This investigation used capsaicin to selectively lesion unmyelinated sensory neurons in rats. Neuronal lesioning induced a loss of trabecular integrity, reduced bone mass and strength, and depleted neuropeptides in nerve and bone. These data suggest that capsaicin-sensitive sensory nerves contribute to trabecular bone integrity.
Introduction: Familial dysautomia is an autosomal recessive disease in which patients suffer from unmyelinated sensory neuron loss, reduced BMD, and frequent fractures. It has been proposed that the loss of neurotransmitters synthesized by unmyelinated neurons adversely affects bone integrity in this hereditary syndrome. The purpose of this study was to determine whether small sensory neurons are required for the maintenance of bone integrity in rats.
Materials and Methods: Ten-month-old male Sprague-Dawley rats were treated with either capsaicin or vehicle. In vivo DXA scanning and μCT scanning, and histomorphometry were used to evaluate BMD, structure, and cellular activity. Bone strength was measured in distal femoral sections. Body weight and gastrocnemius/soleus weights were measured and spontaneous locomotor activity was monitored. Peroneal nerve morphometry was evaluated using light and electron microscopy. Substance P and calcitonin gene-related peptide (CGRP) content in the sciatic nerve and proximal tibia were determined by enzyme immunoassay (EIA). Substance P signaling was measured using a sciatic nerve stimulation extravasation assay.
Results: Four weeks after capsaicin treatment, there was a loss of BMD in the metaphyses of the tibia and femur. In the proximal tibia, the osteoclast number and surface increased, osteoblast activity and bone formation were impaired, and trabecular bone volume and connectivity were diminished. There was also a loss of bone strength in the distal femur. No changes occurred in body weight, 24-h grid-crossing activity, weight bearing, or muscle mass after capsaicin treatment, indicating that skeletal unloading did not contribute to the loss of bone integrity. Capsaicin treatment destroyed 57% of the unmyelinated sensory axons, reduced the substance P and CGRP content in the sciatic nerve and proximal tibia, and inhibited neurogenic extravasation.
Conclusion: These results support the hypothesis that capsaicin-sensitive sensory neurons contribute to the maintenance of trabecular bone integrity. Capsaicin-sensitive neurons have efferent functions in the tissues they innervate, effects mediated by transmitters released from the peripheral nerve terminals. We postulate that the deleterious effects of capsaicin treatment on trabecular bone are mediated by reductions in local neurotransmitter content and release.
FAMILIAL DYSAUTONOMIA IS an autosomal recessive disease occurring primarily among Ashkenazi Jews. These patients suffer from a loss of unmyelinated axons with reduced BMD and increased bone fragility.(1, 2) Neuropeptides such as substance P and calcitonin gene-related peptide (CGRP) are synthesized in unmyelinated sensory neurons and released from their peripheral terminals. Based on evidence of impaired neuropeptide signaling and content in this condition,(3, 4) it has been proposed that decreased local neuropeptide levels in bone adversely affect bone integrity in this hereditary sensory neuropathy.(2, 4)
The aim of this study was to define the contribution of capsaicin-sensitive sensory neurons to the maintenance of bone integrity in skeletally mature rats. Capsaicin, the pungent ingredient in hot chili peppers, activates the vanilloid receptors (TRPV1) that are expressed by most unmyelinated and some small diameter myelinated sensory neurons.(5) Capsaicin activation of the TRPV1 causes an influx of calcium and sodium cations into the sensory neuron, triggering an excitotoxic effect. Vanilloid receptors are not observed on motor neurons, large diameter sensory neurons, or sympathetic neurons.(6, 7) After capsaicin injection in neonatal rat pups (50 mg/kg, SC), the unmyelinated and small-diameter myelinated sensory neurons are destroyed, but the motor, large sensory afferent, and sympathetic fiber functions are unaffected.(8–11) Capsaicin treatment in neonatal and adult rats depletes substance P and CGRP in peripheral nerves, but has no effect on neuropeptide levels in the central nervous system.(9, 12–17) We postulated that a loss of neuropeptide signaling in trabecular bone would disrupt bone metabolism with adverse effects on bone mass, structure, and strength.
In this study, systemic capsaicin treatment was used to induce a selective lesion of the small-diameter sensory fibers in skeletally mature rats. BMD, trabecular structure, histomorphometry, and metaphyseal strength were evaluated after capsaicin treatment. To confirm that there was no loss of motor function or evidence of skeletal unloading after the capsaicin treatment, we evaluated the body weight, 24-h grid-crossing activity, hindpaw weight bearing, hindlimb muscle mass, and peroneal nerve morphometry. The effects of capsaicin treatment on neuropeptide content and signaling in nerve and bone were also determined. The objective of this investigation was to determine whether capsaicin-sensitive sensory neurons contribute to the maintenance of normal bone cellular activity, density, structure, strength, and neurotransmitter content.
MATERIALS AND METHODS
Our institute's Subcommittee on Animal Studies approved these experiments. Skeletally mature (10 month old) male Sprague-Dawley rats (Harlan, Fremont, CA, USA) were used in all experiments. The animals were housed individually in sterile isolator cages with solid floors covered with 3 cm of soft bedding and were fed and watered ad libitum. During the experimental period, the animals were fed Laboratory Diet 5012 (PMI Nutrition Institute, Richmond, IN, USA), which contains 1.0% calcium, 0.5% phosphorus, and 3.3 IU/g of vitamin D3.
Pilot studies indicated that capsaicin injection in 10-month-old rats caused pulmonary edema, resulting in a 30–40% mortality rate within 12 h after treatment. To prevent pulmonary edema, the rats were denied drinking water for 6 h before capsaicin injection, and after each injection, were kept for 3 h in an airtight chamber continuously perfused with 100% oxygen. This protocol eliminated capsaicin-induced mortality in the aged rats. Immediately before injection, the capsaicin was sonicated in vehicle until homogenously suspended, and the rats were injected subcutaneously over the dorsal spine with either capsaicin (Sigma, St Louis, MO, USA) or vehicle (10:10:80 v/v of Tween-80, ethanol, saline). The needle was left in the skin for 60 s after injection to ensure that the injectant did not reflux back out the needle tract. The capsaicin treatment protocol was to initially inject 25 mg/kg, 8 h later inject 50 mg/kg, and 24 h later inject 50 mg/kg. Two weeks later, the capsaicin treatment protocol was repeated. Each rat received a total capsaicin dose of 250 mg/kg, and all experiments were performed at 2 weeks after the second treatment (4 weeks after first treatment). Control animals underwent the same the injection protocol with vehicle. All in vivo tests were performed at baseline and at 4 weeks after the initiation of treatment, and comparisons were made within groups between baseline and week 4. All ex vivo tests were only performed at week 4 with comparisons between treatment groups.
BMD was measured in vivo by DXA scanning, using a Hologic (Waltham, MA, USA) QDR-4000 instrument adapted to measurement in small animals. Before scanning, the rats were anesthetized with dexmedetomidine (300 μg/kg, IP) and taped into position in clear plastic boxes filled with 3 cm of warm water. The animals were positioned on their backs with the hindlimb externally rotated, and the hip, knee, and ankle articulations placed in 90° flexion. The femur and tibia were scanned, and the fibula excluded from the region of interest. Both the femur and tibia were divided into three regions of interest, the proximal and distal metaphyseal regions (each 7 mm in length) and the diaphysis.
In vivo imaging of the right proximal tibia was performed at baseline and at week 4 with a Scanco VivaCT 40 μCT scanner (Scanco Medical, Basserdorf, Switzerland). Each measurement was derived by averaging two scans performed sequentially, with the rat's hindlimb repositioned between scans. Scanning was initiated 1 mm below the proximal tibia growth plate, and a total of 59 consecutive 38-μm-thick sections was analyzed, encompassing a 2.24 mm length of the secondary spongiosa. Cortical bone was excluded from the region of interest with semiautomatically drawn contours. The segmentation values were kept constant at 0.8/1/130. Relative bone volume (BV/TV), trabecular number (Tb.N), thickness (Tb.Th), and separation (Tb.Sp) were calculated by measuring 3D distances directly in the trabecular network and taking the mean over all voxels. The connectivity density (Conn.D) based on the Euler number was also determined. By displacing the surface of the structure by infinitesimal amounts, the structure model index (SMI) was also calculated. The SMI quantifies the plate versus rod characteristics of trabecular bone, in which an SMI of 0 represents a purely plate-like bone and an SMI of 3 indicates a purely rod-like structure.
All animals were subcutaneously injected with calcein (10 mg/kg) at 13 and 5 days before necropsy. The right proximal tibias were dehydrated in ethanol, embedded undecalcified in methylmethacrylate, and sectioned longitudinally with a Leica/Jung 2255 microtome at 4- and 8-μm-thick sections. The 4-μm sections were stained with von Kossa and toluidine blue for collection of bone mass and architecture data with the light microscope, whereas the 8-μm sections were left unstained for measurements of fluorochrome-based indices. Static and dynamic histomorphometry was performed using a semiautomatic image analysis OsteoMeasure system (OsteoMetrics, Decatur, GA, USA) linked to a microscope equipped with transmitted and fluorescence light.
A counting window, measuring 8 mm2 and containing only trabecular bone and bone marrow, was created for the histomorphometric analysis. Static measurements included total tissue area (T.Ar), bone area (B.Ar), bone perimeter (B.Pm), osteoclast surface (Oc.S), and osteoclast number (N.Oc). Dynamic measurements included single- (sL.Pm) and double-labeled perimeter (dL.Pm) and interlabel width (Ir.L.Wi). These indices were used to calculate bone volume (BV/TV), mineralizing surface (MS/BS), and mineral apposition rate (MAR). Finally, surface-based bone formation rate (BFR/BS) and bone volume-based bone formation rate (BFR/BV) were calculated by multiplying mineralizing surface (single-labeled surface/2 + double-labeled surface) with MAR according to Parfitt et al.(18)
After death, the femur was immediately dissected free and stored wrapped in saline-soaked gauze at −20°C. A planoparallel section ∼3.9 mm was sawed from the distal part of the metaphysis just proximal to the patellofemoral joint cartilage using a precision saw (Isomet Plus; Buehler; Lake Bluff, IL, USA).(19) The whole bone density and volume of each section were determined using an electronic balance (Mettler AE 240; Mettler-Toledo, Columbus, OH, USA) equipped with a density determination kit.(18) The height of each section was measured using digital calipers (CD-6 CS; Mitutoyo, Aurora, IL, USA). The cross-sectional area was calculated by dividing the volume by the section height. The distal femoral sections were tested in a servohydraulic materials testing system (MTS 858; MTS, Eden Prairie, MN, USA) incorporating a load cell (MTS 662.10A-02) with a measurement precision of 4.4 N. The bone sections were axially compressed at a constant displacement rate of 2 mm/min, and both displacement and load were recorded for later analysis. Ultimate force was determined from the load-displacement curves, and ultimate stress was calculated by dividing the ultimate force by the cross-sectional area.
Spontaneous locomotor activity
Rats were individually placed in clear plastic monitoring chambers each measuring 72 × 32 × 32 cm. The chamber floors were covered with 0.5 cm of soft bedding, and the rats were fed and watered ad libitum while in the chamber. Spontaneous locomotor activity was measured through seven sets of photoelectric sensors evenly spaced along the length of the monitoring chamber at a level 3 cm above the floor (San Diego Instruments, San Diego, CA, USA). Total grid-crossing activity was recorded as the number of time the rat interrupted the photoelectric sensors during a 24-h monitoring session. The testing room was lighted for 10 h (8:00 a.m. to 6:00 p.m.) and dark for 14 h during the monitoring session, and the room temperature was maintained at 23°C.
Hindpaw weight bearing
Rats were individually placed into the entrance of an open-top narrow wooden chute (9.5 cm wide, 28.5 cm high, 86 cm long), and the entrance was closed. The rat walked the length of the level wooden chute to the opposite end that opened into a dark box enclosure. In the middle of the chute was a plastic plate (4.75 cm wide and 6 cm long) flush with the floor. The plate was attached to a commercial strain gauge (Salter, Fairfield, NJ, USA) that was connected to a computer chart recorder (MacLab8e; ADInstruments, Castle Hill, Australia). Clear windows on both sides at the middle of the chute allowed observation of the plate. When the rat's hindpaw was placed on the plate during the stance phase of gait, the peak weight was recorded. The mean of three consecutive tests was used for the maximum weight bearing value, and this number was converted to a percentage of the total body weight.
Gastrocnemius and soleus weights
At the end of the experiment, the gastrocnemius and soleus muscles were excised, and their combined wet weight was determined.
At 4 weeks after capsaicin or vehicle treatment, the rats were anesthetized with isoflurane, and 2 ml of blood was withdrawn by transcardial puncture. The serum was collected and immediately tested on a Beckman LX 20 Analyzer for calcium, phosphorus, creatinine, albumin, total protein, magnesium, alkaline phosphatase, alanine transaminase, and aspartate transaminase.
Under isoflurane anesthesia, the common peroneal nerve was excised and immediately fixed in isotonic glutaraldehyde for 24 h, postfixed in 2% osmium, uranyl acetate en bloc stained, dehydrated, and embedded in LX resin. Thick sections were cut at 0.5 um and stained with toluidine blue. Thin sections were placed on coated copper slot grids and stained with lead citrate for viewing on a Hitachi H300 electron microscope. The entire cross section of each peroneal nerve was imaged at a 150× magnification, and this digital image was enlarged to 660× to determine the total myelinated axon count. Approximately 35–40% of the entire peroneal nerve cross-sectional area was also digitally imaged using 16–18 nonoverlapping electron microscope images at a magnification of 1000×. These images were enlarged to 9400× to determine the total unmyelinated axon counts and to 2400× to determine myelinated axon area measurements. The nerve and axon areas were measured using RT SPOT software (Diagnostic Instruments, Sterling Heights, MI, USA), and the axon diameters were calculated from the area. The total unmyelinated axon count for each nerve was calculated from ratio of the area encompassed by the electron microscope images compared with the cross sectional area of the nerve.
Enzyme immunoassay assay for neuropeptide concentration of bone and nerve
Under isoflurane anesthesia, the sciatic nerve (∼3 cm, 32–46 mg) and the proximal tibia (∼7 mm, 282–380 mg) were collected, immediately frozen, and weighed. Nerve samples were minced in 1 ml of 3:1 ethanol/0.7 M HCl and sonicated for 20 s. Bone samples were minced in 2 ml of the same solution and were homogenized for 20 s. The samples were shaken for 2 h at 4°C and centrifuged at 3000g for 20 minutes at 4°C. The supernatant was frozen and lyophilized, and the lyophilized product was stored at −80°C. The substance P and CGRP content of the samples was assayed in duplicate using an enzyme immunoassay (EIA) kit (substance P, 900–018; Assay Designs, Ann Arbor, MI, USA, and CGRP (rat), 589001; Caymen Chemical, Ann Arbor, MI, USA).
Neurogenic extravasation procedures
Under isoflurane anesthesia, the sciatic nerve was exposed in the thigh and tightly ligated proximal to the midthigh stimulation site. The incision was filled with warm mineral oil, and a Plexiglas-platinum stimulating electrode with a sliding jaw was gently secured around the nerve (Harvard Apparatus, Holliston, MA, USA). Evans blue dye (50 mg/kg; Sigma) was administered intravenously in a 50-mg/ml solution of 0.9% saline. Ten minutes after dye injection, the left sciatic nerve was stimulated for 5 minutes (5 Hz, 0.5-ms pulse duration, 10 mA). Ten minutes later, the rats were transcardially perfused with 1000 ml of saline (0.9%), suspended 100 cm above the heart (50 mm Hg perfusion pressure). The plantar and dorsal skin on each hindpaw was excised from the base of the heel to the tip of the third digit, and the tissue was placed in 4 ml of 99% formamide in a shaker bath at 55°C for 72 h. The dye concentration was determined spectrophotometrically at a wavelength of 620 nm.
Differences within treatment groups were determined using a paired t-test, and between-group differences were determined using the unpaired Student's t-test. All data are presented as the mean ± SE, and differences are considered significant at p < 0.05.
Reduction in BMD
After baseline DXA scanning of the tibia and femur, all rats were treated with either capsaicin (250 mg/kg, SC, n = 10) or vehicle (n = 8), and changes in BMD were determined 4 weeks later. Figure 1 shows that BMD in the capsaicin-treated animals declined by 8% in the proximal tibia, by 3% in the distal tibia, by 5% in the proximal femur, and by 9% in the distal femur. No bone loss was observed in the diaphyses. No significant changes in BMD occurred in any of the regions of interest after injection of vehicle.
Trabecular structural changes measured by μCT
Table 1 shows the effects of capsaicin treatment on the proximal tibia trabecular bone architecture. BV/TV was reduced by 28%, and Tb.Th was reduced by 13% at week 4 compared with baseline values (p < 0.01). Conn.D decreased by 28% (p < 0.05), and the SMI increased by 19% (p < 0.05), indicating a transformation from plate- to rod-like trabecular morphology. No significant changes over time were observed in any structural index after injection of vehicle.
Table Table 1. Proximal Tibia Trabecular Bone Evaluated by μCT
Figure 2 shows representative stained and unstained proximal tibial sections collected 4 weeks after capsaicin or vehicle treatment, whereas Table 2 presents a summary of the histomorphometric data. All comparisons were made between the capsaicin and vehicle treatment groups. The proximal tibia BV/TV was 19% less in the capsaicin- versus the vehicle-treated group, but this failed to reach significance because of the large variance of the vehicle treatment group values. Osteoclast surface (Oc.S) and number (N.Oc) were increased in the capsaicin group compared with vehicle-treated controls (101% and 75%, respectively, p < 0.05 for both). Both sL/Pm and dL/Pm were reduced in the capsaicin treatment group versus the vehicle-treated controls (31%, p < 0.05 and 67%, p < 0.01, respectively), an indication of diminished osteoblast activity. Furthermore, the BFR/BS was reduced by 44% in the capsaicin group versus the vehicle-treated cohort (p < 0.05).
Reduction in bone strength
Figure 3A shows that capsaicin treatment reduced the distal femur whole BMD by 7% (1.433 ± 0.012 g in capsaicin-treated, n = 11 versus 1.542 ± 0.040 g for controls, n = 10, p < 0.05). Axial loading of distal femur sections showed a 35% reduction in ultimate load values in the capsaicin cohort (532 ± 27N in capsaicin treated versus 821 ± 66N for controls, p < 0.001; Fig. 3B). Capsaicin treatment also reduced the ultimate stress values in the distal femur by 41% (24.7 ± 1.4 MPa in capsaicin-treated versus 42.3 ± 6.0 MPa for controls, p < 0.01; Fig. 3C).
No evidence of skeletal unloading or altered serum chemistry
Capsaicin treatment did not cause weight loss over the 4-week study, and in fact, there was a 2% weight gain in the capsaicin-treated rats (n = 10) compared with a 5% gain in controls (n = 10; Fig. 4A). At 4 weeks, the capsaicin-treated rats (n = 10) had a nonsignificant increase in spontaneous locomotor activity over a 24-h period (7870 ± 786 grid-crossings in the capsaicin-treated, n = 10 versus 6749 ± 374 grid-crossings in controls n = 10; Fig. 4B). The maximum hindpaw weight bearing, as a percentage of total body weight, was identical for the capsaicin (n = 11) and control (n = 7) cohorts at 4 weeks (51 ± 3%, Fig. 4C). At 4 weeks, there was no significant difference in the mean calf muscle mass (gastrocnemius and soleus) of the capsaicin- and vehicle-treated cohorts (3.54 ± 0.23 g in capsaicin-treated, n = 8 versus 3.71 ± 0.19 g in controls, n = 7; Fig. 4D).
Table 3 presents the serum chemistry for the capsaicin (n = 8) and vehicle (n = 10) treatment cohorts at week 4. No differences were observed between treatment groups for any variable. The albumin levels were low in both treatment groups because of the dye binding method used in the Beckman LX 20 Analyzer for albumin determination. This assay is standardized for human samples, and with rat albumin, this methodology results in a bias toward lower values. The salient point is that albumin levels were similar in the capsaicin and vehicle treatment groups (1.13 ± 0.05 versus 1.22 ± 0.07 g/dl, respectively).
Table Table 3. Serum Chemistry After Capsaicin Treatment
Changes in nerve morphometry
Capsaicin treatment caused a 41% reduction in unmyelinated axon counts in the common peroneal nerve (2344 ± 334 versus 3946 ± 119 for controls; Figs. 5A and 6). The unmyelinated axon counts observed in the control rats were similar to those reported by Schmalbruch (3946 versus 4171, respectively).(20) Schmalbruch also compared the axon counts of sympathectomized rats to those of controls and calculated that there were 1125 sympathetic unmyelinated axons in the rat peroneal nerve. Sympathetic neurons do not express TRPV1 receptors and are insensitive to capsaicin treatment.(6, 10) To determine the effect of capsaicin treatment on just the sensory unmyelinated fibers, we calculated the total unmyelinated axon counts in the control and capsaicin-treated rats after removal of the estimated sympathetic axon component and found that capsaicin treatment reduced the number of unmyelinated sensory axons by 57%.
Capsaicin also reduced the peroneal nerve myelinated axon counts by 8% (1915 ± 128 versus 2075 ± 75 for controls; Fig. 5A). The frequency histogram of the myelinated axon diameters for the capsaicin and vehicle treatment groups did not show a difference in the axon diameter distributions (Fig. 5B).
Reduced neuropeptide content and signaling
Substance P content in the sciatic nerve was reduced at 4 weeks after capsaicin treatment by 78% (3.48 ± 0.39 pg/mg in capsaicin-treated, n = 10 versus 16.16 ± 2.03 pg/mg in controls, n = 18, p < 0.001; Fig. 7A). This 78% reduction in the substance P content in the sciatic nerve is probably the basis for the 64% reduction in sciatic nerve stimulation evoked extravasation response observed at 4 weeks after capsaicin treatment (52 ± 9 ng dye/mg skin in capsaicin treated, n = 8 versus 138 ± 36 ng dye/mg skin in controls, n = 13, p < 0.001; Fig. 7B). Capsaicin treatment also reduced substance P content in the proximal tibia by 68% (0.50 ± 0.17 pg/mg in capsaicin-treated, n = 10 versus 1.55 ± 0.21 pg/mg in controls, n = 17, p < 0.001; Fig. 7C). Similarly, the CGRP content of the sciatic nerve was also reduced by 60% at 4 weeks after capsaicin treatment (41.9 ± 7.9 pg/mg in capsaicin-treated, n = 13 versus 105.1 ± 15.4 pg/mg in controls, n = 10, p < 0.01; Fig. 7D). Capsaicin treatment also reduced the CGRP levels in the proximal tibia by 57% (0.23 ± 0.02 pg/mg in capsaicin-treated, n = 10 versus 0.54 ± 0.08 pg/mg in controls, n = 13, p < 0.01; Fig. 7E).
The first aim of this investigation was to characterize the effects of capsaicin treatment on bone in the skeletally mature rat. A reduction in metaphyseal BMD was observed in the femur and tibia at 4 weeks after capsaicin treatment, but no change occurred in the diaphyses (Fig. 1). Trabecular volume, thickness, and connectivity all declined with capsaicin treatment, and the SMI increased, indicating a transformation to a more rod-like conformation (Table 1). The loss of trabecular connectivity and the shift to a more rod-like trabecular structure indicates that capsaicin treatment induced trabecular resorption and perforation; data consistent with the histomorphometric evidence of increased the osteoclast surface and number after capsaicin treatment (Table 2). The reduced single- and double-labeled perimeter and diminished bone formation rate observed after capsaicin treatment indicate impaired osteoblast activity (Table 2). Biomechanical testing showed a reduction in distal femur ultimate stress after capsaicin treatment, an indication of increased bone fragility (Fig. 3C).
Table Table 2. Proximal Tibia Trabecular Bone Evaluated by Histomorphometry
Because motor neurons do not express TRPV1 receptors and are insensitive to capsaicin,(7, 11) we anticipated that the capsaicin-treated rats would show normal locomotor activity and hindpaw weight bearing. There was no evidence that skeletal unloading contributed to the loss of BMD and strength observed after capsaicin treatment. The capsaicin-treated rats gained weight, had increased locomotor activity, had no reduction in hindpaw weight bearing, and had no loss in calf muscle mass (Fig. 4). Furthermore, there is little evidence that capsaicin treatment affects other organ systems. Capsaicin treatment in rats has no effect on exercise tolerance, basal cardiac function, blood pressure, renal and hepatic blood flow, urine output, glomerular filtration rate, renal sodium output, intestinal lipid absorption and transport, and plasma levels of insulin, glucose, norepinephrine, and epinephrine.(21–25) In addition, we failed to observe any changes in serum calcium, phosphorus, magnesium, creatinine, albumin, total protein, alkaline phosphatase, alanine transaminase, or aspartate transaminase levels after capsaicin treatment in skeletally mature rats (Table 3).
We were also concerned that capsaicin lesioning might indirectly induce bone loss by facilitating sympathetic signaling in bone. There is evidence that β-adrenergic receptor activation can induce bone loss in mice,(26) and neonatal capsaicin lesioning in rats can induce sympathetic neuron sprouting in some peripheral tissues, although this has not been examined in adult rats or in bone.(27–29) To determine whether the sympathetic nervous system regulates bone balance, we used 6-hydroxydopamine (6-OHDA) injections to induce a chemical sympathectomy in skeletally mature rats. After baseline DXA scanning, the rats were injected with 6-OHDA (400 mg/kg IP in divided doses over 5 days). We previously observed that this regimen dramatically reduced skin norepinephrine content for at least 4 weeks.(30) Repeat DXA scanning after 4 weeks failed to show any changes in tibial or femoral metaphyseal or diaphyseal BMD (data not shown), evidence that altered sympathetic neuron signaling did not mediate the changes we observed in metaphyseal BMD after capsaicin treatment.
Collectively, these results indicate that capsaicin-sensitive sensory neurons contribute to skeletal homeostasis and that lesioning these neurons caused enhanced bone resorption, a reduction in new bone formation, a subsequent loss of trabecular connectivity and thickness, and ultimately an increase in bone fragility. Capsaicin-sensitive sensory neurons have efferent functions in a wide variety of tissues, effects that are mediated by the peripheral release of neuropeptides. Because capsaicin treatment reduced the substance P and CGRP levels in nerve and metaphyseal bone, we postulated that these neurotransmitters may play a role in bone metabolism. Recently we observed that, after unilateral sciatic nerve transection in rats, there was widespread rapid trabecular bone loss in both hindlimbs, which was exacerbated by chronic administration of the substance P NK1 receptor antagonist LY303870.(31) There was no evidence of disuse of the contralateral hindlimb. However, the contralateral neurogenic extravasation response was diminished, and there was a contralateral reduction in bone substance P content, evidence of a transmedial loss of substance P signaling after nerve injury. These experiments suggest that the inhibition of substance P signaling can have deleterious effects on bone mass.
Neuropeptide regulation of bone metabolism has been previously proposed based on anatomic and cell culture data.(32, 33) Substance P and CGRP immunoreactive axons have been observed in the periosteum, epiphysis, trabecular bone, and bone marrow, whereas innervation of cortical bone is sparse.(34–39) The paucity of cortical neurons suggests that neural-osseal signaling does not play a role in cortical bone metabolism and may explain why there was no loss of diaphyseal bone after capsaicin treatment (Fig. 1). The neurons containing substance P and CGRP usually enter the bone in association with a blood vessel but dissociate from the vessel and terminate as a free ending in the marrow.(33, 39, 40) Whereas fast-acting “classical” neurotransmitters such as acetylcholine, glutamate, and GABA act on postsynaptic receptors immediately after their synaptic release, signal transduction by neuropeptides is nonsynaptic, diffuse, and slow. Neuropeptides are stored for extended periods in dense-core secretary vesicles preferentially located at sites distant from active zones of synapses and are exocytosed at nonsynaptic sites, where they slowly spread through a much greater extracellular volume than the synaptic cleft before encountering their cognate receptors.(41–43) Exocytosis occurs both constitutively and in response to physiologic signals. Substance P and CGRP secreted from skeletal neurons in the metaphyses are in a favorable position to regulate metabolic activities in local bone cells. Immunocytochemical studies have observed substance P (NK1) receptors on the plasma membrane and in the cytoplasm of osteoclasts, osteoblasts, and osteocytes,(37) and RT-PCR has been used to show NK1 receptor mRNA expression in osteoclasts.(44) Furthermore, substance P can increase osteoblastogenesis, bone formation, H3-proline incorporation, and protein accumulation in bone cell cultures.(32, 45, 46)
CGRP inhibits bone resorption in vitro,(47–51) increases cyclic AMP in osteoblasts,(52–54) and increases intracellular calcium in osteoblasts.(55) CGRP also increases the size and number of bone colonies in bone marrow stromal cell cultures.(56, 57) RT-PCR has been used to show CGRP receptor mRNA expression in osteoblasts and osteoclasts.(44) Moreover, administration of CGRP partially inhibits bone loss in ovariectomized rats.(58) Collectively, the in vitro data support the hypothesis that the release of sensory neurotransmitters such as substance P and CGRP in bone could directly regulate cell metabolism by promoting bone formation and inhibiting resorption. The histomorphometry findings in the capsaicin-treated rats show that a similar shift in bone cellular activity occurred after sensory lesioning, with impaired osteoblast activity, reduced bone formation, increased osteoclast numbers and surface, and increased bone resorption.
Two previous studies have examined the effects of capsaicin treatment on periosteal surface resorption induced by maxillary molar extraction.(59, 60) Capsaicin treatment dramatically reduced substance P and CGRP innervation in the mandibular periosteum and modestly inhibited the increase in osteoclast numbers and resorption area induced by molar extraction. We postulate that the antiresorptive effect of capsaicin lesioning in the molar extraction model may be the consequence of a loss in substance P signaling in the cellular inflammatory cascade that develops after molar extraction. Molar movement or extraction induces T-cell activation and cytokine expression that could mediate bone remodeling.(61, 62) T-cells express NK1 receptors, and substance P has direct stimulatory effects on T-cell migration and function, including the production of TNF-α, interleukin 1-β, and interleukin 6.(63–66) T-cells and the inflammatory cytokines they produce play a central role in osteoclastogenesis and bone resorption.(67, 68) Collectively, these data suggest that a reduction in mandibular substance P signaling might inhibit cytokine release from T-cells with a consequent reduction in molar extraction induced resorption.
We have used selective lesioning of the unmyelinated sensory neural pathway to determine the role of the capsaicin sensitive sensory afferents in the maintenance of normal bone balance in skeletally mature rats. Destruction of sensory neurons caused a loss of trabecular integrity and bone mass and strength, and also reduced nerve and bone neuropeptide content. Familial dysautonomia,(2) nerve injuries,(31, 69) glucocorticoid use,(70–72) and aging(73–75) are all conditions associated with bone loss, increased bone fragility, and impaired neuropeptide signaling. The signal transduction pathways by which these heterogeneous interventions cause bone loss are unknown. A parsimonious explanation would be that these various interventions inhibit sensory signaling in bone resulting in the development of osteoporosis. If future experiments confirm this hypothesis, the restoration of neuropeptide signal would be a novel therapeutic target for the treatment of osteoporosis.
The authors thank Marilyn Masek in the Laboratory of Neuropathology at Stanford University School of Medicine, Elijah Min and Anthony Joseph in the Physical Medicine and Rehabilitation Service at the Veterans Affairs Palo Alto Health Care System, and Ron van Groningen in the Pathology and Laboratory Medicine Service at the Veterans Affairs Palo Alto Health Care System for excellent technical support. This work was supported National Institutes of Health Grant R01 GM65345.