Expression Profile of Nerve Growth Factor after Muscle Incision in the Rat

This study was approved by the institutional Animal Care and Use Committee at the University of Iowa (Iowa City, Iowa). Adult male Sprague-Dawley rats, 225–275 g, were used for all the experiments. Postoperatively, the rats were housed in groups of two or three in clear plastic cages with food and water available ad libitum .

Rats were acclimated to the testing environment. After acclimation, 100 mg/kg anti-NGF peptibody12or vehicle was administered subcutaneously at the neck. The next day, the baseline pain score was measured and the rat underwent plantar incision.13Under isoflurane anesthesia (1.5–2.0%) and using aseptic conditions, a 1-cm longitudinal incision was made 0.5 cm from the heel on the plantar aspect of the right hind paw. The underlying flexor muscle was dissected and longitudinally incised, and the wound was closed with 5-0 nylon suture. Immediately after surgery, antibiotic ointment was applied to the incision. Sutures were removed on postoperative day (POD) 2. A separate group of rats underwent incision of skin only; no fascia or muscle was incised.

A cumulative pain score was used to assess nonevoked pain behaviors as described previously.14Unrestrained rats were placed on a plastic mesh floor (8 × 8 mm). The incised and nonincised paws were viewed. Both paws of each animal were closely observed during a 1-min period repeated every 5 min for 1 h. Depending on the position in which each paw was found during the majority of the 1-min scoring period, a 0, 1, or 2 is given. Full weight-bearing of the paw (score = 0) was present if the wound was blanched or distorted by the mesh. If the paw was completely off the mesh, a score of 2 was recorded. If the area of the wound touched the mesh gently without any blanching or distorting, a 1 was given. The sum of the 12 scores (0–24) obtained during the 1-h session for each paw was obtained. The difference between the scores from the incised paw and nonincised paw is the cumulative pain score for that 1 h period. The person performing the behavioral experiments was blinded to the type of incision and the drug (or vehicle) administered.

Total ribonucleic acid (RNA) from incised rat flexor digitorum brevis muscle tissue was prepared according to manufacturer's instructions using the RNeasy mini kit (Qiagen, Valencia, CA). Some muscle samples from sham-incised rats were used as control. The RNA was quantified using Ribogreen reagent (Molecular Probes, Eugene, OR). Two-step RT-PCR was performed on the 7000 Sequence Detection System (Applied Biosystems, Foster City, CA) by using the TaqMan Gold RT-PCR kit (Applied Biosystems). The RNA was reverse transcribed using random hexamers, and the complementary deoxyribonucleic acid (cDNA) was amplified using a primer/ probe set specific for NGF (RatNGF-fwd: TGACTCCAAGCACTGGAACTCAT, complementary to 693–715 nucleotides in the rat NGF mRNA; RatNGF-rev: GTTTGTCGTCTGTTGTCAACGC, complementary to 947–926 nucleotides in the rat NGF mRNA; RatNGF-Tamra: TGCACCACGACTCACACCTTTGTCA, complementary to 718–742 nucleotides in the rat NGF mRNA). The resulting PCR product was 309 bp in length. Thermal cycling was initiated with an initial incubation at 95°C for 10 min. After that, 40 PCR cycles were performed; each cycle consisted of heating at 95°C for 15 s for melting and at 60°C for 1 min for annealing and extension. The samples were analyzed in triplicate from the reverse-transcription step and normalized to the internal control, a housekeeping gene, β-actin. No template samples were used as controls. PCR data were analyzed by the ΔΔCt method described in detail by Livak et al.  15 

Flexor digitorum brevis muscle samples were taken from control rats and injured rats at 4 h, and 1, 2, 5, and 10 days after plantar incision. Muscle tissue from origin to insertion was harvested from each animal after they were euthanized with carbon dioxide. Immediately after extraction, the tissue samples were homogenized and lysed in ice-cold radioimmunoprecipitation assay buffer containing protease inhibitors. The homogenates were centrifuged, and the supernatant was taken. Total protein concentration in each sample was determined using a modified Lowry protein assay. The NGF concentration was measured using a commercially available assay per manufacturer's instructions (NGF Emax Immunoassay; Promega, Madison, WI). This assay exhibits very low crossreactivity with structurally related growth factors, and a minimum of 7.8 pg/ml NGF can be quantified.

After assay of total protein, 20 μg of protein in 20 μl of volume from each sample was electrophoresed on 12% sodium dodecyl sulfate-polyacrylaminde gels and then transferred to polyvinylidene difluoride membranes (Bio-Rad Lab, Hercules, CA). Overnight incubation of the membranes in rabbit anti-NGF polyclonal antibody (1:1000, Cat # sc-548; Santa Cruz Biotechnology Inc., Santa Cruz, CA) at 4°C was followed by incubation in peroxidase-conjugated donkey anti-rabbit IgG secondary antibody (1:10,000; Jackson ImmunoRes, West Grove, PA) for 4 h. Membranes were then developed with SuperSignal West Femto Maximum Sensitivity Substrate (Pierce Biotechnology Inc., Rockford, IL) and scanned by Kodak Image Station 440 scanner (Eastman Kodak Company, Rochester, NY). After the membranes were scanned, they were incubated in stripping buffer at 50°C for 30 min and probed for actin as a loading control. Immunoreactive NGF band intensity was quantified and normalized with actin bands by Image J software (NIH, Bethesda, MD). In a previous study, we demonstrated that the antibody was immunoreactive to mature rat NGF (N2513; Sigma, St. Louis, MO); no immunoreactive bands were noted when we added the blocking hapten peptide (cat. No. sc-548 P; Santa Cruz Biotechnology Inc.) containing the antigen (100 μg/0.5 ml) used to raise the antibody.9 

Briefly, rats were anesthetized with sodium pentobarbital (150 mg/kg) and transcardiacally perfused with 0.1 m phosphate-buffered saline followed by 4% paraformaldehyde. The glabrous plantar skin samples together with deep tissue and flexor muscle were removed from sham and incised hindpaws and postfixed in the same fixative for 4 h. After washing in gradually increasing concentrations of sucrose, samples were rapidly frozen in 2-methylbutane chilled at –80°C. Consecutive sections (10 μm) were prepared and frozen at –80°C until usage. NGF expression was visualized by NGF immunohistochemisty using the ABC method. The following antibodies were used: rabbit anti-NGF antibody (1:2000, Cat # sc-548; Santa Cruz Biotechnology Inc.) and biotinylated anti-rabbit IgG (1:500; Vector Laboratories, Burlingame, CA). For controls, some sections were processed after preabsorption of the anti-NGF antibody with the blocking peptide (4 pg/μl). Other sections were processed without the primary or the secondary antibody as controls.

Double-labeling confocal microscopy and immunofluorescent staining were performed for NGF and protein gene product 9.5 (PGP 9.5, a panaxonal marker), calcitonin gene-related peptide (CGRP, a peptidergic C-fiber marker), isolectin B4 (IB4, a nonpeptidergic C-fiber marker), or neurofilament 200 (N52, a myelinated A-fiber marker) to examine the colocalization of NGF and sensory fibers. To block nonspecific reactions, sections were incubated in 0.1 m phosphate-buffered saline containing 1.5% bovine serum albumin and 0.1% Tween-20 for 1 h. Sections were processed for NGF and PGP 9.5, CGRP, IB4, or N52 double-labeling. The following primary antibodies were used: rabbit anti-NGF polyclonal antibody (1:1000, Cat # sc-548; Santa Cruz Biotechnology Inc), guinea pig anti-PGP 9.5 polyclonal antibody (1:1000, Cat # AB5898; Chemicon, Temecula, CA), rabbit anti-CGRP polyclonal antibody (1:1000, Cat # T-4032; Peninsula Laboratories Inc, San Carlos, CA), biotinylated griffonia simplicifolia lectin I (IB4, 1:8000, Cat # B-1105; Vector Laboratories), mouse antineurofilament 200 monoclonal antibody (1:1000, Cat # N0142; Sigma). Sections were then incubated in the following secondary antibodies, Cy3-conjugated goat anti-rabbit IgG, Cy2-conjugated donkey anti-guinea pig IgG, Cy2-conjugated donkey anti-rabbit IgG, Cy2-conjugated streptavidin and Cy2-conjugated goat antimouse IgG. All the secondary antibodies were obtained from Jackson ImmunoResearch (West Grove, PA) and applied to sections with a concentration of 1:200. Sections were mounted and observed under Zeiss 510 confocal microscopy (Carl Zeiss MicroImaging, Inc., Thornwood, NY). Digital images were collected in the LSM 510 software (Carl Zeiss MicroImaging, Inc). For controls, sections were stained without primary or secondary antibodies.

Rat hind paw flexor digitorum brevis muscle was taken 1 h after incision. Total RNA was isolated using RNeasy mini kit (Qiagen). One microgram of RNA was used for RT-PCR to amplify a 416-bp band of NGF using primers: forward, 5′-CCATGGTACAATCTCCTTCA-3′; reverse, 5′:-TGCAGGCAAGTCAGCCTCTT-3′. The product was purified using QIAquick PCR purification kit (Qiagen) and then subcloned into T-easy vector (Promega). After sequencing, the template was linearized with Sal I digestion and gel purified using QIAquick Gel Extraction kit (Qiagen). Digoxigenin-labeled probes were made using DIG RNA Labeling Kit (Roche, Mannheim, Germany), and the labeling efficiency was quantified using DIG Luminescent Detection Kit (Roche). All procedures were performed following the manufacturers' instruction.

Sections were postfixed in 4% paraformaldehyde for 10 min and permeabilized by treatment with 0.002% proteinase K for 5 min. Fixation in 4% paraformaldehyde again for 5 min and underwent acetylation (0.5% acetic anhydride in 0.1 m triethanolamine) for 10 min. Probes were diluted to 10 ng/ml in hybridization buffer (50% formamide, 4 × standard saline citrate, 0.5 × Denhardt solution, 0.025% yeast tRNA, 0.025% PolyA, 0.025% salmon sperm DNA, 20% dextran sulfate, and 0.1 m dithiothreitol). The sections were hybridized in a humidified chamber at 55°C overnight with sense, antisense or internal control (actin) probes. Some sections were also incubated in a hybridization solution containing no probe. Adult rat cortex sections were hybridized with NGF sense or antisense probes as positive control because it is well accepted that NGF is produced by neurons in cortex.

After hybridization, sections were subjected to post-hybridization washes to remove unbound probes. Endogenous alkaline phosphatase was blocked with blocking reagent (Roche) dissolved in 1 × MABT (0.88% NaCl, 1.16% maleic acid, 0.72% NaOH) containing 0.01% Tween-20 and 20% goat serum for 60 min. Signal was detected by incubating sections with sheep antidigoxigenin-AP, Fab fragments antibody (1:2000, Cat# 11093274910; Roche) at 4°C overnight, visualized with alkaline phosphatase color substrates, nitroblue tetrazolium/5-Bromo-4-Chloro-Indolyl-Phosphatase for 2 h. The color reaction was stopped by immersing sections in 0.1 m phosphate-buffered saline. Sections were then counterstained in nuclear fast red for 1 min, dehydrated, and mounted for microscopic examination.

All statistical analyses were performed using Prism 4.0 (GraphPad Software, Inc., San Diego, CA). Data were expressed as mean ± SEM. Differences between means were determined by 1-way ANOVA followed by Dunnett post hoc  test for comparisons of incision versus  sham. Guarding pain scores were analyzed by ANOVA. Subsequent 1-way ANOVA for repeated measures and post hoc  t  tests compared the effect of drug and Dunnett test versus  baseline examined the effect of incision. P < 0.05 was considered statistically significant.

Incision of skin and treatment with vehicle resulted in a cumulative pain score of 4.3 ± 1.3 (fig. 1A; P < 0.05 vs.  baseline) 4 h after incision. There was no significant guarding thereafter. Treatment with the anti-NGF peptibody (100 mg/kg) and skin incision were not different than vehicle and skin incision.

In rats that underwent incision of skin, fascia, and muscle and were treated with vehicle, guarding was increased at 4 h, 1 day, and 2 days to 11.9 ± 0.6, 8.8 ± 1.2, 5.0 ± 0.7, respectively (fig. 1B; P < 0.05 vs.  baseline). Guarding after pretreatment with anti-NGF peptibody (100 mg/kg) and skin, fascia, and muscle incision was 8.1 ± 1.4, 4.6 ± 0.7, 2.1 ± 0.5, at 4 h, 1 day, and 2 days, respectively. The anti-NGF peptibody decreased guarding by skin fascia and muscle incision at 4 h, 1 day, and 2 days (P < 0.05 vs.  vehicle).

Compared to the β-actin control, NGF mRNA increased 6.6 ± 1.3-fold (P < 0.01 vs.  sham) in incised flexor muscle at 2 h after incision compared with sham control skin (fig. 2A; n = 6 per group). The level of NGF mRNA was not significantly increased on POD 1 and thereafter and was similar to the contralateral, nonincised hind paw flexor muscle.

As shown in figure 2B, basal NGF expression was 149.3 ± 35.9 ng/mg protein in sham-operated muscle. NGF levels increased to 614.3 ± 61.3 ng/mg (P < 0.01 vs.  sham) by 4 h and peaked at 923.4 ± 101.3 ng/mg on POD 1 in incised muscle. After POD 1, levels of NGF expression were 803.3 ± 47.3 ng/mg, and 725.1 ± 135.5 ng/mg on POD 2 (P < 0.01 vs.  sham, n = 6 per group), and POD 5 (P < 0.01 vs.  sham, n = 6 per group), respectively. NGF was not different than control on POD 10 (449.8 ± 131.6 ng/mg) and at the contralateral hind paw muscle (100.1 ± 35.1 ng/mg).

Western blot analysis revealed a single high–molecular weight (MW) NGF increased by incision. Representative blots were shown in figure 2Ctop. This large MW form of NGF protein, approximately 75 kDa, was increased in incised muscle. A summary quantifying the NGF immunoreactive bands for different times after incision is shown in figure 2Cbottom. The data were normalized to actin (n = 6 per group). In comparison to sham, NGF expression in the incised hind paw increased initially 2.4 ± 0.4-fold by 4 h to 7.8 ± 2.6-fold (P < 0.01 vs.  sham) on POD 1. Intensity of NGF expression was increased 7.0 ± 1.9-fold (P < 0.05 vs.  sham) on POD 2, 4.5 ± 1.3-fold on POD 5, and 3.1 ± 0.5-fold POD 10. NGF expression in the contralateral, unincised hind paw flexor muscle (1.0 ± 0.4-fold) was the same as sham.

There was no NGF immunoreactivity detected in sham-operated, control flexor muscle (fig. 3A). NGF immunoreactivity was found in apparent nerve-like structures in incised muscle on POD 1 (fig. 3B). Higher magnifications of the rectangular areas marked in figure 3Bare shown in figure 3C and 3D. NGF immunoreactive elements were localized within transverse and longitudinal sections of nerve-like structures (fig. 3, C and D). After preabsorption of the anti-NGF antibody with the hapten-blocking peptide, there was no NGF immunoreactivity detected in skin, fascia, and muscle (fig. 3E).

NGF and PGP 9.5, CGRP, or IB4 double-labeling was performed to confirm the immunoreactive NGF in nerve structures. NGF-positive staining was found in nerve axons labeled with PGP 9.5 in incised muscle 1 day after incision (fig. 4, A–A2). These NGF-positive axons were colocalized with CGRP-positive fibers (fig. 4, B–B2), but not IB4-positive non-peptidergic C-fibers (fig. 4, C–C2). NGF-positive staining could also be found in large nerve bundles in deep tissue, which innervates the flexor digitorum brevis muscle (fig. 5) and possibly other deep structures. In these bundles, NGF immunoreactivity was colocalized with CGRP-positive fibers (fig. 5, A–A2). There was no colabeling of NGF with IB4-positive C-fibers (fig. 5, B–B2), but colabeling was present on some N52-labeled myelinated axons (fig. 5, C–C2).

Adult rat brain cortex sections were subjected to in situ  hybridization with NGF antisense and sense probes to examine the specificity of the NGF probes (fig. 6). NGF mRNA signals (blue) could be observed in cortex (fig. 6A) by antisense probe hybridization, and most of the signals were localized in cytoplasm surrounding the nucleus (red) of neurons (fig. 6B). There was no NGF mRNA staining in cortex when hybridized with the sense probe (fig. 6C); only nuclear staining (red) was observed (fig. 6D).

NGF mRNA expression could be found in sections obtained from sham-incised hind paw (fig. 7A). The sources of NGF mRNA included keratinocytes in the epidermis and fibroblasts in dermis (fig. 7B). In deeper tissue, fibroblasts in dense connective tissue (fig. 7C), perimysium of muscle (fig. 7D), perivascular connective tissue (fig. 7E), and a large nerve bundle (fig. 7F) innervating the muscle were positive for NGF mRNA.

NGF mRNA labeling was observed in sections obtained at 2 h after plantar incision (fig. 8A) when NGF mRNA by RT-PCR was increased. In addition to structures labeled in the unincised tissue (fig. 8, B–F), fibroblasts in the subcutaneous region adjacent to the dissected tissue and the incision (fig. 8G) and fibroblasts adjacent to the muscle incision (fig. 8H) produced NGF mRNA. In situ  hybridization with NGF sense probe at 2 h after incision did not show NGF mRNA–positive signals (fig. 8I).

In this study, incision that included hind paw fascia and flexor digitorum brevis muscle produced guarding behavior that was reduced by blockade of NGF. NGF expression was examined in normal and in incised hind paw muscle. NGF mRNA was increased at 2 h after incision and returned to normal on POD 1. NGF protein increased 4 h after muscle incision, was increased for 5 days, and was not different than control on POD 10. A large MW form of NGF was produced and increased by incision. NGF immunoreactivity was not identified in normal muscle; however, NGF-immunoreactive structures were observed in nerve bundles from incised muscle. Further delineation of the immunoreactive structures revealed that NGF staining was present in CGRP- and N52-positive axons. IB4-positive axons did not stain for NGF. NGF mRNA was present in normal hind paw sections. Fibroblasts were the sources of NGF mRNA in perimysium, perivascular connective tissue, and nerve adjacent to the incised muscle.

NGF is an important pain mediator in skin,2but NGF and muscle have been studied to a lesser extent. NGF is synthesized in normal skeletal muscle tissue, but the cellular elements producing NGF are not known.16,17Studies by others indicate that NGF contributes to muscle healing after injury.18NGF is also upregulated in muscle after intramuscular formalin injection.19 

Previously conducted human studies indicate the potential role of NGF in muscle pain. Local injection of NGF into the masseter muscle produces long-term local signs of mechanical allodynia and hyperalgesia and is associated with pain during jaw movement.20In trials of parenterally administered NGF for patients with diabetic neuropathy, NGF did not improve symptoms. Rather, administration of NGF elicited whole-body myalgias in patients.4The duration and severity of these myalgias were dose-dependent.

Muscle nerves contain motor axons, sensory axons, and sympathetic axons; the sensory fibers are both myelinated and unmyelinated.21Our data have shown immunocytochemical labeling for NGF in CGRP-positive peptidergic axons that may also express the high-affinity NGF receptor trkA.22–24NGF labeling was not found in IB4-positive nonpeptidergic axons. In incisions, NGF was colabeled with the myelinated axon marker N52-positive in a few fibers. There is approximately 10% overlap between CGRP- and N52-positive axons; therefore, these NGF-positive, N52-positive fibers may also have the trkA receptor.25In animal studies, intramuscular injection of NGF excites primary afferent muscle nociceptors26and dorsal horn neurons transmitting nociception from muscle.27,28One study was not able to detect sensitization of nociceptors by NGF injected into rat masseter muscle.29Muscle afferents transmit both mechanosensitivity and chemosensitivity.30–33There is also evidence that muscle afferents have thermosensitivity.34 

NGF released by muscle incision likely influences the sensitivity of nociceptors to a variety of mediators, although no systematic studies of NGF and muscle sensory ion channels and receptors have been described. In other tissues, NGF regulates a number of ligand-gated ion channels, including purinergic receptor of ATP (P2X3), acid-sensing ion channel 3, transient receptor potential vanilloid 1, and G protein–coupled receptors, including bradykinin B2 receptors.35NGF also increases voltage-gated sodium channels, including tetradodotoxin-sensitive and tetradodotoxin-resistant sodium channels (Nav 1.3, 1.8, and 1.9). Whether these mechanisms through which NGF contributes to nociception in other tissues also influence nociception from muscle is not known. We speculate that NGF influences chemosensitivity of muscle afferents after incision, potentiating its responses to mediators like prostaglandins, acid, and lactate.

NGF is produced in several forms by a variety of tissues and cells depending on the conditions. A large MW form of NGF, approximately 75 kDa, was shown after muscle incision in the current study and after skin incision in a previous study.9Mature NGF is approximately 16 kDa; therefore, this large MW form is likely a glycosylated precursor, like preproNGF, or a dimer of a precursor form.36,37Isaacson et al.  37,38demonstrated that mature forms of the lower MW NGF proteins are uncommon in tissues, and each tissue exhibits a characteristic NGF expression pattern. In adults, muscle and skin have the same large MW form, and both increase in response to incision.

In vitro  studies of primary cell cultures indicate that NGF is produced by fibroblasts,39keratinocytes,40and T cells.41In inflammatory pain models, NGF is likely produced by macrophages42and mast cells43; after nerve injury, Schwann cells are a source of NGF.44 

In the current study, NGF mRNA was identified using in situ  hybridization in normal plantar skin and flexor muscle as well as in incised tissue. Fibroblasts from fascia and connective tissues surrounding muscle express NGF in the area injured by the incision. NGF mRNA was found in nonneuronal cells, whereas NGF protein was localized in nerve fibers. NGF expression by inflammatory cells was not prominent. The different NGF source in our incisional pain model compared with other inflammatory pain models supports the unique characteristics for incisional pain mechanisms versus  inflammatory pain. Our data are supported by another study of cutaneous wound healing,45which demonstrates that NGF is produced by keratinocytes and fibroblasts and contributes to cutaneous wound healing.

In previously performed studies in skin,9we demonstrated that NGF protein is present in the area surrounding incised skin. In transverse hind paw skin incisions, immunoreactive NGF was prominent in distal axons in skin but not found in axons proximal to the incision. On the basis of these previous results, the current study, and work by others,46,47we propose the following model. NGF is synthesized in fibroblasts and connective tissue in incisions, is retrogradely transported to the dorsal root ganglia in intact axons, and sensitizes these nociceptors. The distal portions of cut axons take up NGF but do not transport it retrogradely.46These distal portions of the cut axons are immunoreactive for NGF. The colabeling of NGF with CGRP indicates that these are Trk A–positive axons. In incised skin, NGF produces heat sensitization.10In deep muscle incision, we propose that NGF produces spontaneous activity resulting in ongoing pain.

In summary, NGF contributes to guarding pain behaviors caused by deep muscle incision. In incised muscle, NGF mRNA increases 2 h after incision, and increased protein expression is evident for several days. Immunohistochemical studies of incisions reveals NGF protein in CGRP-positive but not IB4-positive nerve fibers. Fibroblasts adjacent to the injury are sources of NGF in incised muscle. NGF is not only important for cutaneous pain mechanism but also for pain from incised muscle.

The authors thank Alberto Subieta, B.S., Research Assistant, Department of Anesthesia, University of Iowa, Iowa City, Iowa, for technical assistance.