Neurons responsive to stimulation of the temporomandibular joint (TMJ) region were recorded from superficial laminae at the trigeminal subnucleus caudalis/upper cervical cord (Vc/C2) junction region of cycling female rats under barbiturate anesthesia. To determine if receptive field (RF) properties or sensitivity to algesic chemicals of TMJ units vary over the estrous cycle, animals were selected from proestrous (high estrogen) or early diestrous (low estrogen) stages. More than 90% of TMJ units from each group received convergent nociceptive input [wide dynamic range (WDR) or nociceptive specific (NS)-like] from facial skin. The cutaneous high-threshold RF areas of WDR units from proestrous rats were 30% larger than diestrous units, while RF areas of NS units were similar. Bradykinin (BK, 0.1–10 μM) injection into the TMJ region excited a high percentage of units (>80% of total) from both groups in a dose-related manner. However, BK-evoked response magnitude (R mag, +140%) and duration (+64%) were greater for proestrous than diestrous units. Both WDR and NS-like TMJ units of proestrous females displayed enhanced BK-evokedR mag values and response duration. Glutamate or mustard oil excitation of TMJ units was not affected by stage of the estrous cycle. Several TMJ units from proestrous and diestrous females were activated antidromically from the contralateral posterior thalamus, indicating that projection and nonprojection units were included in the sample population. These results were consistent with the hypothesis that factors related to stage of the estrous cycle modify the processing of deep craniofacial inputs by superficial dorsal horn neurons at the spinomedullary junction, a key region for the initial integration of sensory signals from the TMJ.
Temporomandibular disorders (TMD) comprise a family of conditions that often present with diffuse pain in the temporomandibular joint (TMJ) region and masticatory muscles that spreads to adjacent areas of the face and neck (Denucci et al. 1996; Dworkin and LeResche 1992; Turp et al. 1998). Although the basis for TMD pain is not well defined, the signs and symptoms suggest a prominent role for central neural mechanisms. TMD patients display lowered sensory thresholds, altered temporal summation to experimentally induced pain (Maixner et al. 1995, 1998), and a reduced capacity to recruit endogenous pain inhibitory controls (Bragdon et al. 2002).
Women are over-represented for TMD (Bush et al. 1993;LeResche 1997; Stohler 1997) and other painful arthritic/arthralgic conditions (Berkeley 1997;Buckwalter and Lappin 2000; Fillingim and Maixner 1995; Gabriel 2001; Unruh 1996). A role for sex hormones in TMD pain is supported by reports of higher prevalence among women of reproductive age that diminishes after menopause and recurs after estrogen replacement therapy (LeResche et al. 1997). Also, pain due to TMJ disk displacement (Suenaga et al. 2001) and muscle pressure pain thresholds vary systematically over the menstrual cycle (Isselee et al. 2002). Numerous animal studies have demonstrated that pain-like behavior evoked by cutaneous (Martinez-Gomez et al. 1994), deep, or visceral tissue stimulation (Giamberardino et al. 1997; Kayser et al. 1996; Ness et al. 2001) varies over the estrous cycle. However, few experimental studies have examined the effect of reproductive status on nociceptive processing of input from articular tissues.
Current understanding of the central neural basis for articular pain derives mainly from studies using the knee and ankle joints as models (see Kidd et al. 1996; Schaible and Grubb 1993), while fewer studies have used models specifically related to the TMJ region (see Sessle et al. 1993). The TMJ capsule, masticatory muscles, and overlying facial skin are supplied by branches of the trigeminal, upper cervical, and vagus nerves (Casatti et al. 1999; Denny-Brown and Yanagisawa 1973; Kido et al. 1993;Klineberg 1971; Uddman et al. 1998;Widenfalk and Wiberg 1990). Converging lines of evidence indicate that the subnucleus caudalis/upper cervical cord junction region (Vc/C2 region) plays a significant role in nociceptive processing related to the TMJ. Afferent nerves that supply the TMJ region terminate in common laminae at the Vc/C2 junction region (Jacquin et al. 1983; McNeill et al. 1991; Pfaller and Arvidsson 1988). Selective lesion of the Vc/C2 junction region, but not of more rostral areas of the trigeminal sensory complex, blocks the increase in masseter muscle activity caused by inflammation of the TMJ region (Hu et al. 1997). c-fos immunocytochemical studies reveal a high density of Fos-positive neurons produced at the Vc/C2 junction after acute injury to the TMJ region (Bereiter 2001; Bereiter and Bereiter 2000; Hathaway et al. 1995) and proestrous females produce greater numbers of Fos-positive neurons than diestrous females or males (Bereiter 2001). Recently, we reported that TMJ units in superficial laminae at the Vc/C2 junction region of male rats were excited by bradykinin (BK) injection into the joint space in a dose-dependent manner (Takeshita et al. 2001). BK is a potent algogenic agent whose levels are elevated after acute or chronic inflammation (Hargreaves et al. 1988) as well as within the TMJ after acute injury (Swift et al. 1998). The present study used localized injections of BK in the TMJ region of cycling female rats to test the hypothesis that the encoding properties of TMJ units in superficial laminae at the Vc/C2 junction region vary over the estrous cycle.
The protocols were approved by the Institutional Animal Care and Use Committee of Rhode Island Hospital and conformed to the established guidelines set by The National Institutes of Health guide for the care and use of laboratory animals (PHS Law 99-158, revised 2002). Beginning at 7–10 days prior to the experiment, the stage of the estrous cycle was determined by daily microscopic examination of vaginal smears taken by gentle lavage. Females were selected in proestrous or early diestrous stages of the estrous cycle when sex steroid levels are expected to be at the highest and lowest, respectively (Smith et al. 1975). The stage of estrous cycle was determined by monitoring vaginal smear cytology daily for two consecutive cycles and defined as proestrus or early diestrus (metestrus) only if normal cycling was confirmed and on the day of the experiment >70% of total cells were the expected cell type (see Adler 1981).
Female rats (Sprague-Dawley, Harlan, 250–300 g) were anesthetized initially with pentobarbital sodium (60 mg/kg, ip). After tracheostomy, animals were respired artificially with oxygen-enriched room air and catheters were placed in the right femoral artery (blood pressure monitor) and right jugular vein (anesthesia and drug infusions). Anesthesia was maintained by a continuous infusion of methohexital sodium (18–25 mg kg−1h−1) and switched to a mixture of methohexital and a paralytic agent, gallamine triethiodide (15–20 mg kg−1 h−1) after completion of all surgical procedures, just prior to the recording session. Adequate depth of anesthesia was confirmed by absence of corneal and hindlimb withdrawal reflexes prior to gallamine, fully constricted pupils, and constant arterial blood pressure and heart rate throughout the experiment. The animal was placed in a stereotaxic frame and the C1 and a portion of the C2 vertebrae were removed to expose the upper cervical dorsal horn. The brain stem surface was bathed in warm mineral oil. A portion of parietal bone was removed from the right skull, contralateral to the recording electrode, for placement of a dual stimulating electrode array in the sensory thalamus. The left temporalis muscle was reflected partially to expose the external pterygoid muscle and connective tissue overlying the dorsal surface of the posterior mandibular condyle. Expiratory end-tidal CO2 (3.5–4.5%) and mean arterial pressure (MAP, 100–130 mmHg) were monitored throughout the experiment. Body temperature was maintained at 38°C with a heating blanket and thermal probe.
The caudal portion of trigeminal subnucleus caudalis (Vc) and the upper cervical (C1–C2) spinal cord, 4 to 7 mm caudal to the obex, was explored for TMJ-responsive units using the entrance of the C2rootlet as a landmark. A tangential approach (approximately 43° off vertical, 60° off midline) was used to record single units extracellularly with tungsten microelectrodes (9 MΩ, Frederick Haer, Bowdoinham, ME) within 500 μm after surface penetration. Unit activity was amplified, discriminated (model DIS-1, BAK Electronics, Mount Airy, MD), and then stored and analyzed off-line on an Macintosh (Apple G3) computer using a DAQ interface board and LabVIEW software (National Instruments), as described previously (Hirata et al. 1999). The window discriminator allowed single units to be isolated on the basis of amplitude and time. Spike amplitude and shape were monitored throughout each experiment and displayed on a digital oscilloscope and stored on tape (model CDAT4, Cygnus Technology, Delaware Water Gap, PA) for subsequent reconfirmation during off-line data analyses. To confirm that only a single spike contributed to the overall spike count during each stimulus trial, a discriminated action potential waveform was displayed continuously on a digital oscilloscope and compared against the original shape and amplitude for that spike at the outset of the recording period.
The search for single units began by gentle mechanical palpation of the skin over the posterior aspect of the TMJ with a cotton-tipped applicator and then with a small metal burnisher (site D in Fig.1 A). However, the critical test for inclusion of a unit for further analysis was a vigorous response to direct probing of the surgically exposed dorsal aspect of the posterior condyle and surrounding muscles with a blunt metal probe (site C in Fig. 1). TMJ units that received convergent input from facial skin were classified further as wide dynamic range (WDR) or nociceptive specific (NS). WDR units were excited by brush (camel hair) or indentation of the skin surface with low-threshold von Frey filaments (1.2 g force) and showed a greater response to press or pinch (see example in Fig. 1 B). An arterial clip (approximately 20 mm2) was used for “press” stimuli, while “pinch” stimuli were delivered by a shorter, more stiff arterial clip (approximately 15 mm2). Brush, press, and pinch stimuli were applied for 10 s. When applied to the investigator's skin, the intensity of the press stimulus was near threshold for pain sensation, while the pinch stimulus was clearly above threshold. NS units were activated by press or pinch of the skin but not by brushing. TMJ units activated by deep indentation of the TMJ region alone with no apparent cutaneous receptive field (RF) were classified as deep-only cells. However, no systematic attempt was made to stimulate deep tissues distant from the TMJ region or within the oral cavity. No TMJ units were found that also responded only to brushing the skin [i.e., low threshold mechanoreceptive (LTM) units]. Electrical stimuli (0.1–5 mA, 0.2–2.0 ms, square-wave pulses) were applied to the TMJ region by a dorsal approach to estimate the type of fiber input. A bipolar electrode (Frederick Haer; 2 mm separation) was inserted approximately 2 mm deep into the tissue surrounding the posterior aspect of the mandibular condyle after retracting the temporalis muscle.
A single TMJ-responsive unit was recorded from each animal preparation. After confirming the response to condyle stimulation, the entire surface on the face and neck was explored for possible cutaneous inputs. Intraoral regions were not tested systematically and jaw movement was not used since this manipulation caused movement artifacts in pilot experiments. After cell classification and mapping of the cutaneous high-threshold RF for area measurements, a guide cannula (26 gauge) was inserted into the TMJ region (approximately 3 mm deep) by a dorsal approach directed at the posterior edge of the mandibular condyle near the zygomatic arch. Test solutions were delivered from a Hamilton syringe attached by polyethylene tubing to an inner cannula (33 gauge) that protruded approximately 0.5 mm from the end of the guide cannula. Test solutions consisted of phosphate-buffered saline (PBS), glutamate (GLU), BK, and mustard oil (MO). Each solution was injected from a separate inner cannula-tubing assembly to prevent drug mixing. Injections were delivered slowly over 30 s (total volume = 20 μl) with an inter-injection interval of 30 min to minimize tachyphylaxis, especially for BK. Typically, four to eight injections were made while recording from each unit. In pilot studies (n = 3) injection of PBS followed by five repeated injections of 1 μM BK (20 μl each) into the TMJ region at 30-min intervals evoked reproducible responses from single TMJ units and did not cause sensitization. The protocol for chemical stimulation of the TMJ region was as follows: injection of PBS (pH 7.4) followed by three doses of BK (0.1, 1, and 10 μM, pH 7.3, acetate salt, Sigma) and thenl-glutamate (200 mM, pH 7.1, monosodium salt, Sigma) using a cumulative dose design. MO (allyl isothiocyanate, 20% solution) was injected into the TMJ region at the end of the experiment after all other chemical agents were tested (n = 32). In several cases (n = 5), the selective B2antagonist, HOE 140, 10 μg in 20 μl of PBS (Sigma), was co-injected with BK into the TMJ region in proestrous rats.
An array of two concentric bipolar stimulating electrodes (SNE-100, Rhodes Medical Instruments, Tujunga, CA), separated rostrocaudally by 2 mm, was directed at the contralateral ventral posteromedial nucleus (VPM), and posterior nuclear group (PO) of the thalamus. The approximate stereotaxic coordinates in millimeters for VPM were AP = –2.5 from bregma, ML 2.5, DV 4–9 from the cortical surface; for PO, AP = –4.5, ML 2.5, DV 4–9. Antidromic activation was defined by a constant latency (<0.1 ms jitter), high-frequency following (0.1-ms pulse, 200–300 Hz, 20-ms train duration), and collision with an orthodromically driven spike within a critical time interval. The maximum allowable current used to define an antidromic response was 500 μA. Brain loci for antidromic responses were marked electrolytically (30 μA, 30 s).
Neural activity and arterial pressure were acquired and displayed by LabVIEW as peristimulus time histograms (PSTHs) of spikes per 1-s bins, exported to a spreadsheet, and analyzed off-line. Spontaneous activity (spikes/s) was calculated by averaging the spike counts during a 1-min epoch immediately preceding each stimulus. Evoked responses were assessed statistically by first calculating the response magnitude (R mag).R mag was determined by subtracting the mean plus two times the standard deviation (SD) of the background activity from the total spike count for each bin. The totalR mag for a given stimulus was defined as the cumulative sum of spikes for those contiguous bins in which the spike count minus the background gave a positive value. The totalR mag can be viewed as equivalent to the “area under the curve” for each stimulus (compare, for example, Fig. 4, A and B). Accordingly, response duration was defined as the time interval after stimulus onset until three consecutive bins with a positive spike count occurred above background (initial latency) and until the value of three consecutive bins no longer exceeded the mean + 2SD above background activity. This approach was described previously for caudal Vc units responsive to corneal stimuli (Hirata et al. 1999). A unit that failed to show three consecutive bins with a positiveR mag value within 100 s after the onset of the stimulus was considered unresponsive to that condition. Units were classified as BK-responsive (BK+units) if the total R mag exceeded that after injection of PBS by >50%, independent of BK dose. The threshold concentration of BK was defined as the lowest concentration that produced a total R mag that exceeded that to PBS by >50%. Responses to mechanical and chemical stimuli were assessed statistically by analysis of variance (ANOVA) corrected for repeated measures (Winer 1971) and individual comparisons were made by Newman-Keuls after ANOVA. Chi-square analysis determined if different classes of neurons (e.g., WDR, NS, deep only) or frequency of responsiveness to different test chemicals were represented equally between proestrous and diestrous females. Cutaneous high-threshold RF areas were determined using a small serrated forceps (approximately 3 mm2) and mapped onto a series of standardized rat face outlines (see examples in Figs. 4, 5, and 6). No attempt was made to selectively map the low-threshold RF area for WDR units. The RF areas were digitized and quantified by a planimetric method using National Institutes of Health Image software (v. 1.61). Cutaneous RF areas were not mapped blindly; however, in several cases the RF area was confirmed by a second investigator without knowledge of the stage of cycle and found not to differ significantly. The frequency distributions of RF areas for proestrous and diestrous units were compared using the K-S two-sample test, while the mean RF area for NS and WDR units were compared by ANOVA.
At the end of the experiment Sudan black dye (20 μl) was injected into TMJ region through the guide cannula to verify placement in the joint space. The animal was given a bolus dose of methohexital sodium (60 mg/kg, iv) and perfused through the heart with 10% formalin containing potassium ferrocyanide to identify antidromic stimulation sites by the Prussian blue reaction. Recovered sites were drawn on a standard series of brain outlines adapted from the atlas ofPaxinos and Watson (1997). The recording site was marked electrolytically (<5 μA, 20 s, see Fig. 7A inTakeshita et al. 2001).
A single neuron from each of 46 proestrous and 33 diestrous female rats was recorded from the superficial laminae at the Vc/C2 junction region as defined by a vigorous response to probing the dorsal aspect of mandibular condyle (Fig. 1). All TMJ units from proestrous (4.5 ± 0.5 spikes/s) and diestrous (4.6 ± 0.6 spikes/s) rats were spontaneously active and most displayed an irregular burst-like pattern of activity. Spontaneous activity rates for different classes of units (e.g., WDR, NS, deep only) were similar. Since mechanical stimuli were used to classify TMJ units, it was necessary to determine if stage of the estrous cycle affected mechanical sensitivity. As seen Fig. 1 C, TMJ units classified as WDR-like from proestrous and diestrous rats encoded mechanical stimuli with a progressive increase in firing rate with increasing stimulus intensity. Proestrous units displayed a greater numerical increase in firing than diestrous units; however, this difference was not significant. No TMJ units were inhibited by mechanical stimulation of deep tissues in TMJ region or the overlying facial skin. The average depth of recording for proestrous (192 ± 12 μm) and diestrous units (208 ± 30 μm) was similar and all sites were at or within 1.5 mm rostral to the entrance of the C2 rootlets (Fig.2). Nearly all TMJ units recorded from proestrous (42 of 46, 91.4%) and diestrous females (30 of 33, 90.8%) also received convergent nociceptive input from the overlying facial skin, while the remaining cells were classified as deep only (see Table1). Although a greater percentage of WDR units was sampled from proestrous than diestrous females, the relative percentage of WDR, NS, and deep cells was not significantly different (χ2 = 4.37, df = 2, P > 0.05). The depth reading from the microdrive indicated that, on average, WDR units were recorded from deeper locations in the dorsal horn than NS units in both animal groups (Fig. 2 B).
TMJ units from proestrous females had significantly larger cutaneous high-threshold RF areas than units from diestrous animals (Fig.3 A, χ2= 8.23, df = 2, P < 0.02). The average cutaneous RF area of proestrous units was 30.1% greater than diestrous units. This difference was due mainly to enlarged RF areas for WDR units, while the RF areas for NS units were similar for proestrous and diestrous females (Fig. 3 B). Typically, the convergent cutaneous RF extended rostral and ventral from the TMJ region to include the maxillary and mandibular divisions of the trigeminal nerve (see examples in Figs. 4 C and5 C). In all but two cases, at least a portion of the cutaneous RF was above the deep RF for the TMJ region. Although not tested extensively in all cases, no units were activated by mechanical stimulation of the contralateral TMJ region or facial skin.
A bipolar stimulating electrode was inserted into the TMJ region to estimate the type of fiber input that relayed information from the TMJ to the Vc/C2 region. In both proestrous and diestrous females input was mainly from small myelinated A-delta fibers since the conduction velocities averaged 3.9 ± 0.3 and 4.5 ± 0.5 m/s, respectively. Only 6 of 74 TMJ units (proestrus,n = 2; diestrus, n = 4) displayed long-latency activity (<2 m/s conduction velocity) consistent with C-fiber input, assuming a distance of 20 mm from the dorsal condyle surface to the Vc/C2 junction region.
Responses to PBS and BK
Injection of PBS alone activated 36 of 46 (78.3%) TMJ units in proestrous and 19 of 32 (59.4%) units in diestrous females (i.e.,R mag = >10 spikes per stimulus). Thirty-eight of 46 (82.6%) units from proestrous and 25 of 30 units (83.3%) from diestrous females were excited by BK injection into the TMJ region. BK+ units (i.e., totalR mag > 50% above that to PBS injection) were represented among all cell classes of proestrous (WDR,n = 19; NS, n = 16; deep only,n = 3) and diestrous females (WDR, n = 6; NS, n = 18; deep only, n = 1). The lowest concentration of BK (0.1 μM) tested was sufficient to excite 59.1% of BK+ units in proestrous females and 87.5% of BK+ units in diestrous females. Figures4 and 5 present examples of BK+ units from a proestrous and diestrous female, respectively. The raw spike counts are shown in Figs. 4 A and 5 A with confirmation of constant spike shape and amplitude for each unit. TheR mag values (spike count minus background activity) are displayed in Figs. 4 B and5 B. Co-injection of the selective B2receptor antagonist, HOE 140, greatly reduced BK-evoked activity supporting the notion of a receptor-mediated response (Fig.6 A). Each of five units tested from proestrous females displayed a decreased response of >95% after co-injection of BK plus HOE 140 compared with BK alone (not shown).
Figure 7 compares the average totalR mag for all BK+units (top, i.e., WDR, NS, and deep only cells), WDR units alone (middle), and NS units alone (bottom) from proestrous and diestrous females after injection of PBS and cumulative doses of BK. The total R mag for proestrous units increased to near-maximal values after only 0.1 μM BK and remained elevated after higher doses and was similar for WDR and NS units. The total R mag among all proestrous units was significantly greater that that of diestrous units at the lowest (0.1 μM) and highest doses of BK (10 μM). The largest overall difference in total R magbetween proestrous and diestrous units occurred among WDR cells (Fig.7, middle). At each cumulative dose of BK the totalR mag for WDR units from proestrous units was significantly greater than for diestrous units (P < 0.05). By contrast, the increase in totalR mag among NS units was not different for proestrous and diestrous units over the cumulative doses of BK (Fig. 7, bottom). Figure 8compares the response duration for all BK+ units (top, i.e., WDR, NS, and deep only cells), WDR units alone (middle), and NS units alone (bottom) from proestrous and diestrous females after injection of PBS and cumulative doses of BK. The response duration for all proestrous units increased progressively with increasing doses of BK, while the maximum increase in response duration for diestrous units was seen after 1 μM BK. Response duration evoked by the highest dose of BK (10 μM) was significantly longer for proestrous than diestrous units and this effect occurred among WDR and NS units.
Responses to GLU and MO after BK
Injection of GLU (200 mM) into the TMJ region, after cumulative doses of BK, excited many proestrous (40.9%) and diestrous units (64.7%). Examples of responses to GLU after BK for proestrous and diestrous units are shown in Figs. 4 and 5. To determine the relative responses of TMJ units to BK, GLU, and MO, the average totalR mag caused by injection of each agent was compared among BK+ units (Fig.9). Within each animal group the peak total R mag to BK, independent of BK dose, was not different from for GLU, while the totalR mag after MO was significantly (P < 0.01) greater than after BK or GLU. Between group comparisons (i.e., proestrus versus diestrus) of the totalR mag after BK was greater for proestrous than diestrous units, as expected; however, the totalR mag to GLU was not different. Similarly, although GLU evoked a significant increase in response duration (P < 0.01) for proestrous (51.1 ± 6.3 s) and diestrous units (38.6 ± 6.2 s), this group difference was not significant. The selective small fiber excitant, MO, was injected into the TMJ region after all other agents had been tested. A majority of TMJ units from proestrous (93.5%) and diestrous females (77.3% of units tested) responded vigorously to MO as shown by the example in Fig. 6 A. This example also suggested that MO-evoked TMJ unit activity was not dependent on the bradykinin B2 receptor since responses occurred within 60 min after administration of HOE 140. The totalR mag to MO among proestrous and diestrous units was not different among TMJ units classified as BK+. Correlation analyses revealed that the totalR mag to BK was not correlated with the response to GLU or MO (Spearman rank-order analysis,r s < 0.3) for either proestrous or diestrous units (data not shown).
Several TMJ units from each animal group were tested for antidromic activation from the contralateral sensory thalamus. Five of 20 (25%) units from proestrous females and 3 of 7 (43%) from diestrous females were activated antidromically from the posterior thalamus. Successful activation of units occurred only from the caudal-most electrode of the array, centered in the PO, and none from the more rostral electrode that passed through the ventro-posteromedial nucleus were consistent. Effective antidromic sites in PO were similar to those that activated cornea-responsive units recorded from caudal Vc (see example in Fig. 7 of Hirata et al. 2000). The average latency for activation from the PO was similar for proestrous (5.9 ± 0.3 ms) and diestrous (7.3 ± 1.0 ms) units. The average current for activation was 178 ± 64 and 250 ± 95 μA for proestrous and diestrous units, respectively.
Three main findings supported the hypothesis that the response properties of TMJ units in superficial laminae at the Vc/C2 junction region vary over the estrous cycle. First, injection of low doses of BK (0.1 μM) into the TMJ region evoked greater increases in totalR mag (i.e., total spikes per stimulus above background) and response duration of proestrous than diestrous units. Second, injection of high doses of BK (10 μM) caused further increases in total R mag and response duration above that seen after lower doses in proestrous units, while diestrous units displayed reduced responses consistent with desensitization. Third, convergent cutaneous high-threshold RF areas for proestrous units were larger than diestrous units, due mainly to an increase in RF area for WDR-like units. These results indicated that TMJ unit responses to chemical and mechanical stimulus modalities, especially among WDR-like units, were enhanced during proestrus, a stage of the estrous cycle when circulating levels of sex steroids normally are elevated.
These results extend those of a previous study in which the RF properties of TMJ units in superficial laminae at the Vc/C2 junction region of male rats were determined (Takeshita et al. 2001). TMJ units from proestrous and diestrous females shared many properties with those of males; however, there were differences. More than 90% of TMJ units from proestrous or diestrous females received convergent nociceptive (WDR or NS-like) input from facial skin, while only 74% of units from males were classified as WDR or NS (χ2 = 14.8, df = 4, P < 0.01). This suggested that the degree of spatial convergence onto second-order TMJ units was greater for females than males and did not depend on stage of the estrous cycle. Post-hoc analyses of cutaneous RF areas for WDR units from males (1.4 ± 0.18 cm2, n = 19,P < 0.01) were smaller than those from proestrous but not diestrous units. The percentages of TMJ units responsive to BK, GLU, or MO were similar in females and males. However, post-hoc analysis (ANOVA) of the BK dose-effect relationship indicated greater total R mag values (P< 0.01) for proestrous units after injection of low doses (0.1 μM) of BK (425 ± 82 spikes per stimulus) than males (262 ± 53) or diestrous females (174 ± 33). Similarly, response duration to 0.1 μM BK was longer (P < 0.01) for proestrous units (42 ± 4 s) than male (26 ± 3 s) or diestrous units (28 ± 5 s). Since most TMJ units responded to the lowest dose of BK tested, it was not possible to determine if stage of the estrous cycle significantly altered response thresholds to algesic chemicals injected into the TMJ region.
Although various etiologies likely underlie the onset of TMD pain, elevation of pro-inflammatory factors within the joint space may be a common contributing factor (Denucci et al. 1996;Milam and Schmitz 1995; Stohler 1997). Synovial fluid collected from clinical cases and animal preparations after TMJ injury revealed increased levels of BK, serotonin, substance P, and cytokines (Kopp 1998; Kubota et al. 1998; Swift et al. 1998). Sex steroids exert significant influence on inflammation and immune function (seeMcKay and Cidlowski 1999). Correspondingly, TMD patients display altered pressure pain thresholds for masticatory muscles (Isselee et al. 2002) and pain severity in relation to positional changes in TMJ disk attachment, an index of TMJ inflammation that varies over the menstrual cycle (Suenaga et al. 2001). These results support the use of an inflammatory model for TMJ injury and the assessment of reproductive status on processing nociceptive signals from deep craniofacial tissues.
Sex steroids, principally estrogen, have been implicated as causative factors underlying the influence of reproductive status on nociceptive processing and could act at peripheral or central loci. Evidence to support peripheral estrogen-dependent actions include the following: enlarged tactile RF areas of single trigeminal ganglion neurons during estrogen-evoked behavioral estrus (Bereiter and Barker 1980); expression of estrogen receptor-alpha in dorsal root ganglion (Sohrabji et al. 1994) and trigeminal ganglion neurons (this laboratory, unpublished observations); and increased glutamate-evoked TMJ afferent activity in female compared with male rats (Cairns et al. 2001). Also, a high percentage of TMJ primary afferents from male rats were excited by BK; however, female subjects were not included in that study (Takeuchi et al. 2001). Although estrogen levels have been linked to peripheral inflammation (see Bjorling and Wang 2001), it is not clear if differences in neural activity between naı̈ve and inflamed animals share mechanisms with those seen between proestrous and diestrous females. For example, ovariectomy normalized the sex differences in BK-evoked plasma extravasation in the knee joint and estrogen replacement returned these values to those of normal female rats (Green et al. 1999). Administration of exogenous estrogen for 2 weeks increased the expression of the BK B2-receptor in other tissues (Madeddu et al. 1997). By contrast, neither RF area nor thresholds for primary afferent neurons supplying the knee joint were altered significantly by prior inflammation (Dowd et al. 1998;Messlinger et al. 1994). Similarly, RF areas of primary afferent neurons that innervate the skin were not different between naı̈ve and inflamed rats, while RF areas of dorsal horn neurons were enlarged after inflammation (Hylden et al. 1989). Estrogen receptor expression in the TMJ disc was similar in symptomatic and asymptomatic TMD female and male patients (Abubaker et al. 1993). Together these results argue against a peripheral mechanism to explain the differences in response properties of TMJ units at the Vc/C2 junction in proestrous and diestrous females.
The signs and symptoms of TMD pain are most consistent with a central neural dysfunction since peripheral pathology of the joint region often is not apparent. Evidence to support central neural mechanisms in TMD pain include the following: spread of pain sensation from the TMJ to adjacent regions of the face (Turp et al. 1998); autonomic and endocrine disturbances (Jones et al. 1997;Korszun et al. 2002); lower thresholds and tolerance for thermal and ischemic pain stimuli applied to other body regions (Maixner et al. 1995); enhanced temporal summation for thermal pain (Maixner et al. 1998); and reduced ability to recruit endogenous pain controls (Bragdon et al. 2002). Also, previous recording studies noted an apparent increase in convergent input onto single brain stem neurons in cats (Rose 1975) and rats (Bradshaw and Berkley 2000) during periods of high circulating estrogen. Several lines of evidence suggest that the Vc/C2 junction region, the initial site of integration for nociceptive signals from the TMJ region, plays a critical role in TMD pain and may be a target for sex hormone modulation. The Vc/C2 junction receives convergent input from all sensory nerves that supply the TMJ region, i.e., trigeminal (Jacquin et al. 1983), upper cervical rootlets (Pfaller and Arvidsson 1988), and vagus (McNeill et al. 1991). A high density of Fos-positive neurons are produced at the superficial laminae at the Vc/C2 junction after acute injury to the TMJ region, proportional to stimulus intensity, and in greater numbers among proestrous than diestrous females or males (Bereiter 2001). The superficial laminae near the Vc/C2 junction region express a high density of estrogen receptor-positive neurons (Amandusson et al. 1996; this laboratory, unpublished observations). The present results further suggested that factors related to stage of the estrous cycle preferentially targeted WDR-like TMJ units. Cutaneous high-threshold RF areas for WDR units were enlarged in proestrous compared with diestrous units, while RF areas for NS units were similar. Iwata et al. (1999) also reported an increase in high-threshold RF areas of WDR neurons in medullary dorsal horn 24 h after inflammation of the TMJ region compared with naı̈ve animals. However, in that study most WDR units were recorded from deeper laminae and only male rats were included. The difference in BK-evoked total R mag for proestrous and diestrous units was greatest for WDR units. Also, response duration of WDR units after high doses of BK (10 μM) was greater for proestrous than diestrous units, while among proestrous units alone response duration for WDR units was greater than for NS units (see Fig. 8). Although these results suggested a prominent role for WDR-like units, NS units also likely contributed to the processing of chemical inflammatory input from the TMJ region since NS units from both female groups encoded BK concentration. Furthermore, in acute trigeminal recording studies, inflammation of the tooth pulp cavity (Chiang et al. 1998) or deep masseter muscle (Hu et al. 1992) increased the RF area of NS as well as WDR units in subnucleus caudalis of male rats 1–2 h after injury. It cannot be excluded that factors related to the estrous cycle led to apparent phenotypic changes and a shift in the percentage of TMJ units classified as WDR and NS based on cutaneous RF properties. Such changes would not be detected by the “between group” design of the present study.
The mechanisms by which sex steroids modify neural systems involve long-term genomic as well as rapid nongenomic actions (McEwen and Alves 1999) and have been well studied in the hippocampus. Long-term effects of estrogen include alteration in neural structure as well as function. Estrogen administration for 2 days increased dendritic spine density and synaptic connectivity in the hippocampal CA1 region similar to the changes observed over the estrous cycle (Woolley and McEwen 1992; Yankova et al. 2001). Estrogen may act directly throughN-methyl-d-aspartate (NMDA)-dependent neurotransmission (Pozzo-Miller et al. 1999) or indirectly through brain-derived neurotrophic factor and reduced GABAergic neurotransmission (Murphy et al. 1998) to increase spine density, and presumably, to enhance the excitability of hippocampal neurons. Rapid effects of estrogen on hippocampal neurons also have been reported and 10-min exposure to estrogen increased NMDA-dependent LTP in tissue slice preparations (Foy et al. 1999). Inflammation-induced neuroplasticity among trigeminal caudalis neurons has been shown in electrophysiological (Chiang et al. 1998) and c-fos studies (Bereiter and Bereiter 2000) to involve NMDA-dependent mechanisms. To determine if acute estrogen administration was sufficient to alter the response properties of TMJ units, in preliminary studies we monitored BK-evoked activity after injection of estradiol benzoate (20 μg, sc) in ovariectomized rats and noted no significant changes in either R mag or response duration over 2 h (unpublished observations). More than 2 days of estrogen administration were required to increase the cutaneous low-threshold RF area of single trigeminal ganglion neurons (Bereiter and Barker 1980). It is noteworthy that studies designed to examine the influence of estrogen replacement therapy on pain-related behavior or inflammation (Dina et al. 2001; Green et al. 1999; Khasar et al. 2001) have reported significant effects only after several days of treatment. Collectively, these findings are most consistent with the view that sex steroids modulate pain-related neural circuits to affect behavior by longer term genomic rather than rapid nongenomic mechanisms.
Although existing evidence suggests that the Vc/C2 junction is a critical relay for nociceptive input from the TMJ region, it cannot be excluded that other regions of the trigeminal brain stem complex also play a role in processing deep craniofacial pain. Fos-positive neurons are produced bilaterally at the ventrolateral pole of the subnucleus interpolaris/caudalis (Vi/Vc) transition region after acute TMJ injury (Bereiter and Bereiter 2000; Hathaway et al. 1995). However, the number of cells produced at the Vi/Vc transition is not proportional to TMJ stimulus intensity and is similar in females and males (Bereiter 2001). TMJ units also have been recorded in rostral portions of Vc, mainly in deep laminae, from cats of either sex and were excited by different classes of algesic chemicals including BK (Broton et al. 1988).Koijima (1990) used electrical and thermal stimuli and found that 80% of TMJ units throughout laminae I-V of Vc in female rats received convergent input from both the TMJ capsule and the masseter muscle; however, chemical stimuli were not tested and the stage of the estrous cycle was not reported.
These results suggested that factors related to the stage of the estrous cycle, rather than sex differences per se, contribute to changes in excitability of TMJ units in laminae I–II at the Vc/C2 junction region. Furthermore, these results indicated that neurons that integrate deep craniofacial input, located in the superficial laminae Vc/C2 junction region, demonstrate considerable capacity for neuroplastic change. Based on cutaneous RF properties and responsiveness to local injection of pro-inflammatory agents into the TMJ region, the present results were consistent with the hypothesis that central neural mechanisms contribute to enhanced activity by TMJ units during proestrous, the stage of the estrous cycle when estrogen levels are elevated.
The authors thank Dr. James W. Hu for helpful comments in preparing the manuscript and A. Benetti and D. Bereiter for excellent technical assistance.
This study was supported in part by Grant DE-12758 from National Institute of Dental and Craniofacial Research.
Address for reprint requests: D. A. Bereiter, Depts. Surgery and Neuroscience, Brown Medical School, Rhode Island Hospital, 222 Nursing Arts Bldg., Providence, RI 02903. (E-mail:).
- Copyright © 2003 The American Physiological Society