Estrogen Status and Trigeminal Ganglion Responses to Jaw Movement

Abstract

Chronic temporomandibular joint disorders (TMDs) present with pain in the temporomandibular joint (TMJ) and muscles of mastication. Risk factors for TMD include localized joint/muscle inflammation and estrogen status. This study determined whether mild tissue inflammation and estrogen status influenced the responses of trigeminal ganglion neurons to jaw palpation or jaw movement, 2 key diagnostic features of clinical TMD, in adult rats. Neuronal activity was recorded from male rats, ovariectomized (OvX) female rats, and OvX female rats injected with 17β-estradiol 24 h prior to testing (OvXE). Neurons were tested for responses to deep press over the TMJ region and jaw movement in 3 directions (open, protrusion, lateral) 10 d after intra-TMJ injection of a low dose of complete Freund’s adjuvant (CFA) or vehicle (sham). Deep press evoked similar responses in all treatment groups. The response magnitude to jaw opening and protrusion was significantly greater for neurons recorded from OvXE CFA-treated rats than from OvX CFA-treated or OvXE sham rats. The responses to lateral movement of the jaw were similar across all treatment groups. Most neurons (70% to 90%) displayed a static response pattern to jaw movement independent of direction. Estradiol treatment also increased the proportion of neurons that were excited by jaw movement in >1 direction as compared with untreated OvX females or males. These results suggest that mild localized inflammation in the TMJ region during periods of elevated estrogen were sufficient to increase the peripheral driving force for jaw movement–evoked hyperalgesia.

Introduction

Chronic temporomandibular joint disorders (TMDs) are nonprogressive conditions that present with fluctuating bouts of pain in the temporomandibular joint (TMJ) and muscles of mastication (Ohrbach and Dworkin 1998). A primary diagnosis of TMD is based largely on symptoms in the TMJ region (Ohrbach et al. 2011; Schiffman et al. 2014). Chronic pain is driven by a variable level of ongoing peripheral nerve activity and maintained by central neural mechanisms (Baron et al. 2013; Berta et al. 2017). Although patients with TMD often present with overlapping chronic pain conditions suggesting the involvement of central neural mechanisms (List and Jensen 2017), the contribution of peripheral nerve activity from the TMJ and jaw muscles to persistent TMJ-related hyperalgesia is not well defined.

The sensory innervation of the TMJ and masticatory muscles is supplied mainly by trigeminal ganglion (TG) neurons (Uddman et al. 1998). Preclinical studies have described the properties of TMJ and jaw muscle afferent neurons in naive animals (Klineberg et al. 1971; Kawamura and Abe 1974; Cairns et al. 2001a; Carins 2001b; Tsuboi et al. 2009), yet the response to jaw movement under persistent inflammatory conditions has received less attention. Since a limiting factor in translating preclinical results into clinical practice has been the lack of appropriate animal models (Yezierski and Hansson 2018), a primary aim of this study was to record the activity of TG neurons to jaw palpation and movement under conditions of minimal tissue inflammation. Low to moderate levels of synovial inflammation with minimal signs of tissue damage are common features of chronic TMD (Ernberg 2017). To address the issue of persistent inflammation of the TMJ region, a low dose (10 µg) of complete Freund’s adjuvant (CFA) was injected into the joint space; this dose has been reported to induce mild hypertrophy of the TMJ synovium but was sufficient to alter feeding behavior (Harper et al. 2001). To minimize tissue injury due to animal preparation, neural activity was recorded without surgical exposure of the TMJ region or the trigeminal brainstem by lowering the electrode vertically into the TG. The second aim of this study was to determine whether sex differences or estrogen status influenced the neural responses to jaw movement since considerable preclinical (Bereiter and Okamoto 2011) and clinical (Cooper and Kleinberg 2007; Isong et al. 2008) evidence has suggested that sex hormone status is a risk factor for TMJ hyperalgesia.

Materials and Methods

General Animal Preparation

Experiments were performed with 61 adult rats (200 to 300 g; Harlan Sprague-Dawley), which consisted of males, ovariectomized (OvX) females, and OvXE females that were given an injection of estradiol benzoate (E2, 30 µg/kg, subcutaneous, in sesame oil) 24 h before the experiment to simulate the preovulatory surge of estrogen in cycling females. A vaginal lavage sample confirmed the estrogen status of females on the day of the experiment (Naftolin et al. 1972). OvX females were used within 2 wk of gland removal. Animals received an injection of CFA (10 µg/10 µL) or saline under isoflurane (5%) into the left TMJ capsule and survived for 10 d prior to the recording session. This dose of CFA has been reported to cause minimal tissue damage to the TMJ synovium but did affect patterns of feeding behavior in adult rats (Harper et al. 2001). Rats received a single analgesic dose of carprofen (25 mg/kg, intraperitoneal) immediately after the intra-TMJ injection and displayed no overt signs of injury or behavior. Animals were assigned to different treatment groups in a random order. Animals were housed in pairs and given free access to food and water. Climate and lighting were controlled (25 ± 2 °C, 12:12-h light/dark cycle, light on at 7:00 am). The animal protocols were approved by the Institutional Animal Care and Use Committee of the University of Minnesota and according to guidelines set by the National Institutes of Health’s Guide for the Care and Use of Laboratory Animals (Public Health Service law 99-158, revised 2015). The author guidelines for ARRIVE 2.0 were followed for in vivo experiments.

Extracellular Single-Neuron Recording

Animals were anesthetized initially with urethane (1.2 g/kg, intraperitoneal) and after tracheotomy were maintained under isoflurane (1.5% to 2.0%) in oxygen-enriched air. A catheter was placed in the right femoral artery to monitor blood pressure. An adequate depth of anesthesia was confirmed by the absence of corneal and hind limb withdrawal reflexes. Expiratory end-tidal CO₂ (3.5% to 4.5%), arterial blood pressure (90 to 120 mm Hg), and body temperature (37 °C) were monitored continuously.

Animals were placed in a stereotaxic frame; the scalp was reflected; and a small hole was drilled on the left side of the skull to allow a microelectrode to be slowly lowered through the brain and into the TG (Fig. 1). A tungsten microelectrode (5 to 7 MΩ; FHC) was positioned under stereotaxic control (4 to 4.5 mm posterior to bregma, 3 to 4 mm lateral to the midline) and lowered until unit discharges were observed after brush of the posterior-lateral craniofacial skin. Next, to test for responses to mechanical input, a blunt probe (1-mm diameter) was applied to the muscle dorsal to the zygomatic arch overlying the lateral surface of the condyle. TMJ-responsive units were found 8 to 9 mm below the cortical surface, and none were excited by pinch of the skin overlying the TMJ. Jaw movement was made by rapidly lowering the mandible (~1 s) until resistance was met (average measured distance, 5 to 8 mm), held for 9 s, and then allowed to close passively. Similarly, the mandible was protruded until resistance was met (3 to 4 mm) and held for 10 s. Finally, the mandible was pulled to the left until resistance was met (4 to 5 mm), held for 10 s, and then pulled to the right for 10 s (Fig. 1C). At least 3 min elapsed between each jaw movement. Up to 4 neurons were recorded in each preparation (n = 129 neurons). At the end of the jaw movement session, many neurons (117 of 129) were tested for chemical sensitivity by intra-TMJ injection of ATP (1 mM, 20 µL).

Figure 1.

Recording sites in the trigeminal ganglion (TG) and setup for jaw movement. (A) Example of a recording site (*) in the TG at the junction of the ophthalmic/maxillary (V1/V2) and mandibular (V3) branches of the trigeminal nerve. (B) Summary of recovered recording sites in the TG from all experiments. M, male; OvX, ovariectomized females; OvXE, ovariectomized females injected with E2 on the day prior to recording. Arrows indicate medial (M) and rostral (R) orientations. (C) Rat skull outline indicating the range of jaw movements for opening (5 to 8 mm) and protrusion (3 to 4 mm). Lateral movement was 3 to 5 mm to either side (not shown). M, mandible; Z, zygomatic arch.

Data Analysis

Neural activity was amplified, discriminated, digitized, stored, and analyzed offline via LabChart software (AD Instruments). Neural data were displayed as peristimulus time histograms of spikes per 1-s bins. Background activity (spikes/s) was calculated as the average spike count over 1 min immediately preceding each jaw movement stimulus. The evoked responses were assessed by calculating the average spike count during over the 10-s stimulus minus the background activity. A neuron was considered responsive if the discharge rate to a specific directional movement was ≥1.5 times the background activity. Power analysis based on previous TG recording studies indicated that a sample size of 8 was required for 0.80 power. Statistical tests for differences in firing rates across animal groups was performed as a 1-way analysis of variance, and individual comparisons were made by a Tukey’s test (Prism version 9; GraphPad). The distributions of interspike intervals were compared by the Kolmogorov-Smirnov 2-sample test across pairs of treatment groups. Tests for differences in the frequency of occurrence of neurons responsive to individual jaw movements (i.e., open, protrusion, lateral) across treatment groups were determined by chi-square analyses. Data were presented as mean ± SEM, and the significance level was set at P < 0.05.

Histology

At the end of each experiment, rats were injected with pentobarbital (100 mg/kg, intraperitoneal) and perfused through the heart with 10% formalin. The recording site for the last recorded neuron was marked by electrolytic lesions (20 µA, 20 s) and confirmed histologically. Separate groups of animals were prepared for immunohistochemical analyses of the synovium (3 to 6 rats per group). Cells counts within the synovium were made at 40× after hematoxylin and eosin staining without prior knowledge of treatment. Cells counts were compared by analysis of variance, and P < 0.05 was considered significant. A separate group of OvXE rats received an injection of Evans blue dye (100 µL, 1% in phosphate-buffered saline) into the left ventricle under pentobarbital (70 mg/kg), followed by 150 mL of phosphate-buffered saline to assess plasma extravasation in TMJ tissues at 10 d after CFA (n = 5) or saline (n = 3) injection into the ipsilateral joint. The TMJ and surrounding muscles were removed and placed in formamide (1 mL/g of tissue) for 24 h. Absorbance readings were taken at 600 nm with a spectrophotometer. Absorbance values were compared by t test, and P < 0.05 was considered significant.

Results

The cellular thickness and density in the outer layers of the synovium appeared similar for sham and CFA-treated rats (Fig. 2). Cell counts across the outer layers of the synovium were similar across all treatment groups (28.1 ± 2.53 cells/0.1 mm, n = 26, F_5,20 = 0.74, P > 0.1). Evans blue dye absorbance revealed a significant increase at 10 d post-CFA versus controls (0.11 ± 0.002 vs. 0.03 ± 0.001, respectively) indicating a persistent increase in plasma extravasation (t = −6.63, P < 0.001).

Figure 2.

Micrograph examples of the temporomandibular joint synovial lining in ovariectomized rats injected with E2: (A) sham and (B) 10 d after complete Freund’s adjuvant. Note that the cellular density and synovial thickness were not significantly affected by adjuvant treatment. Hematoxylin and eosin stain. Scale = 50 µm.

A total of 129 TG neurons, identified by a robust response to deep press over the TMJ region, were recorded at the junction of the maxillary and mandibular branches of the trigeminal nerve (Fig. 1A, B). The responses to deep press ranged from 5.2 ± 2.3 to 7.3 ± 4.0 spikes/s and were not different across treatment groups (F_5,123 = 1.02, P > 0.1). The ongoing discharge rate of neurons was low (0.16 ± 0.26 spikes/s, n = 129) and similar across treatment groups. Most neurons were excited by jaw movement in >1 direction (Fig. 3); yet, as seen in the Table, approximately 10% of neurons did not respond to movement in any direction. The frequency of occurrence of neurons excited by movement in multiple directions was significantly greater for the OvXE group (χ² = 13.37, P < 0.001), whereas neurons from males (χ² = 0.13, P > 0.1) and OvX females (χ² = 1.3, P > 0.1) were as likely to respond to movement in none or 1 direction as to multiple directions. Since >1 neuron was sampled in each preparation, we confirmed that a neuron recorded late in the recording session was not more likely to display multidirectional responses than a neuron recorded earlier (χ² < 1.0, P > 0.1). Most neurons were slowly adapting and displayed a continuous firing pattern to jaw movement over the 10-s sampling period, as shown by the example in Figure 3. A slowly adapting firing pattern was dominant for neurons excited by jaw opening (range, 69% to 78% of neurons), jaw protrusion (range, 91% to 94%), or lateral movement (range, 81% to 96%). The percentage of slowly versus rapidly adapting neurons was similar within each animal group (male, OvX, OvXE) and across different treatments (sham, CFA). However, comparison of the cumulative interspike interval distribution for OVX and OVXE CFA-treated rats revealed a shift to shorter intervals to jaw opening for OvXE rats (Kolmogorov-Smirnov test, F[x] = 0.26, P < 0.05).

Figure 3.

Example of a temporomandibular joint neuron excited by jaw movement in 3 directions recorded in an ovariectomized rat injected with E2 at 10 d after complete Freund’s adjuvant injection. (A) Note that activity is sustained throughout the stimulus period. Red scale bars: 10 s. (B) Confirmation of spike isolation by coincident spike shape of 7 superimposed traces. Dotted blue line indicates threshold for spike inclusion. This figure is available in color online.

Table.

Number of Trigeminal Ganglion Neurons Excited by Jaw Movement in Different Axial Directions Under Sham, Inflamed, and Stress Conditions.

	No. of Jaw Directions
Treatment	1	2	3	None	Total
Male
Sham	2	4	12	3	21
CFA	2	6	9	4	21
OvX
Sham	3	2	14	5	24
CFA	4	3	12	1	20
OvXE
Sham	1	7	12	0	20
CFA	0	6	16	1	23
Total	12	28	75	14	129

CFA, complete Freund’s adjuvant; OvX, ovariectomized; OvXE, ovariectomized with 17β-estradiol.

Jaw opening evoked a significant overall increase in firing rate (F_5,123 = 3.58, P < 0.005). As shown in Figure 4A, individual comparisons revealed that activity evoked by jaw opening was significantly greater for neurons of OvXE CFA-treated rats than for OvX CFA-treated rats (P < 0.01) and sham OvXE rats (P < 0.05). Jaw protrusion also evoked an overall increase in neural activity (F_5,123 = 3.96, P < 0.002) that was greater for OvXE CFA-treated rats than for OvX CFA-treated rats (P < 0.01) and for sham OvXE rats (P < 0.05; Fig. 4B), whereas activity evoked by lateral jaw movement was similar across all groups (F_5,123 = 1.14, P > 0.11; Fig. 4C). An intra-TMJ injection of ATP (1 mM, 20 µL) excited 12.8% of TG neurons (15 of 117 units tested), consistent with a recent report in mice that few TG neurons that supply the masseter muscle were excited by ATP (Lindquist et al. 2021).

Figure 4.

Summary of the temporomandibular joint neuronal responses to jaw movement: (A) jaw opening, (B) jaw protrusion, and (C) lateral jaw movement. Sample sizes: male sham, n = 21 neurons; male CFA, n = 21; OvX sham, n = 24; OvX CFA, n = 20; OvXE sham, n = 20; OvXE CFA, n = 23. *P < 0.05 and **P < 0.01 across treatments; mean ± SEM. CFA, complete Freund’s adjuvant (10 d postinjection); FR, Firing rate; OvX, ovariectomized; OvXE, ovariectomized with 17β-estradiol; S, sham.

Discussion

The main findings of this study indicated that moderate but persistent inflammation of the TMJ and high estrogen status combined to increase 1) the frequency of occurrence of TG neurons that encoded jaw movement in multiple directions and 2) the magnitude of the firing rate by neurons to jaw opening or protrusion. High estrogen status in female rats further altered the firing pattern of TG neurons during inflammation by shifting the distribution of interspike intervals to shorter intervals. The significance of these findings was that moderate inflammation of the TMJ region alone was not sufficient to enhance the sensitivity of TG neurons that encode jaw movement but rather became effective under conditions of increased estrogen levels.

Previous reports indicated that myogenic TMD cases can be classified on the basis of tender point scores and sensory testing, as those displaying evidence of peripheral sensitivity versus those more likely to have altered central pain mechanisms (Pfau et al. 2009; Bair et al. 2016). Microneurography in awake control subjects revealed a causal relationship between trigeminal nerve activity and masseter muscle activity (Turker et al. 2006). However, the effects of ongoing craniofacial inflammation on the properties of nerve fibers that supply the TMJ or jaw muscles have not reported in clinical studies. Earlier studies in sham animals demonstrated that many TG neurons that responded to jaw movement were slowly adapting and were activated by movements in >1 direction (Klineberg et al. 1971; Kawamura and Abe 1974; Tsuboi et al. 2009) in agreement with the present results. Other studies have noted that TG neurons were activated by glutamate injection into the TMJ capsule (Cairns et al. 2001a) or jaw muscle (Cairns et al. 2001b), and in both studies, responses were greater in sham female than male rats, although responses to jaw movement were not assessed. By contrast, no previous in vivo recording studies have determined if local inflammation altered the response properties of TG neurons to jaw movement. The present study was designed to quantify the responses of TG neurons in vivo to jaw movement and during minimal persistent TMJ inflammation in male rats and OvX and OvXE female rats. Previous in vitro studies indicated that acute inflammation increased the mechanical sensitivity of TMJ afferents (Takeuchi et al. 2004) and the excitability of isolated TG neurons identified by prior anatomic tracing methods (Flake et al. 2005), while long-term estrogen treatment and TMJ inflammation increased the excitability of TG neurons in a synergistic manner (Flake et al. 2005). Interestingly, a recent study stated that intra-muscular injections of glutamate and nerve growth factor evoked similar peak pain intensities in control male and female subjects; yet, by 10 d postinjection, females displayed significantly greater decreases in pressure pain thresholds and pain due to chewing when compared with males (Alhilou et al. 2021).

A key finding was that estrogen treatment increased TG responses to jaw opening and protrusion but not to lateral movement. This directional preference matched well the pattern of innervation of the rat TMJ in which nerve terminal densities were highest anteriorly and posteriorly and less in the medial and lateral planes (Kido et al. 1995; Shinoda et al. 2003). The present study did not address the site and/or mechanisms by which exposure to elevated estrogen for 24 h enhanced the activity of TMJ-responsive neurons. However, given the short exposure time, it is likely that nongenomic as well as genomic mechanisms contributed to the enhanced TG responses to jaw movement by estrogen. Estrogen receptors are expressed by TG neurons (Bereiter et al. 2005; Liverman et al. 2009), and acute exposure to E2 can rapidly enhance bradykinin signaling in TG neurons (Rowan et al. 2010). Acute E2 treatment increased trigeminal brainstem neuronal responses to intra-TMJ injections of ATP in an NMDA-dependent manner (N-methyl-D-aspartate; Tashiro et al. 2009), while long-term E2 treatment increased NMDA-evoked masseter muscle afferent nerve activity and increased the expression of the NMDA subunit NR2B in TG neurons that supply the masseter muscle (Dong et al. 2007). The NR2B receptor subtype has been linked to long-term potentiation in other brain regions in an estrogen-dependent manner (Smith et al. 2016). The functional significance of an increase in the number and response magnitude of TG neurons that encode jaw movement and their role in TMJ hyperalgesia are not well defined. However, previous studies revealed that relatively brief increases in peripheral nociceptor activity were sufficient to cause prolonged central sensitization (Hathway et al. 2009). Recently, we reported that local blockade of connexin 43, a key gap junction protein, in the TG prevented the enhancement of TMJ-evoked jaw muscle activity following TMJ inflammation (Ahmed et al. 2021), consistent with a role for peripheral nociceptor input in TMJ hyperalgesia. Thus, it is possible that even small changes in TG neural activity after inflammation may be sufficient to produce significant long-term changes in central brain pathways during periods of elevated estrogen.

The limitations of preclinical studies as translational models for human pain are well known (Yezierski and Hansson 2018). This may be particularly challenging for animal models of TMJ hyperalgesia in which species differences in the biomechanical properties of jaw movement may play a role. Although the rat TMJ shares several anatomic and histologic features with the human (Porto et al. 2010), the rodent jaw displays greater protrusive movements (Herring 2003). The present study used a low dose of CFA to avoid overt tissue damage, since clinical examinations of patients with TMD typically reveal no overt signs of tissue injury or inflammation; however, more invasive methods, such as synovial fluid sampling, often find evidence of ongoing inflammation suggesting that TMD may be defined as a low-grade inflammatory condition (Ernberg 2017). Thus, we injected a low dose of CFA, but even this low dose resulted in an increase in plasma extravasation in tissue samples from the TMJ region at 10 d postinjection. To limit the influence of surgical preparation on these results, we positioned the recording electrode in the TG with minimal surgery, restricted to the scalp, and placement of a tracheal tube and catheter in the femoral artery. The limitations of this study included both the method to apply jaw movement and the search strategy used to identify a responsive neuron. Jaw movements were made manually just to the point of resistance, and only neurons with a rigorous response to deep press to tissues overlying the TMJ were included for further analysis. Although jaw movement beyond the range of passive resistance may have increased the number and magnitude of responsive neurons, a goal of this project was to assess TG encoding properties without inducing TMJ tissue damage to better match the clinical signs of many patients with TMD. We cannot exclude that neurons not responsive to deep press of tissues overlying the TMJ may have been responsive to jaw movement. These limitations may have contributed to the variability of the responses seen across treatment groups, while the search strategy restricted to those neurons responsive to deep press may have underestimated the number of TG neurons responsive to jaw movement.

The heterogeneity of conditions grouped under the diagnostic heading of TMD has likely contributed to the variable effectiveness of specific treatments for jaw pain (List and Jensen 2017). Several studies have reported that treatments based on intramuscular or intra-articular injections of drugs with anti-inflammatory, analgesic, or structural effects were more effective than placebo (Machado et al. 2018; Liu et al. 2020); however, none have specifically targeted the TG, as has been suggested for spinal pain (Berta et al. 2017). The present results lend support to the notion that activation of TG neurons plays a role in driving TMJ hyperalgesia, which may be amplified by inflammation and enhanced by elevated estrogen conditions.

Author Contributions

X. Zhang, contributed to design and data acquisition, drafted and critically revised the manuscript; M. Rahman, contributed to design and data interpretation, drafted and critically revised the manuscript; D.A. Bereiter, contributed to conception, design, data analysis and interpretation, drafted and critically revised the manuscript. All authors gave final approval and agree to be accountable for all aspects of the work.