Effective relief of neuropathic pain by adeno-associated virus-mediated expression of a small hairpin RNA against GTP cyclohydrolase 1

For knockdown of rGCH1 expression, nine siRNAs targeting various rGCH1 exon regions were designed (Fig. 1). The GCH1 silencing effect was examined using 293T cells exogenously expressing rGCH1, which were transfected with rGCH1-specific or control non-specific siRNA. Western-blotting analysis revealed that all specific siRNAs markedly inhibited rGCH1 expression, while control siRNAs did not. For further analysis, we selected a siRNA sequence targeting exon 4 of rGCH1 and generated a shGCH1 construct expressing shRNA under the control of a H1 promoter [15]. As expected, shGCH1 dramatically suppressed rGCH1 expression, similar to its corresponding siRNA (Fig. 2a). To achieve persistent shRNA expression, we constructed rAAV-shGCH1 by inserting this shGCH1 expression cassette into the rAAV2 backbone. rAAV-shGCH1 additionally encodes the GFP reporter gene driven by the CMV promoter to allow the simultaneous tracking of shGCH1 expression (Fig. 2b). rAAV-GFP lacking the shGCH1 expression cassette was employed as a control. The rGCH1 silencing effect was viral dose-dependent in HeLa cells (Fig. 2c). As expected, the rGCH1 level was not affected by the control virus. A similar degree of GFP expression was observed with equivalent amounts of either virus. Our data collectively suggest that rAAV-mediated shGCH1 synthesis effectively suppresses GCH1 expression.

Recent studies propose that sciatic nerve injection of rAAVs is a promising strategy for delivery of therapeutic genes to DRG [21, 22]. In view of these findings, we attempted to deliver transgenes into DRG by administering 6.0 × 10⁷ of rAAV-shGCH1 or rAAV-GFP to the sciatic nerve 9 days after spared nerve injury (SNI). Gene transfer efficiency was confirmed by the presence of GFP-positive DRG neural cells under a fluorescent microscope (Fig. 3). Intense green signals were discerned in all infected mice from both experimental groups on day 14 post virus injection (n = 4 for each group), indicative of effective GFP transgene expression. In contrast, green fluorescence was absent in control groups (n = 5). In equivalent experiments involving treatment with viruses prior to pain surgery, we observed analogous patterns of notable GFP expression in experimental groups (data not shown). The data confirm that the sciatic nerve is a useful delivery route for transducing the DRG with the genes of interest, in both cases of pre- and post-injection.

Spared nerve injury was generated by tight ligation and transaction of the tibial and common peroneal nerves while leaving sural nerves intact [23], and promptlly developed pain behavior which was determined by mechanical von Fray threshold. Following surgery, the mechanical threshold began to drop sharply from normal range to pain threshold levels (Fig. 4). All the animals showed pain symptoms within few days which were maintained thereafter (on day 7; 2.2 ± 0.9 at the threshold level). To explore the beneficial effects of GCH1 down-regulation on neuropathic pain, related behavior was initially examined under blinded conditions. After introducing rAAV-shGCH1 (n = 10) or control rAAV-GFP (n = 10) into rats (9 days post pain surgery), the mechanical von Fray test was performed at the designated time-points. As shown in Fig. 5, the average von Fray threshold was 2.8 ± 0.3 at the threshold level on day 9 after pain surgery. Following rAAV-shGCH1 treatment, the pain threshold gradually declined over time, except in the case of rAAV-GFP as expected. To ensure that GCH1 down-regulation was specifically promoted by rAAV-shGCH1, GCH1-specific immunohistochemistry and western-blotting assay was performed in ipsilateral L4/5 DRGs. In immunohistochemistry, shGCH1 introduction via rAAV2 significantly suppressed GCH1 expression to a similar level as the no-pain control group, while the up-regulation of GCH1 remained unchanged in the rAAV-GFP injected group (Fig. 6a), which supports previous report showing the positive association of pain symptoms with GCH1 activation [10]. The relative GCH1 levels were 2.3 ± 0.4 and 1.2 ± 0.3 in GFP- and shGHC1-treated pain groups, respectively (Fig. 6b). In addition, the immunohistochemical staining against GCH1 showed a prominent negative correlation with the mechanical threshold of respective animals (Spearman rho test, Correlation coefficient = -0.470), though there was no statistical significance mainly due to small sample numbers. Similar pattern of GCH1 expression was also observed in western-blotting analysis (Fib. 6c). These results collectively illustrate that rAAV-shGCH1 treatment of the sciatic nerve effectively suppresses GCH1 expression and leads to effective pain relief.

As predicted, injection of rAAV-shGCH1 prior to SNI prevented pain induction whereby several rats experienced no differences in the pain test compared to control animals (Fig. 7). Similar to post-injection, the anti-pain effect upon pre-treatment reached maximal levels on day 14 after virus injection. In conjunction with data obtained from Fig. 3, these results suggest that shGCH1 introduction into the sciatic nerve via rAAV2 delivery leads to marked attenuation of pain symptoms, regardless of treatment times.

Previous studies have demonstrated that inflammatory activation of microglia in the ipsilateral spinal cord plays a crucial role in the development of neuropathic pain [24]. To determine whether GCH1 downregulation affects microglial activation, we performed immunohistochemistry with the Iba-1 antibody, a microglial marker (Fig. 8). Indeed, nerve injury activated microglia, as evident from both the increased number and the morphological changes of microglia in the ipsilateral dorsal horn (data not shown). The GFP pain group experienced similar microglial alterations. In contrast, microglial activation was significantly inhibited in the shGCH1 pain group, and a similar pattern of microglial inactivation was observed in case of virus pre-treatment. These results suggest that pain attenuation following shGCH1 introduction is associated with a decrease in inflammation.

293T and HeLa cells were purchased from the American Type Culture Collection (Manassas, VA). Cells were maintained in Dulbecco's modified Eagle's medium (Gibco BRL, Carlsbad, CA) supplemented with 10% fetal bovine serum, L-glutamine (2 mM), penicillin (100 IU/ml) and streptomycin (50 μg/ml) at 37°C in a 5% CO₂ incubator.

All siRNAs against GCH1 were designed using software from MWG Biotech AG http://​www.​mwg-biotech.​com. In total, nine duplex siRNAs (Supple. 1) with a dTdT 3' overhang and control scrambled siRNA were manufactured by DHARMACON (Lafayette, CL) with the "ready-to-use" option. Cells were transfected with 100 nM siRNA complexed with Oligofectamine reagent (Invitrogen, Carlsbad, CA) in OPTI-MEM medium (Invitrogen). After 3 h, media were replaced with growth medium containing 10% serum. The next day, cells were harvested and extracted for Western blotting.

The human H1 polymerase III promoter (pH1) for shRNA expression was amplified from HeLa cells using PCR-based methods, as described previously [36]. The shRNA sequences against rGCH1 (shGCH1) are 5'-GAT CCC GTG GAA ATC TAC AGT AGA ATT CAA GAG A TT CTA CTG TAG ATT TCC ACT TTT TTG GAA A-3' (sense) with a Bam HI linker and 5'-AGC TTT TCC AAA AAA GTG GAA ATC TAC AGT AGA ATC TCT TGA A TT CTA CTG TAG ATT TCC ACG-3' (anti-sense) with a Hind III linker. Nucleotides specific for rat GCH1 are underlined, and the loop structures presented in bold. The pH1-shGCH1 sequence was inserted adjacent to the CMV promoter of the vector expressing GFP, resulting in a bi-directional promoter (Figure 1a). All clones were verified by sequencing (data not shown).

To produce rAAV2, the viral backbone, pH1-shGCH1-pHpa-trs-SK, was co-transfected with pRepCaps and pXX6 (Stratagene, La Jolla, CA) encoding adenoviral helper genes [37, 38]. For large-scale rAAV preparations, 293T cells were transfected in 20 × 10 cm dishes. rAAVs in cells were released and purified from cell lysates (including supernatant fractions) using two sequential CsCl gradient steps. After dialysis in 10 mM Tris buffer (pH 7.9) containing 2 mM MgCl₂ and 2% sorbitol, rAAVs were aliquoted and stored at -80°C. The number of total virus particles was estimated with an ELISA kit (Progen Inc., Heidelberg, Germany) and real-time qPCR, as described in a previous report [38].

Adult male Sprague Dawley rats weighing 180-200 g were housed with free access to food and water. With the approval of the University committee on the use and care of animals, rats were maintained in an optimal temperature and humidity-controlled room with a 12 hr light/dark cycle.

To inject rAAVs to the sciatic nerve, rats were anesthetized with Zoletil (50 mg/kg) and Rompun (10 mg/kg). Next, a segment of the left sciatic nerve was exposed at the mid-thigh level. Surrounding tissues were carefully removed under a surgical microscope, exposing sciatic nerve exposed (Olympus, Japan). A 3 μl viral solution containing rAAV-shGCH1 (6.0 × 10⁶ viruses in 3 μl) or rAAV-GFP (4.0 × 10⁶ viruses in 3 μl) was slowly injected into the nerves through a glass micropipette connected to a Hamilton syringe for 5 min. The pipette was pulled out after 5 min.

To generate the spared nerve injury (SNI) model, the common peroneal and tibial nerves were tightly ligated and completely transected while the sural nerve was left intact [39]. Mechanical allodynia in rats was assessed by applying a series of Von Frey filaments to the plantar surface of the hind paw (range 0.01 - 1.66 g) in ascending and descending order (up-and-down method). Brisk withdrawal of the hind limb was considered a positive response. A withdrawal threshold of 4.0 g or less was classified as allodynia (pain-like behavior).

L4-5 DRG tissues were homogenized with lysis buffer containing 10 mM Tris-HCL (pH 8.0), 150 mM NaCl, 1% Triton-X100 and protease inhibitor cocktails (Sigma, St. Louis, MO), and centrifuged for 30 min at 4°C. Lysates (10-40 μg of proteins per lane) were separated by 10% SDS-PAGE and transferred onto PVDF membranes (Bio-Rad, Hercules, CA). Blots were incubated with the specified primary antibodies. The following antibodies were used: GCH1 (Abnova, Taiwan), β-actin (Sigma), and GFP (Santa Cruz Biotechnology, Santa Cruz, CA). Following incubation with peroxidase-conjugated anti-mouse (or anti-goat) IgG (Vector Laboratories, Burlingame, CA), proteins were detected using the chemiluminescence method (Pierce, Rockford, IL).

Frozen tissue sections of 10 μm in thickness were prepared and Nissl-staining was carried out, as described previously [40]. Immunostaining was also performed using frozen section. Briefly, after perfusion with 4% paraformaldehyde, DRG (L4, L5) and spinal cord (T13-L2) were removed and cryoprotected in 30% sucrose solution (dissolved in phosphate-buffered saline) for 24 h at 4°C before embedding in OCT compound (Sakura Finetek, Torrance, CA). Sections were incubated at 4°C with primary antibodies overnight. Antibodies used for immunohistochemistry were as follows: GCH1 (Abnova), NeuN (Chemicon, Temecula, CA), Iba-1 (Wako, Japan), phosphor-p38 (Cell signaling Technology, Beverly, MA), and OX-42 (Chemicon). Sections were subsequently incubated with Alexa 488-conjugated (or Alexa 555) secondary antibodies at room temperature for 60 min. After mounting using Vectashield with DAPI (Vector laboratories), images were observed by fluorescence microscopy (Leica, Germany).