Mechanisms of PDGF siRNA-mediated inhibition of bone cancer pain in the spinal cord

Patients with tumors that metastasize to bone frequently suffer from debilitating pain, and effective therapies for treating bone cancer are lacking. This study employed a novel strategy in which herpes simplex virus (HSV) carrying a small interfering RNA (siRNA) targeting platelet-derived growth factor (PDGF) was used to alleviate bone cancer pain. HSV carrying PDGF siRNA was established and intrathecally injected into the cavum subarachnoidale of animals suffering from bone cancer pain and animals in the negative group. Sensory function was assessed by measuring thermal and mechanical hyperalgesia. The mechanism by which PDGF regulates pain was also investigated by comparing the differential expression of pPDGFRα/β and phosphorylated ERK and AKT. Thermal and mechanical hyperalgesia developed in the rats with bone cancer pain, and these effects were accompanied by bone destruction in the tibia. Intrathecal injection of PDGF siRNA and morphine reversed thermal and mechanical hyperalgesia in rats with bone cancer pain. In addition, we observed attenuated astrocyte hypertrophy, down-regulated pPDGFRα/β levels, reduced levels of the neurochemical SP, a reduction in CGRP fibers and changes in pERK/ERK and pAKT/AKT ratios. These results demonstrate that PDGF siRNA can effectively treat pain induced by bone cancer by blocking the AKT-ERK signaling pathway.

genes can partially relieve pain 16,20 , indicating that HSV-based constructs carrying siRNA have the potential to effectively treat bone cancer pain.
Platelet-derived growth factor (PDGF), a cationic disulfide-linked polypeptide that exists as a homo-or heterodimer, can promote cell migration, proliferation, and differentiation 21,22 . PDGF is involved in several pathologies, including oncogenesis, atherosclerosis, and lung and kidney fibrosis [23][24][25] . Moreover, hyper-active PDGF results in tactile allodynia, and intrathecal administration of either a selective inhibitor of PDGF receptor (PDGFR) phosphorylation or an antibody blocking endogenous PDGF suppresses thermal hyperalgesia and tactile allodynia 26,27 . These results suggest that PDGF knockdown might relieve mechanical and thermal hyperalgesia.
In the present study, we developed a clinically relevant model of bone cancer pain and investigated the potential role of PDGF in mediating bone cancer pain. An HSV vector carrying 2 artificial PDGF small interfering RNAs was intrathecally injected into the L4-6 spinal cord segments of animals with bone cancer pain. The injection of tumor cells into the tibia induced an increase in PDGF levels in the L4-6 spinal segment. In addition, immunofluorescence staining demonstrated that PDGF localized to microglia cells, astrocytes and neurons, cell types expected to be infected by HSV. These data demonstrate that reducing PDGF expression using RNAi inhibits the expression of glial fibrillary acidic protein (GFAP), substance p (SP), calcitonin gene-related peptide (CGRP) and the phosphorylation of ERK and AKT and therefore represents a novel strategy for treating bone cancer pain.

Results
Anatomical and behavioral characterization of the bone cancer pain model. Histological and radiological analyses were performed to confirm tumor formation following the injection of tumor cells into the bilateral tibia in the bone cancer pain model. Hematoxylin and eosin (H&E) staining 28 days after injection demonstrated that the bone matrix in the bone marrow of these animals was damaged and filled with tumor cells and osteoclasts; these features were not observed in the normal control rats (Fig. 1A). Furthermore, radiographs of the animals' hind limbs revealed bone destruction in the bilateral tibia compared with the control rats (Fig. 1B).
To evaluate the development of pain-related behaviors induced by the intra-tibial injection of tumor cells, tail-flick latency (TL) and paw withdrawal threshold (PWT) were assessed at baseline and 1, 2, 3 and 4 weeks after tumor cell inoculation. The animals in the bone cancer pain group exhibited a decrease in TL or PWT over time.
There was a significant decrease in TL and PWT 4 weeks after tumor cell injection (Fig. 1C). The decrease in TL was observed as early as 21 days after the injection of tumor cells (P = 0.001) (Fig. 1C) and worsened over the time course of experiment (Fig. 1C). The decrease in TL following thermal stimulation indicated the development of thermal hyperalgesia in the bone cancer pain rat model.

Bone cancer pain enhances PDGF expression in neurons and glia cells in the spinal cord.
To investigate the molecular mechanisms by which bone cancer induces pain, we examined the expression of PDGF in the spinal cord. Changes in the levels of PDGF in animals with bone cancer pain were detected at the mRNA and protein levels. PDGF mRNA expression was elevated in animals with bone cancer pain (0.00113 ± 0.000204) compared with the control rats (0.00031 ± 0.000390) (P = 0.002) (Fig. 1D). Consistent with these findings, PDGF protein levels were elevated in the bone cancer pain group (1.17 ± 0.174790) compared with the control group (0.53 ± 0.060825) (P = 0.004) (Fig. 1D). Immunofluorescence staining demonstrated that PDGF was expressed in neurons, astrocytes and microglia cells (Fig. 2). The number of cells expressing PDGF and the astrocyte marker GFAP increased in the rats with CIBP compared with the control rats ( Fig. 2A). Similar results were observed for staining of the microglia cell marker OX-42 and PDGF (Fig. 2B). The number of NeuN-positive cells should remain unchanged if neuronal cell death or neurogenesis does not occur. No detectable changes in the number of neurons were observed, but there was an increase in PDGF-positive cells in the rats with CIBP compared with the control rats (Fig. 2C). Characterization of the HSV-1-PDGF siRNA vector. To investigate the effect of increased levels of PDGF in rats with CIBP, we constructed a replication-incompetent HSV-1 vector carrying a siRNA to inhibit PDGF gene expression via RNA interference (RNAi) (Fig. 3A). Semi-quantitative RT-PCR analysis using the U6-F/H1-F primer demonstrated that pNX01-U6H1-rPDGFi vector samples 1-4 contained the approximately 477-bp PDGF siRNA sequence (Fig. 3b). Sample 4 of the virus of NX01-U6H1-rPDGFi also included the target sequence of the PDGF siRNA as demonstrated by semi-quantitative RT-PCR analysis using the same primer (Fig. 3C). The infection efficiency of the HSV virus was evaluated by immunofluorescence assays 7 days following the injection of the viruses into the rat spinal cords. The red signal in Fig. 3 represents red fluorescent protein (RFP), the marker used to label the HSV vector. Cells throughout the dorsal horn of the spinal cord including astrocytes, microglia cells and neurons, were infected with the virus as demonstrated by the colocalization of RFP and cell type-specific markers (green; Fig. 3D-F).

Inhibiting PDGF expression attenuates tumor-evoked mechanical and thermal hyperalgesia.
Treatment with PDGF siRNA reduced PDGF protein expression in the L4-L6 spinal cord segments of animals with CIBP compared with the negative control treatment (P = 0.000) (Fig. 4A). The second week following tumor cell inoculation, the rats with CIBP exhibited an increase in thermal hyperalgesia from baseline (blue and green asterisk in Fig. 4C), whereas 7 days after the injection of the PDGF siRNA, thermal hyperalgesia was attenuated in the treated animals compared with the negative control vector (NC)-treated animals (P = 0.000) (purple asterisk, Fig. 4C). Fourteen days after the injection of PDGF siRNA in L4-L6, we observed improvements in PWT and TL compared with the NC-treated group (P = 0.000 and P = 0.002, respectively) (purple asterisk, Fig. 4B,C). Treatment with morphine also attenuated mechanical and thermal hyperalgesia 7 days after administration compared with saline treatment (P = 0.012 and P = 0.000, respectively, for PWT; P = 0.002 and P = 0.004 for TL, respectively) (red asterisk, Fig. 4B,C). PDGF siRNA was as effective as morphine in alleviating mechanical and thermal hyperalgesia.

PDGF down-regulation and treatment with morphine reverses bone cancer pain-induced changes in the expression of the neurochemicals GFAP, SP and CGRP.
To investigate the mechanisms by which PDGF induces bone cancer pain, we measured changes in the levels of the neurochemicals SP and CGRP and evaluated astrocyte hypertrophy in the spinal cord. Rats injected with tumor cells exhibited increased levels of SP in laminae I-II, and the RNAi-induced reduction of PDGF levels suppressed SP levels in  In addition, SP immunoreactivity decreased by 1.7 ± 0.3434 in animals treated with morphine compared with animals treated with saline (P = 0.000) (Fig. 4F,G). These results are consistent with the inhibition of SP protein expression in the spinal cord by PDGF down-regulation in the bone cancer pain group compared with the control group (P = 0.025). Morphine treatment also reduced the expression of SP compared with saline treatment (P = 0.029) (Fig. 4J,K).
In addition to the increase in neurochemicals, we also observed an increase in astrocyte hypertrophy (an indicated by increased GFAP-immunoreactivity) in the L4-L6 spinal segments of animals with CIBP. PDGF siRNA reduced GFAP immunofluorescence (0.81 ± 0.1417) compared with NC siRNA and morphine (P = 0.000 for both) (Fig. 4D,E). Morphine treatment, however, did not influence GFAP immunofluorescence (Fig. 4D,E). In addition, treatment with PDGF siRNA reduced GFAP expression compared with the NC control (P = 0.049) (Fig. 4J,K).

Reducing PDGF expression inhibits the phosphorylation of PDGF receptors and suppresses pain induced ERK and AKT signaling.
In an effort to identify the downstream mechanisms by which PDGF induces bone cancer pain, we used WB to evaluate the activation (phosphorylation) of PDGFR, ERK and AKT by measuring levels of pPDGFRα /β , pERK and pAKT 14 days after treatment with PDGF siRNA. The increase in PDGF expression in the spinal cord segments L4-6 of mice injected with tumor cells coincided with increases in pPDGFRα /β and pAKT/AKT expression compared with the control group (Fig. 5A,B). As shown in histogram B, pPDGFRα /β levels increased in the saline-and NC-treated groups compared with the normal group (P = 0.000 for both in the left illustration and P = 0.001, 0.002 for the right, respectively), whereas histogram C demonstrates that pAKT/AKT levels were elevated in the saline-and NC-treated groups compared with the normal group (P = 0.001 and P = 0.002, respectively). pPDGFRα /β levels decreased in the HSV-siPDGF-treated group compared with the NC siRNA-treated group (P = 0.000 and P = 0.017, respectively). A similar trend was observed in the morphine-treated group but was not statistically significant. PDGF siRNA treatment reduced the pERK:ERK and pAKT:AKT ratios compared with NC siRNA treatment (P = 0.01 and P = 0.09, respectively) ( Fig. 5C,D). Morphine treatment also reduced the pERK:ERK and pAKT:AKT ratios compared with saline treatment (P = 0.000 and P = 0.001, respectively) (Fig. 5C,D).

Discussion
Here, we investigated the analgesic effect of intrathecal injection of PDGF siRNA in a bone cancer pain animal model generated by injecting MRMT-1 tumor cells into the tibias of female rats. First, we developed a bone cancer pain animal model that mirrored features of clinical bone cancer pain 28 , including evoked and spontaneous pain. The number of tumor cells used for tibia injections (4 × 10 5 cells), host immune suppression by daily cyclosporine administration, and the intrathecal delivery of the HSV virus carrying the PDGF siRNA were critical factors in establishing this bone cancer pain model 29 . Moreover, sterilized medical bone wax was used to close the injection site to prevent backrush of tumor cells after injection into the marrow cavity 30 . Female rats were used because they are more sensitive to pain stimuli and exhibit more pronounced behavioral effects 6,31 . Twenty-eight days after the injection of the tumor cells in the tibia, the animals developed thermal and mechanical hyperalgesia, as demonstrated by the reduction in TL and PWT. H&E staining revealed the presence of tumor cells and activated osteoclasts in the tibia, and radiography confirmed bone destruction. These characteristics of the bone cancer pain model are consistent with the findings of previous studies 32, 33 .
Morphine, an opioid receptor agonist, is a commonly used clinical analgesic 34 , but its clinical use is limited by tolerance, addiction and adverse effects 35 . A single dose of morphine for analgesia does not affect GFAP expression or activate astrocytes in the spinal cord 36,37 , consistent with our recent data. Acute administration of morphine blocks spinal NK1r internalization (SP release) and abolishes the nerve-evoked release of SP 38,39 . Welch S P et al. previously reported that acute administration of morphine suppresses the expression of CGRP in the corpus striatum 40 . Morphine treatment also markedly attenuated phosphorylation of ERK1/2, which could induce the hyperalgesia 41 . Therefore, the mechanisms by which morphine and PDGF siRNA induce analgesia may involve little kind of same identical molecules as the above described. In the present study, we compared the efficacy of PDGF siRNA and morphine in treating bone cancer pain. The mechanical and thermal hyperalgesia induced by the injections of tumor cells into the tibias of rats were reversed by RNAi-mediated inhibition of PDGF or by treatment with morphine. These results demonstrate that intrathecal administration of PDGF siRNA can alleviate pain associated with bone cancer.
Central sensitization is a type of activity-dependent plasticity in spinal cord neurons 42 that results in hyperalgesia under pathological conditions 42,43 . In our model of bone cancer pain, we observed astrocyte hypertrophy and increases in the levels of SP, CGRP, pERK/ERK, and pAKT/AKT in the spinal cord segments L4-6, consistent with previous reports 44,45 . Astrocyte hypertrophy is observed in multiple inflammatory and neuropathic pain models 46,47 , suggesting that astrocytes partially mediate pain in these contexts 48 . Increased expression of SP is an indirect marker of sensitization of primary afferents 45 , whereas an increase in CGRP in laminae I-II of the spinal cord dorsal horn might be associated with the mechanisms underlying peripheral and spinal pain 49 . Our study provides novel evidence indicating that treatment with PDGF-siRNA attenuates astrocyte activation and reduces the expression of SP and CGRP in lamina I-II of the dorsal horn. These findings suggest that down-regulation of PDGF via treatment with PDGF siRNA is an effective strategy to attenuate the sensitization of the primary afferent centers in the spinal cord dorsal horn that are associated with the release of chemicals that mediate pain.
PDGF signaling contributes to the pathogenesis of neuropathic pain 26 . PDGF isoforms bind to PDGF receptors to induce receptor dimerization and autophosphorylation and, ultimately, the activation of downstream signaling pathways. The PDGF-B chain activates both PDGFRα and PDGFRβ 50 ; therefore, we evaluated PDGFRα and PDGFRβ . The levels of phosphorylated PDGFRα and PDGFRβ were down-regulated in the PDGF siRNA-treated group. The ERK and AKT signaling pathways are frequently reported to be involved in pain and are verified therapeutic targets for treating diseases associated with pain 51,52 . The increased levels of pERK and pAKT in the bone cancer model evaluated in this study are in agreement with results from previous studies 53,54 demonstrating that pERK and pAKT levels in the spinal cord increase under conditions of inflammation and neuropathic pain 55,56 . Phosphorylation of ERK has been reported as a marker of central sensitization 54,57,58 . Activation of the ERK and AKT signaling pathways is associated with the modulation of nociceptive information and peripheral and central sensitization in the spinal cord 53,59 . In the current study, intrathecal injection of PDGF siRNA inhibited ERK and AKT signaling, indicating that PDGF siRNA alleviated pain by decreasing the phosphorylation of PDGFRs and down-regulating the ERK-AKT signaling pathway, events that ultimately result in the down-regulation of GFAP, SP and CGRP expression.
In conclusion, RNAi-mediated down-regulation of PDGF inhibits bone cancer pain in the spinal cord. Therefore, reducing PDGF levels in the spinal cord using RNAi might be an effective therapeutic approach for relieving CIBP.

Materials and Methods
Induction of cancer and bone cancer pain. Animal care procedures and all experimental protocols were approved by the Sichuan Medical Experimental Animal Care Commission. The methods were conducted in accordance with the guidelines of the International Association for the Study of Pain 60 . A total of 70 female Sprague-Dawley rats (200-250 g) (provided by Department of Experimental Animal Center, Sichuan University) were used in this study. Throughout the experiments, 5-6 animals were housed in each cage with a 12-hour alternating light-dark cycle at 25 °C and free access to water and food. Ten animals were randomly assigned to the normal control group. Immunosuppression in the tumor-injected rats was initiated on the day of tumor cell injection via subcutaneous administration of cyclosporine (1 mg/100 g weight; 250 mg/5 mL, diluted in 50 mL of normal saline (NS); n = 60). Cyclosporine was administered daily for 28 days after the tumor cells were injected. The cell injection protocol used in this study has been previously described 61 . Briefly, 1 × 10 7 mammary rat metastatic tumor (MRMT-1) cells (Haibo Whole Seoul Biotechnology Co., Lt) were harvested by centrifugation (10 min at 1,000 rpm) and diluted to a final concentration of 4 × 10 5 cells/mL in NS solution. Anesthesia was administered via intraperitoneal injection of 3.6% chloral hydrate (1 mL/100 g), and 10 μ L of a tumor cell solution was injected into the bilateral tibial bone cavity using a micro-syringe (n = 60). The pinhole was sealed with bone wax to prevent the carcinoma cells from leaking out along the injection track. Four weeks after the injection of tumor cells, PDGF expression in the spinal cord and bone destruction in the tibia were assessed by radiology and histological staining in 10 randomly selected rats. Subarachnoid catheterization was performed in the surviving rats (n = 32).

Intrathecal insertion of the catheter at the lumbar 4-6 level of the spinal cord.
We used a previously described procedure with some modifications 62 . Briefly, after anesthesia was administered, the fourth lumbar spinous process was removed, and the correct location was confirmed by puncturing the dura mater with a needle. After inserting a polyethylene catheter (PE-10, 15 cm) filled with NS 2cm into the subarachnoid space, the remaining portion of the catheter was sutured to the surrounding tissue to prevent slipping. More than 4 cm of the catheter was left outside the nape cervical skin, and the remainder was fixed and remained under the skin. The wound was closed in layers. Three days later, 10 μ L of a 2% lidocaine solution was injected into each animal through the catheter. Animals that did not experience temporary paralysis of the hind limbs following lidocaine injection were excluded from the study. Ultimately, 22 rats were included for further analysis.
HSV-1-based construct of the small interfering RNA targeting PDGF. The sequences of the PDGF siRNA expressed from the U6 promotor (21 bp) were as follows: rPDGFi-F1, agctG-CAGGTGAGAAAGATCGAAATcgaaAT T TCGATCT T TCTCACCTGCtttttAT, and rPDGFi-R1, C G ATa a a a a G C AG G T G AG A A AG AT C G A A ATt t c g AT T T C G AT C T T T C T C AC C T G C . T h e sequences of the PDGF siRNA expression from the H1 promoter (19 bp) were as follows: rPDGFi-F2, GAGACTCCGTAGACGAAGAttcaagacaTCTTCGTCTACGGAGTCTCTTTTTAT and rPDGFi-R2, CGATAAAAAGAGACTCCGTAGACGAAGAtgtcttgaaTCTTCGTCTACGGAGTCTCgtac. These sequences were verified by chip-Seq technology. The 2 vectors used in the study were a non-replicating negative control HSV-1 (NC-siRNA) and pNX01-U6H1-rPDGFi (HSV-siPDGF). The vectors were co-transfected into the OG01 cell line, which expresses the ICP27 and ICP4 proteins required for the packaging of the virus harboring the PDGF siRNA. The titer of NC-siRNA and HSV-siPDGF injected was 5 × 10 9 and 5.6 × 10 9 , respectively. The PDGF-B chain was specifically targeted by HSV-siPDGF in this study because expression of the PDGF-B chain is induced by peripheral nerve injury 63 .
Drug Administration. Drugs were administered by connecting a 50-μ L micro-syringe pre-filled with drug to the catheter. Morphine was dissolved in endotoxin-free physiological saline at a final concentration of 1 μ g/mL for subarachnoid administration (20 μ L; n = 5). Endotoxin-free physiological saline (n = 5), HSV-siPDGF (n = 6) and NC-siRNA (n = 6) were administered in equal volumes.
Behavioral analysis. The latency of the response of the rat tail to a radiating thermal stimulus was measured using an automatic tail-flick analgesiometer (Ugo 7360, Italy) to indicate the degree of thermal hyperalgesia. Metal fixators were used to restrain the rat while their tails hung freely to receive the radiant heat source on the third caudal position. The intensity of the radiant heat source was 80 Hz and the cutoff time was 20 s. The test was performed after the animal became stable, i.e., the rat adapted to the restraining condition. Then, TL was automatically recorded when an abrupt flick of the tail occurred. To ensure consistency, each measurement was repeated 3 times in a blinded manner.
Mechanical hyperalgesia was measured by a paw withdrawal test using a Randall Selitto paw pressure device (Bioseb, France) that applied a linearly increasing mechanical force with an acrylic stylus. The stylus was placed on the middle of the dorsum of the rat's paw. The PWT was recorded when the rat withdrew its paw. Data for both the left and right paws of each animal were collected and recorded in grams. Each measurement was repeated 3 times in a blinded manner. The data were averaged for statistical analysis.
The 2 behavioral tests were performed prior to surgery to determine the baseline values for each rat and 7, 14, 21 and 28 days after the tumor cell injections. Hyperalgesia was assessed on days 7 and 14 after intrathecal catheterization and gene administration for all groups, with the exception of the morphine-treated group, for which the tests were performed prior to the injection morphine and 10 min post-morphine injection at each indicated time point. The rats were acclimated for more than 30 min before the tests. To measure the threshold of mechanical and thermal hyperalgesia, the interval for each test was at least 10 min.
Radiology. Destruction of the tibia bone was assessed 28 days after tumor cell injection. The hind limbs were isolated from the body, and the tibias were placed on Industrex X-ray film (Kodak, Italy) and exposed to an X-ray source (Faxitron) for 1 min at 30 kVp.
Tissue preparation. Four weeks after tumor injection, the tibias of 10 randomly selected rats were isolated and immersed in 4% paraformaldehyde (PFA) for decalcification and histological staining. Two weeks later, the rats from each group were perfused and processed for immunofluorescence, western blot analysis and quantitative PCR (qPCR). After anesthetization, the animals were intracardially perfused with NS. Spinal cord segments L4-L6 were harvested. Half of the harvested spinal cords were fixed in 4% PFA, and the other half were frozen at − 80 °C.

Quantification of immunofluorescence.
Immunofluorescence was evaluated in the spinal cord at the lumbar 4 (L4) level because L4 is one of the main projection sites of primary afferent fibers innervating the hind limbs 45,65 . The saline-and morphine-treated groups included 5 animals each, and the NC-and HSV-siPDGF-treated groups included 6 animals each. Immunofluorescence intensity was measured from images captured by a Leica DMI 6000B fluorescence microscope. SP and CGRP immunoreactivity was quantified in laminae I-II, and GFAP immunofluorescence intensity was analyzed in laminae I-V of the dorsal horn in 5 randomly selected L4 sections per mouse. Briefly, 400× images of the dorsal horn of the spinal cord were acquired using an exposure time of 150 ms. Three independent experiments using the same method and software (Image J) were performed, after background subtraction, the mean fluorescence gray intensity of each image was determined. The results are presented as the mean gray intensity values of GFAP, CGRP and SP immunofluorescence normalized to normal.
Histological staining. Tibia bones were fixed in 4% PFA for 1 week, and decalcified in PFA containing 5% formic acid for an additional 4 weeks. Prior to wax embedding, tibia bone tissue was rinsed and dehydrated. 5-μ m longitudinal sections were cut from the paraffin-embedded blocks using a rotary microtome, and the sections were stained with hematoxylin and eosin to visualize the extent of tumor infiltration and bone loss.
Statistical analyses. All data were recorded as the mean ± standard error of the mean (SEM), and statistical analysis was performed using SPSS 16.0 software. The statistical significance of differences between groups was analyzed with a one-way analysis of variance (ANOVA) and the Bonferroni/Dunn multiple-comparison test. P < 0.05 was considered statistically significant.