HEK-293T cells (Figure 2). All promoters were able to drive the expression of the fluorescent protein, however, important differences in their transcriptional activity were found. Of the neuronal promoters, BM88 was stronger than CHNRB2, while the astrocyte promoter mGFAP was better than hGFAP, and the level of eGFP expression was not diminished upon deletion of the D region. Non-transfected controls were eGFP negative whereas the majority of cells transfected with plasmid containing eGFP under the control of CAG pr were strongly positive.

After testing the functionality of the different plasmids in vitro, we produced the recombinant AAV8 vectors for in vivo studies. A dose of 4 × 10⁹ vp of each AAV8-eGFP vector was injected into the left striatum by stereotactic injection (n = 3 per group). The mice did not display any adverse reaction or behavioral changes after the intracranial surgery or during the subsequent period until sacrifice. However, no long term studies were performed to test the potential toxicity of sustained transgene expression. Three weeks after vector injection mice were euthanized and eGFP expression was analyzed in both the right and left striatum. In all groups eGFP expression was detected in the left striatum as well as along the injection tract but never in the right brain hemisphere (See Supplementary Figure 1). The number of eGFP-positive cells/μm² varied depending on the promoter. The highest number of transduced cells was observed with the vector carrying the neuron-specific promoter BM88 pr (Figure 3). Transduction was significantly less efficient using the three variants of GFAP, CHNRB2, or CAG as promoter. The lowest levels of expression were consistently found with the construct containing CHRNB2 pr.

CNS tissue is highly heterogeneous and consists of different cell types including neurons and glia cells. Following the delivery of eGFP expressing vectors under the control of the constitutive promoter CAG, eGFP-expressing cells with different morphologies were observed. To further identify the type(s) of cells transduced by this vector, a triple immunofluorescence stain using both anti-eGFP as well as cell-specific markers was performed. Astrocytes were identified by their expression of GFAP, whereas for neurons the pan-neuronal marker of neuronal nuclei (NeuN) was used (Figure 4). AAV8-CAG-eGFP mainly transduced neuronal cells and to a lesser extent also astrocytes and oligodendrocytes. Quantification of eGFP-expressing cells revealed that the number of eGFP-positive neuronal cells was eight-fold higher than the number of astrocytes (Figure 5). These results indicate that both neurons and astrocytes are transduced by AAV8 after stereotactic injection into the striatum, albeit with a greatly varying efficacy.

Next we wanted to analyze the specificity of the vectors containing either neuron or astrocyte-specific promoters. In mice injected with either of the AAV-GFAPpr variants most of the eGFP-positive cells were astrocytes (Figure 6), while neurons were never found to express eGFP (Figure 7). Interestingly, a population of eGFP-positive cells lacking GFAP expression was detected. These cells were lacking the morphological features that typically characterize astrocytes and their small size and morphology were consistent with an oligodendroglial phenotype. In an attempt to properly identify the exact nature of these cells, we labeled brain sections with anti-eGFP, anti-Olig 2, an oligodendroglial marker, and anti-Ibal, a microglial marker. As shown in Figure 8, eGFP co-localized with Olig2-expressing cells but not with Ibal.

Mice injected with the vector carrying the mGFAP promoter had the highest levels of transduction. A strong fluorescence was observed in cell bodies throughout the dorsal area of the striatum. Furthermore, the number of positive cells was similar in the groups injected with either hGFAP or hGFAPΔD, indicating that the D element is dispensable for the transcriptional activity of the promoter (no significant difference was observed between these two promoters; Figure 9A). This is in line with what was suggested by our in vitro results. Three weeks after viral injection, 81.5% of astrocytes in AAV8-mGFAP recipients were eGFP positive. With 55.5% of positive cells in AAV8-hGFAP- and AAV8-hGFAPΔD-injected mice, respectively, these promoters were less efficient (Figure 9A). Moreover, co-localization of reporter gene expression and the oligodendroglia marker Olig2 in the AAV8-mGFAP group was the lowest with 17.4% (Figure 9B). With the other variants 20% (AAV8-hGFAPΔD) and 26.6% (AAV8-hGFAP), respectively, stained positive for both eGFP and Olig2. These results indicate that the mGFAP promoter works best for astrocyte-specific transgene expression.

Selective neuronal transduction was observed in mice having received the vector in which eGFP had been placed under control of the CHNRB2 or BM88 promoters (Figure 10). Widespread expression of the transgene was observed throughout the dorsal region of the striatum. At higher magnification specific eGFP expression was identified within NeuN-positive cells (Figure 10), suggesting that eGFP-expressing cells accounted for a neuronal phenotype. This is supported by the fact that co-localization of eGFP and GFAP was never found. The percentage of NeuN-positive cells co-expressing eGFP was significantly higher in animals injected with the BM88 vector (63.4%) than in those injected with the CHNRB2 promoter (15.9%), thus identifying the former as the more useful one for expressing transgenes in neuronal cells (Figure 11).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.