Physiological temperatures drive glutamate release onto trigeminal superficial dorsal horn neurons

Abstract

Trigeminal sensory afferent fibers terminating in nucleus caudalis (Vc) relay sensory information from craniofacial regions to the brain and are known to express transient receptor potential (TRP) ion channels. TRP channels are activated by H⁺, thermal, and chemical stimuli. The present study investigated the relationships among the spontaneous release of glutamate, temperature, and TRPV1 localization at synapses in the Vc. Spontaneous excitatory postsynaptic currents (sEPSCs) were recorded from Vc neurons (n = 151) in horizontal brain-stem slices obtained from Sprague-Dawley rats. Neurons had basal sEPSC rates that fell into two distinct frequency categories: High (≥10 Hz) or Low (<10 Hz) at 35°C. Of all recorded neurons, those with High basal release rates (67%) at near-physiological temperatures greatly reduced their sEPSC rate when cooled to 30°C without amplitude changes. Such responses persisted during blockade of action potentials indicating that the High rate of glutamate release arises from presynaptic thermal mechanisms. Neurons with Low basal frequencies (33%) showed minor thermal changes in sEPSC rate that were abolished after addition of TTX, suggesting these responses were indirect and required local circuits. Activation of TRPV1 with capsaicin (100 nM) increased miniature EPSC (mEPSC) frequency in 70% of neurons, but half of these neurons had Low basal mEPSC rates and no temperature sensitivity. Our evidence indicates that normal temperatures (35–37°C) drive spontaneous excitatory synaptic activity within superficial Vc by a mechanism independent of presynaptic action potentials. Thus thermally sensitive inputs on superficial Vc neurons may tonically activate these neurons without afferent stimulation.

the trigeminal nucleus caudalis (Vc) receives sensory information relayed predominantly by glutamate released from trigeminal primary afferent fibers from the head and neck, including highly specialized tissues like the cornea and dura (Gibbs et al. 2011; Hiura and Nakagawa 2012). On peripheral stimulation, action potentials invade Vc central terminals and trigger evoked glutamate release, but substantial release of vesicular glutamate also occurs spontaneously (Jennings et al. 2003; Travagli and Williams 1996). Recently, it has been suggested that evoked and spontaneous glutamate release may arise from different pools of vesicles that are distinctly regulated and have unique physiological importance (Kavalali et al. 2011).

Primary afferent neurons discharge little in the absence of peripheral stimulation or injury (Jankowski et al. 2012; Pitcher and Cervero 2010), and, therefore, their central terminals might be expected to have similarly low rates of neurotransmitter release in basal conditions. However, at central synapses measured in slices, neurotransmitter release occurs spontaneously, and such events are commonly viewed as arising from infrequent and stochastic release from the same pools of vesicles as action potential-evoked release (Ermolyuk et al. 2013; Katz 1971). In contrast to peripheral activation characteristics, the rates of spontaneous vesicle release are often much higher from primary afferent terminals (Grudt and Williams 1994; Shoudai et al. 2010; Uta et al. 2010). Indeed, primary afferents that contact second-order neurons in the solitary tract nucleus (NTS) often express the transient receptor potential (TRP) vanilloid type 1 channel (TRPV1) and have substantially higher spontaneous glutamate release rates than neurons with afferents lacking TRPV1 (Peters et al. 2010). The spontaneous excitatory postsynaptic current (sEPSC) rates in NTS strongly depend on the ambient temperature for the TRPV1-expressing afferents (Shoudai et al. 2010) and are independent of voltage-activated calcium channels yet strongly modulated by G protein-coupled receptors (GPCRs; Fawley et al. 2011). The high sensitivity near 37°C contrasts to the relatively weak thermal sensitivity at physiological temperatures of neural excitability and conduction more generally (Hardingham and Larkman 1998; Klyachko and Stevens 2006; Pyott and Rosenmund 2002).

Thermally responsive TRPs, including TRPV1, are present in trigeminal sensory afferent fibers and mark the superficial layers in Vc as primary nociceptor terminations (Cavanaugh et al. 2011a; Jennings et al. 2003). Noxiously hot (>45°C) peripheral temperatures evoke action potentials in trigeminal afferents (Cuellar et al. 2010) that can be conducted to superficial Vc neurons and are generally attributed to TRPV1 expression (Neubert et al. 2008). Here, we measured the thermal sensitivity of spontaneous glutamate release in Vc to test whether temperature sensitivity corresponded to TRPV1 expression centrally. In horizontal slices, we recorded sEPSCs in neurons in the superficial laminae of Vc. We found two distinct sets of neurons based on their basal sEPSC rate; high rates of spontaneous glutamate release at near-physiological temperatures (36°C) was greatly reduced with modest decreases in bath temperature (30–35°C) in one group of neurons. The remaining neurons exhibited low spontaneous glutamate release that was thermally insensitive, but TRPV1 expression was not exclusively related to either group. The findings demonstrate substantial heterogeneity in Vc spontaneous glutamatergic transmission and the existence of TRPV1-independent thermal coupling to excitatory transmitter release.

METHODS

Animals.

All animal procedures were approved by the Institutional Animal Care and Use Committee and conformed to the guidelines of the National Institutes of Health publication Guide for the Care and Use of Laboratory Animals. Hindbrains of male Sprague-Dawley rat pups (P9–28; n = 120; Charles River Laboratories) were prepared as described previously (Grudt and Williams 1994) under isoflurane anesthesia (5%, 2 l/min in air). Horizontal slices (200–230 μm) were cut with a sapphire knife (Delaware Diamond Knives, Wilmington, DE) mounted in a vibrating microtome (Leica VT1000 S; Leica Microsystems, Bannockburn, IL). Slices were submerged in a perfusion chamber and placed in artificial cerebrospinal fluid (aCSF) containing (in mM): 125 NaCl, 3 KCl, 1.2 KH₂PO₄, 1.2 MgSO₄, 25 NaHCO₃, 10 d-glucose, 2 CaCl₂, bubbled with 95% O₂-5% CO₂; pH 7.4; 293–300 mosM at 32°C for 45–60 min before recording. Bath temperature was controlled within 1°C using an inline heating system (TC2BIP with HPRE2 and TH-10Km bath probe; Cell MicroControls, Norfolk, VA) and was continually measured with a thermistor placed immediately downstream from the slice (Fawley et al. 2011). For light microscopy studies, slices were obtained as described above and then immediately fixed in 4% paraformaldehyde in 0.1 M phosphate buffer (PB) using a standard microwave for 5 s at 50% power. Sections were rinsed in 0.1 M PB and stored in 30% sucrose-30% ethylene glycol storage solution in 0.1 M PB at −20°C until immunocytochemistry was performed.

Voltage-clamp recordings.

Neurons selected for recording had their cell bodies located within the outer lamina (L_I/IIo) of the Vc [a translucent band ≤200 μm medial to spinal trigeminal tract (spV)]. Patch electrodes were pulled from borosilicate glass (outside diameter = 1.5 ± 0.05 mm; inside diameter = 1.0 ± 0.05 mm; Garner Glass), fire-polished, and had input resistances of 3.5–5.0 MΩ when filled with a low Cl⁻ (10 mM, E_Cl = −69 mV) intracellular solution consisting of (in mM): 6 NaCl, 4 NaOH, 130 K-gluconate, 11 EGTA, 2 CaCl₂, 2 MgCl₂, 10 HEPES, 2 Na₂ATP, and 0.2 Na₂GTP; pH 7.31–7.33; 286–294 mosM. Neurons were visualized using infrared differential interference contrast optics (Axioskop FS2; Carl Zeiss, Thornwood, NY) as previously described (Peters et al. 2010). Neurons were recorded using whole cell configuration and held in voltage-clamp at holding voltage (V_H) = −60 mV. Signals were sampled at 20 kHz and filtered at 10 kHz using pCLAMP software (version 9.2; Molecular Devices, Sunnyvale, CA) with a MultiClamp 700A, Digidata 1322A analog-to-digital converter (Axon Instruments, Union City, CA). Under our conditions, both glutamatergic and GABAergic synaptic currents were inward at a V_H of −60 mV. Liquid junction potentials were not corrected. The GABA_A receptor antagonist, gabazine (GBZ; 3 μM), was present in all experiments to isolate EPSCs. Miniature EPSCs (mEPSCs) were measured in TTX (1 μM) to prevent action potentials. Thermally evoked increases in EPSC frequency were evaluated using small step changes in bath temperature (30–36°C) by electronically controlling (Master-9; AMPI, Jerusalem, Israel) the inline heating system. Capsaicin (CAP; 100 nM) was added at the end of experiments in the presence of TTX to determine whether inputs expressed TRPV1 (TRPV1+).

Current-clamp recordings.

In some recordings, neurons were classified as described above based on sEPSCs before switching the recording configuration to current-clamp to assess action potential firing. Both excitatory postsynaptic potential and action potential frequency were recorded during temperature changes in the presence of GBZ to eliminate fast inhibitory transmission and to focus on excitatory mechanisms. Temperatures were controlled as in voltage-clamp recordings. Following temperature steps, the configuration was changed back to voltage-clamp, and TTX was added to confirm whether the thermal sensitivity was mediated directly or indirectly.

Drugs.

All drugs were added to the aCSF bath solution and perfused for at least 2 min (bath volume = 0.5–0.7 ml, flow = 1.9–2.1 ml/min). GBZ (SR-95531), TTX, 6,7-dinitroquinoxaline-2,3-dione disodium salt (DNQX), and CAP were purchased from Tocris Bioscience (Ellisville, MO).

Light microscopy.

Immunocytochemistry for light microscopic peroxidase detection of TRPV1 in the Vc was performed using a method modified from that previously described (Hegarty et al. 2007). Following incubation in 0.5% BSA in 0.1 M Tris solution (TS) for 30 min, 200-μm horizontal sections were incubated in the goat anti-TRPV1 primary antibody [1:500, cat. no. sc12498, Santa Cruz Biotechnology, Dallas, TX; 0.1% BSA, 0.25% Triton X-100 (Sigma, St. Louis, MO) in 0.1 M TS] for 2 nights at 4°C with continuous agitation. Bound TRPV1 primary antibody was visualized by incubating the Vc tissue sections in biotinylated horse anti-goat IgG secondary antibody (1:400, cat. no. BA-9502, Vector Laboratories, Burlingame, CA; 0.1% BSA in 0.1 M TS) for 30 min followed by 30-min incubation in avidin-biotin (VECTASTAIN Elite ABC Kit; Vector Laboratories) and then diaminobenzidine-hydrogen peroxide (DAB-H₂O₂) solution for 2 min. Tissue sections were mounted on slides, serially dehydrated in increasing ethanol concentrations and xylenes, and then coverslipped using DPX mounting medium (Sigma). Images from sections containing Vc and spV were collected using an Olympus BX51 microscope equipped with a DP71 camera (Olympus America, Center Valley, PA) and associated software.

Data analysis.

Data (frequency, amplitude, and bath temperature) were analyzed offline using Mini Analysis (Synaptosoft Systems) and Ophys (courtesy of Charles Jason Frazier, University of Florida) in OriginLab 8.6. Individual cells were evaluated for normal distribution using the appropriate parametric (Student's t-test) or nonparametric tests (Kolmogorov-Smirnov, K-S). Temperature sensitivity was determined by one-way ANOVA (post hoc, Student-Newman-Keuls), by two-way ANOVA (Bonferroni post hoc), and by calculating Arrhenius relationships with the thermal coefficient, Q₁₀ (Ni et al. 2006). Data were plotted and then fitted by linear regression, and the slope values (slope) and the correlation coefficient (r²) were calculated to quantify thermal sensitivity. Group data are presented as means ± SE and were statistically compared using either the Student's t-test (2 groups) or repeated-measures ANOVA (>2 groups). Statistical significance was set at P ≤ 0.05.

RESULTS

sEPSCs were evident in all Vc neurons (Fig. 1). At a bath temperature of 36°C, the frequencies of sEPSCs were quite high in most neurons (Fig. 1A), but cooling to 30°C markedly reduced this rate. Changes in bath temperature were closely tracked in the sEPSC rate in such thermally responsive neurons (e.g., increases in temperature evoked increases in sEPSC rate; Fig. 1C). In the remaining neurons, the basal sEPSC rates were substantially lower even at warm, physiological temperatures and poorly tracked changes in bath temperatures (Fig. 1, B and D). Thus the basal sEPSC rates identified two distinct groups of Vc neurons with inputs that had high and low rates of release (Fig. 2A): High (≥10 Hz) or Low (<10 Hz) neurons, respectively, when compared at 35°C. Note that High neurons responded to changes in temperature with substantial changes in sEPSC rate, whereas Low neurons showed little change in sEPSC rate. This separation by frequency extended to the full temperature-sEPSC rate relations for individual neurons such that the sEPSC rate of High neurons did not overlap with relations of individual Low neurons and vice versa (Fig. 2A); therefore, the standard recording temperature was set at 32°C. The non-N-methyl-d-aspartate (non-NMDA) receptor-selective antagonist DNQX (5–20 μM in GBZ; n = 4) blocked 99.9 ± 0.01% of sEPSCs in either type of neuron at all temperatures. Temperature changes did not alter the amplitudes of sEPSCs in the two groups of neurons (Fig. 2B), suggesting that temperature acts at a presynaptic site to promote selectively the frequency of glutamate release. The magnitude of the changes in sEPSC frequency also suggests that temperature is the predominant factor driving tonic glutamate release onto superficial Vc neurons at normal temperatures.

Fig. 1.

Physiological temperatures drive spontaneous glutamate release in trigeminal nucleus caudalis (Vc). Traces represent basal and temperature responses (30–36°C) in Vc neurons with High (≥10 Hz) spontaneous activity (A) and Low (<10 Hz) spontaneous activity (B). Individual diary plots of spontaneous excitatory postsynaptic current (sEPSC) frequency (black bars) and bath temperature (Temp; red line) over the 1st 15 min of recording are shown in High (C) and Low (D) for cells shown in A and B, respectively. sEPSC rate tracked with increases or decreases in temperature accordingly.

Fig. 2.

Temperature acts presynaptically to drive glutamate release. A: neurons segregated into 2 distinct groups: High (red squares, n = 7) with sEPSC rates ≥5 Hz or Low (blue squares, n = 5) with sEPSC rates <5 Hz at 32°C. Significant increases in sEPSC rate were observed at temperatures ≥34°C regardless of basal rate (P < 0.001, ANOVA). B: amplitudes were not different between High and Low neuron groups (P > 0.05, Student's t-test), and temperature did not significantly change event size (P > 0.05, ANOVA). Large symbols represent the mean value (±SE); individual cells are represented by colored lines.

To test the physiological relevance of these sEPSCs, we turned to current-clamp conditions to record action potentials. At 36°C, most High neurons discharged spontaneously, and lowering the temperature of the bath decreased and then eliminated the action potential frequency (Fig. 3A). As might be expected from the synaptic recordings, Low neurons displayed modest instantaneous action potential discharge rates that decreased to silence by lowering the temperature (Fig. 3B). This subset of cells (n = 21) was initially classified as High (12.9 ± 1.1 Hz, n = 11) or Low (2.1 ± 0.4 Hz, n = 10) under the voltage-clamp protocols. Under current-clamp, 67% of recorded superficial Vc neurons had temperature-sensitive action potential responses that resembled their sEPSC responses (i.e., n_High = 7, n_Low = 7). Action potentials in the remaining seven neurons were not temperature-sensitive (n_High = 4, n_Low = 3), evidence of the specific and restricted site of action of high thermal sensitivity. Note that the action potential response to current injection was not altered by temperature changes, a finding that is consistent with a presynaptic action and postsynaptic electrical excitability. These data suggest that thermally evoked, spontaneous glutamate release is sufficient to initiate postsynaptic action potentials, and, therefore, this information will be conducted within these circuits.

Fig. 3.

Temperature-evoked sEPSCs generate postsynaptic action potentials. Representative current-clamp recordings of superficial Vc neurons with sEPSC rates were classified as High, ≥5 Hz (n = 7; A), or Low, <5 Hz (n = 7; B) sEPSC rates at 32°C. As temperature increased from 30 to 36°C, the number of action potentials generated increased regardless of basal sEPSC rate; the magnitude of elicited action potential responses was dependent on sEPSC category as observed at 36°C (left).

Spontaneous release can either arise from mechanisms within presynaptic terminals or be generated by local circuit pathways (Kaeser and Regehr 2014). To test whether conducted presynaptic action potentials might contribute to temperature-induced sEPSC responses, we generated thermally responsive sEPSCs and then added TTX (1 μM), a voltage-gated Na⁺ channel blocker, to the aCSF and repeated the temperature challenges (Fig. 4A). On average, in High neurons (n = 13), the sEPSC rates without TTX were not different from the rates of mEPSCs measured in TTX in these same neurons (P > 0.05, Student's t-test; Fig. 4B). Note that thermally driven changes in mEPSC rate, like sEPSCs, closely tracked bath temperature and were readily reversible (Fig. 4A). These observations suggest that, in High neurons, thermal sensitivity is intrinsic to the glutamate terminals directly contacting the recorded neurons (Fig. 4B). In Low neurons (n = 5), TTX did not alter basal sEPSC rates (sEPSCs: 1.6 ± 0.8 Hz; mEPSCs: 2.2 ± 1.5 Hz; P = 0.91, Student's t-test).

Fig. 4.

Blockade of voltage-gated Na⁺ channels in High and Low neurons reveals selective drive of thermally evoked glutamate release. A: the histogram of a representative High neuron where sEPSC frequency tracks closely and is reversible with bath temperature in control solution (left; 0–5 min). Addition of TTX (1 μM) to block voltage-gated Na⁺ channels isolated activity-independent glutamate release and did not alter the rate of thermally evoked EPSCs (6–10 min). B: miniature EPSC (mEPSC) frequency from High neurons (n = 13) is not significantly different in the control solution (red squares) vs. TTX (red circles; P > 0.05, Student's t-test). Increases in EPSC rate were intact after exposure to TTX, suggesting High neurons were directly activated by temperature. C: a histogram of EPSC frequency and bath temperature for a Low neuron in control solution (sEPSCs, 0–5 min) and in TTX (mEPSCs, 6–10 min). D: for Low neurons, inclusion of TTX (blue circles) in the bath solution significantly attenuated thermally evoked responses but not basal release in control solution (blue squares), suggesting that Low neurons (n = 5) were insensitive to temperature changes from 30 to 36°C (**P ≤ 0.01, ANOVA). Data in B and D are shown as means (±SE).

In some Low neurons, the warmest temperatures increased the sEPSC rates significantly above sEPSC rates recorded at lower temperatures (Fig. 1B). Note, however, that the average absolute frequencies of sEPSCs were quite low, and any increases at warmer temperatures were small and often inconsistent (Figs. 1D, 2A, and 4C). Application of TTX reduced the sEPSC rates and eliminated temperature-dependent changes in Low neurons (Fig. 4D). This result suggests that sEPSC rates at 36°C arose from conducted action potentials (Fig. 4D). Thus glutamate terminals directly contacting Low neurons are not themselves intrinsically thermosensitive, but the neurons received conducted action potentials from other thermosensitive neurons (e.g., High neurons). Our evidence suggests that physiological temperatures directly generate a high tonic level of glutamate release at the majority of central terminals within the superficial lamina of Vc that can activate local circuits.

To understand better the intrinsic thermal synaptic responses at Vc neurons, all subsequent tests measured mEPSCs in TTX (n = 54). On average, 3/4 of Vc neurons were classified as High with mEPSC rates 2.8-fold greater at 32°C than the remainder, which were Low neurons (Fig. 5A). The thermal relations for mEPSC frequencies of High neurons (n = 36) were well fit by a simple, least-squares linear regression (slope = 2.0 ± 0.2 Hz/°C, r² = 0.94; Fig. 5B) and showed substantial temperature sensitivity (Q₁₀ = 17.8). In contrast, the average mEPSC rate in Low neurons (n = 18) was not altered by temperatures between 30 and 36°C as evidenced by a horizontal slope (0.0 ± 0.1 Hz/°C; r² = −0.20) and thermal coefficient (Q₁₀ = 3.2 ± 1.2; Fig. 5, A and B). However, it should be noted that at the lowest temperatures tested (30°C), the mEPSC rate of High, temperature-sensitive neurons equaled that of Low, temperature-insensitive neurons, i.e., cooling made these very different neurons appear similar.

Fig. 5.

Glutamate release in Vc is either temperature-sensitive or -insensitive. A: TTX-isolated, direct inputs to Vc for High (red, n = 36) or Low (blue, n = 18) neurons; mEPSC frequency is temperature-sensitive or -insensitive, respectively. Significant increases in mEPSC rate over the basal frequency at 32°C were observed when the bath was ≥34°C (*P ≤ 0.05, ANOVA) for High inputs. Increasing or decreasing bath temperature to 36 or 30°C, respectively, from 32°C did not significantly change mEPSC rates in Low neurons (blue bar; P > 0.05, ANOVA). Data are represented as means (±SE). B: linear regression analysis of Vc neurons confirmed thermal sensitivity of spontaneous glutamate release from High neurons (red) from Low neurons (blue). Dotted lines represent the 95% confidence interval. C: localization map of recording sites for temperature-sensitive (red circles) and temperature-insensitive (blue triangles) neurons. Thermally defined populations overlapped throughout the Vc. Scale bar = 100 μm. D: transient receptor potential vanilloid type 1 channel (TRPV1) immunoreactivity is restricted to the superficial lamina of Vc in the horizontal slice, paralleling the recording zone in C. Scale bar = 250 μm. C and D are orientated such that rostral is to the left and midline is at the bottom. spV, spinal trigeminal tract; Vi, trigeminal nucleus interpolaris. E: summary pie charts for temperature (left; n = 54)- and capsaicin (CAP; right; n = 26)-responsive neurons in Vc. Color coding is such that: red, temperature-sensitive (Temp+); blue, temperature-insensitive (Temp−); white, CAP-sensitive (CAP+); and yellow, CAP-insensitive (CAP−).

Histological plots of the location of High and Low neurons showed an indistinct, intermixed distribution within the superficial laminae of Vc (Fig. 5C). Staining for TRPV1 showed a band of immunoreactivity largely limited to within the superficial lamina of Vc, and this TRPV1 distribution largely overlapped with that of the recorded neuron locations (Fig. 5D). This proximity suggested that CAP sensitivity may correlate with thermal sensitivity as previously observed in NTS (Shoudai et al. 2010). Although the majority of Vc neurons (67%, n = 36 of 54; Fig. 5E) were thermally responsive to our tests, these temperatures were well below the conventional minimum threshold of 42°C for TRPV1 activation. Exposing neurons to CAP (100 nM; TTX) significantly increased the frequency of mEPSCs in nearly 3/4 of Vc neurons by an average of ≥5-fold (32°C, TTX: 6.3 ± 2.9 Hz, TTX+CAP: 45.8 ± 9.1 Hz; P < 0.01, Student's t-test). Event amplitudes were unchanged during CAP in most Vc neurons (73%, n = 19 of 26; Fig. 5E), and the holding current was not significantly altered (TTX: −55.0 ± 24.3 pA; CAP: −95.4 ± 34.6 pA; P = 0.31). Both of these observations are consistent with a presynaptic localization of and action at TRPV1 receptors.

To examine the association of TRPV1 with thermal release of glutamate, we next tested CAP and temperature challenges in the same Vc neurons. In detailed studies, 20 Vc neurons completed the full test protocol on mEPSCs (Figs. 6 and 7). CAP activated mEPSCs, indicating that ∼60% of the neurons received TRPV1-expressing inputs (TRPV1+ neurons; Fig. 6). Surprisingly, only about half of these TRPV1+ neurons had High basal mEPSC rates and significant temperature sensitivity (Fig. 6, A, C, and D; n = 6). The remaining TRPV1+ neurons were thermally insensitive, Low neurons with minimal basal mEPSC rates (Fig. 6, B, E, and F; n = 6). CAP increased mEPSC rates similarly in these subsets of High and Low neurons (Fig. 6, C and E, respectively). The remaining 40% of neurons in this group were TRPV1− (insensitive to CAP), but, like the dichotomy observed in TRPV1+ recordings, these TRPV1− inputs were either temperature-sensitive (Fig. 7, A, C, and D; n = 6) or not (Fig. 7, B, E, and F; n = 2). TRPV1− neurons included both High and Low based on basal mEPSC rates (Figs. 7 and 8). Interestingly, the temperature required to increase mEPSC rate during warming in neurons with thermally sensitive inputs occurred at a lower value in TRPV1+ neurons (34°C) than in TRPV1− neurons (36°C; Figs. 6D and 7D; P < 0.05). Such paired observations of Vc neurons indicate that neither high intrinsic mEPSC rate nor high temperature sensitivity was exclusively associated with TRPV1 expression and vice versa, resulting in the highly heterogeneous groups of glutamatergic synapses (Fig. 8).

Fig. 6.

TRPV1 expression does not correlate to intrinsic rates of release or thermal sensitivity. Representative traces are of superficial Vc mEPSC recordings with High (A; red bars) or Low (B; blue bars) basal release rates in TTX at 32°C (top), at 36°C (middle), and in 100 nM CAP (bottom). Inputs were considered to be CAP-sensitive and, therefore, TRPV1-expressing if application of CAP resulted in a mEPSC rate more than twice that observed in TTX, 32°C. CAP evoked a ≥5-fold increase in mEPSC rate High (C; n = 6). In these same recordings, temperature manipulation directly induced increases or decreases in mEPSC rate when warming or cooling steps, respectively, were applied to High neurons (D). Despite similar responses to CAP application (E), Low neurons with TRPV1 did not have altered rates of glutamate release across 30–36°C (F). CAP (100 nM) was applied at the end of Vc recordings in the presence of TTX (Student's t-test). CAP effects on mEPSCs were analyzed independently from thermally evoked responses (ANOVA). Data in C–F represent the means ± SE. Statistical significant is denoted by *P ≤ 0.05, **P ≤ 0.01.

Fig. 7.

Thermal sensitivity persists in neurons lacking TRPV1. Representative traces are of superficial Vc mEPSC recordings insensitive to CAP with High (A; orange bars) or Low (B; green bars) basal release rates in TTX at 32°C (top), at 36°C (middle), and in 100 nM CAP (bottom). Inputs were considered to be CAP-insensitive if application of CAP did not increase the mEPSC rate more than twice that observed in TTX, 32°C. CAP failed to evoke enhanced mEPSC rates in 8 recordings. In TRPV1− recordings with a High mEPSC rate (C; n = 8), temperature changes directly influence mEPSC rate when warming or cooling steps were applied, respectively (D). E and F: rarely, Low neurons were encountered without TRPV1 that did not have altered rates of glutamate release across 30–36°C (F). CAP (100 nM) was applied at the end of Vc recordings in the presence of TTX (Student's t-test). CAP effects on mEPSCs were analyzed independently from thermally evoked responses (ANOVA). Data in C–F represent the means ± SE. Statistical significance was *P ≤ 0.05.

Fig. 8.

Vc synaptic populations with respect to temperature-sensitive glutamate release and TRPV1 expression. Summary of the proportion of neurons with mEPSC rates sensitive to modest temperature (Temp+) changes and/or CAP application (TRPV1+) is shown. CAP sensitivity was not predictive of thermal sensitivity such that 4 populations were identified: Temp+/TRPV1+ (red), Temp−/TRPV1+ (blue), Temp+/TRPV1− (orange), and Temp−/TRPV1− (green). CAP (100 nM) was applied at the end of Vc recordings in the presence of TTX, and subsequent effects on mEPSCs were analyzed independently from thermally evoked responses.

Spontaneous glutamate release increased with age in neurons in the spinal dorsal horn (Baccei et al. 2003). In Vc neurons, we compared mEPSC rates among postnatal development ages: P7–13 (week 2), P14–20 (week 3), and P21–28 (week 4). At physiological temperatures (36°C), the absolute mEPSC rates were significantly higher in recordings from P14–20 (n = 15) and P21–28 (n = 8) compared with those at P7–13 (n = 15; P < 0.05, 2-way ANOVA; Fig. 9A). No differences in mEPSC rates were observed from thermally insensitive inputs (Temp−; P > 0.05, 2-way ANOVA; Fig. 9B). Event amplitudes were similar with temperature or age (P > 0.05, 2-way ANOVA; Fig. 9C). The proportion of neurons that were Temp+ and Temp− was uniform across this developmental range (60–65%; Fig. 9D). Last, CAP evoked similar significant increases in mEPSC rates across all age groups (P < 0.01 vs. TTX, P = 0.31 among ages, 2-way ANOVA, Bonferroni post hoc; Fig. 9E). Together, these data suggest that similar cohorts of temperature-sensitive afferents contact Vc neurons but that the thermal mechanism matures during the 1st 2 wk of life.

Fig. 9.

Thermally sensitive miniature glutamate release is enhanced during postnatal development. A: the rate of thermally evoked miniature glutamate release is significantly greater in rats older than 14 days (gray, P14–20 postnatal days, n = 15; black, P21–28, n = 8) compared with P7–13 (white, n = 15) at 36°C (*P < 0.05, 2-way ANOVA). B: in contrast, postnatal development does not change the rate of miniature release onto neurons from thermally insensitive inputs (P7–13, n = 8; P14–20, n = 9; P21–28, n = 4). C: amplitudes of mEPSCs were similar across developmental ages and input types. D: the percentage of Vc cells with either temperature-sensitive (black; Temp+) or -insensitive (white; Temp−) inputs was unchanged across the development age range from P7 to P28. E: responses of mEPSC rates to CAP (100 nM) were similar regardless of age (P7–13, n = 8; P14–20, n = 16; P21+, n = 3; P > 0.05, ANOVA). CAP increased rates in all TRPV1+ neurons (*P < 0.05). All experiments were performed in TTX (1 μM) and gabazine (3 μM). NS, not significant.

DISCUSSION

Transmission of sensory information from craniofacial regions to Vc neurons relies on the release of glutamate. Measuring sEPSCs and mEPSCs of Vc neurons, we found that: 1) basal rates of glutamate release identified two distinct neuron populations, either High- or Low-activity Vc inputs; 2) the starkly higher mEPSC rate of High neurons near physiological temperature (36°C) was eliminated by cooling neurons to 30–32°C; 3) thermally triggered glutamate release on warming was responsible for most spontaneous release at High neurons via a presynaptic action in which amplitudes were unchanged; 4) thermal sensitivity of Low neurons was minor and indirect (eliminated by TTX); and 5) a non-TRPV1-mediated mechanism drives thermally regulated glutamate release in some TRPV1− Vc neurons. Our tests suggest at least four distinct groups of Vc neurons based on the intrinsic characteristics regulating glutamate release within the superficial laminae of Vc. These glutamate terminals may be a new, central integrative locus in Vc in which presynaptic mechanisms excite postsynaptic neurons even in the absence of conducted afferent discharge. This ongoing activity in normal conditions might be modified by other signaling mechanisms, including GPCRs.

Rate of spontaneous glutamate release classifies Vc synapses.

Spontaneous neurotransmitter release has conventionally been thought to reflect a low-probability, spontaneous expulsion of synaptic vesicles from a pool of readily releasable vesicles primed for release by action potentials (Kaeser and Regehr 2014). Growing evidence suggests that spontaneous release in some central synapses represents neuronal communication independent of evoked release (Glitsch 2008; Kavalali et al. 2011). Basal release is exceedingly low (<0.02 Hz) at many central neurons (Kavalali et al. 2011). In the superficial Vc, in contrast, basal EPSC frequencies ranged from 1 to nearly 20 Hz (Grudt and Williams 1994; Inoue et al. 2012; Jennings et al. 2003). We report that, within this broad range of basal sEPSC rates, Vc neurons distinctly divided into either High (≥10 Hz) or Low (<10 Hz), but this distinction disappeared with subphysiological cooling. Modest temperature changes altered the sEPSC rates only in High Vc neurons and likewise altered postsynaptic action potential rates. In TTX, only mEPSC frequency changed, indicating an intrinsic presynaptic mechanism with high thermal sensitivity (High Q₁₀ = 17.8) that drives glutamate release. Since cooling to 30°C rendered the two populations indistinguishable, it is unlikely that differing numbers of synaptic contacts was responsible for the High/Low distinctions (Sedlacek et al. 2007). This thermally sensitive release mechanism in Vc resembles that of TRPV1+ solitary tract afferents in medial NTS neurons (Shoudai et al. 2010).

Origins of central thermal sensitivity.

Most biological processes including neurotransmission are affected by temperature but often only weakly (Q₁₀ < 3) or not at all between 33 and 38°C (Klyachko and Stevens 2006). In contrast, the high thermal coefficient of glutamate release on Vc neurons suggests a relatively narrow set of potential molecular drivers, chiefly the most sensitive thermoTRPs (Caterina 2007); the best-characterized member of these is TRPV1. TRPV1 is expressed throughout the trigeminal afferents, extending to their presynaptic terminals in the superficial Vc, thereby identifying primary afferent fibers (Bae et al. 2004; Braz and Basbaum 2010; Cavanaugh et al. 2011a,b; Davies and North 2009; Jennings et al. 2003). In our system, 70% of Vc neurons were CAP-sensitive (i.e., TRPV1+). Such TRPV1+ neurons in Vc were evenly divided in their afferent sensitivity to physiological temperatures and basal rates of release. Thermal sensitivity of afferents in some Vc neurons might reasonably be linked to TRPV1 similar to those in the NTS or dorsal motor nucleus of the vagus (Anwar and Derbenev 2013; Shoudai et al. 2010). However, some TRPV1+ Low Vc neurons failed our modest thermal challenges; yet we cannot rule out the possibility that TRPV1 was thermally active at >42°C, which is more typical of TRPV1 in heterologous expression systems and dorsal root ganglion neurons (Caterina et al. 1997). An additional set of neurons had quantitatively similar thermal sensitivity as TRPV1+ neurons but tested negatively with CAP. The presence of thermally sensitive sEPSCs without TRPV1 points to expression of another thermosensor as responsible in this subpopulation in trigeminal inputs. A comparative analysis of TRP mRNA in trigeminal ganglion (TG) and dorsal root ganglion revealed regionally distinct patterns of expression of individual TRP channels (Vandewauw et al. 2013). Transcripts in TG identified TRPV4, TRPM5, and TRPM3, all of which are activated at physiological temperatures (Caterina 2007; Talavera et al. 2005; Vriens et al. 2011). Unfortunately, these additional thermoTRPs are difficult to study functionally outside of expression systems due to lack of selective agonists and antagonists, and thus their evaluation will necessitate genetic models and other strategies.

Sources of afferent heterogeneity.

The marked heterogeneity in Vc synaptic responses likely reflects differences in terminal phenotype. We classified one fundamental aspect of primary afferent neurons (TRPV1+), but our data support at least four subclassifications of synaptic responses. Primary trigeminal afferents are quite diverse (Hegarty et al. 2010; Kobayashi et al. 2005), and, therefore, the variety of responses is not surprising. Many peripheral sensory responses, however, can show primary afferent nerve activation with decreases in temperature described in the spinal cord (Wrigley et al. 2009) and during extracellular in vivo recordings from the Vc (Kurose and Meng 2013), something that we rarely observed in the central responses (1 of 151 recordings). Remarkably, increases in temperature directly increased glutamate release in most of the Vc neurons tested. In addition to primary trigeminal afferents, Vc neurons receive input from cervical dorsal root ganglia neurons or IX and X cranial nerves with soma in the jugular and nodose ganglia (Beckstead and Norgren 1979; Pfaller and Arvidsson 1988). This raises the interesting possibility that the peripheral and central portions of primary afferents may have different thermal thresholds and sensitivities that correspond with their systemic source. Alternatively, these temperature-sensitive and -insensitive, TRPV1− neurons may originate from central neurons creating a complex synaptic arrangement. Glutamate released in Vc may arise from interneurons (Kato et al. 2009) comprising internuclear and -laminar connections (Hamba and Onimaru 1998; Han et al. 2008; Onodera et al. 2000) as well as descending pathways (Aicher et al. 2012; Sato et al. 2013). Knowledge of the degree of synaptic convergence of such afferent nerves is limited in Vc but the varied thermal and CAP responses could reflect these diverse sources (Zanotto et al. 2007). Combining anatomic approaches with patch-clamp recordings may help identify distinct Vc synapses and pathways (Gracheva et al. 2011; Malin et al. 2011).

Physiological implications and pharmacological considerations.

The physiological ramifications of spontaneously released glutamate are unclear. The very small temperature shifts and the physiological range suggest that central brain temperature will significantly impact spontaneous excitatory transmission, and those signals will generate postsynaptic action potentials at normal temperatures in Vc. Our observations that thermal sensitivity increases with age further indicate an important role for such tightly regulated spontaneous glutamate release in Vc synaptic relay that is not an artifact or paradigm-dependent. Recent evidence has suggested that such synaptically active terminals, even in the absence of evoked afferent activity, serve to shape and maintain synaptic connections (McKinney et al. 1999) and can be regulated in isolation from evoked release (Kavalali et al. 2011). Since trigeminal afferents share some synaptic characteristics with both solitary tract and spinal afferents, it is plausible that spontaneous release is subject to modulation independent of evoked release; studies combining examination of electrically evoked and thermally regulated spontaneous release within recordings will address this question directly. This thermally driven, spontaneous release is modulated by GPCRs (Fawley et al. 2011), suggesting that, despite being autonomous, it can be influenced by other neural inputs and may participate in network signaling activity to affect other brain regions.

Conclusions.

Prevailing normal temperatures tonically activate thermally sensitive afferents in most superficial Vc neurons without action potential-induced afferent activation. Thus synaptic transmission in the superficial lamina of Vc is strikingly diverse compared with other regions that receive primary afferent input such as the spinal dorsal horn (Bereiter et al. 2000) and the NTS (Shoudai et al. 2010). Distinct properties of spontaneous glutamate release including basal rate, thermal sensitivity, and CAP responses identified subpopulations of Vc synapses. The autonomous nature of spontaneous glutamate signaling may provide unique synaptic signals differentiating particular circuits that may be altered in pathological states and subject to pharmacological regulation.

GRANTS

This project was supported by grant numbers F32-DE-022499 (T. M. Largent-Milnes) and DE-12640 (S. A. Aicher) from the National Institute of Dental and Craniofacial Research and HL-105703 (M. C. Andresen) from the National Heart, Lung, and Blood Institute. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Dental and Craniofacial Research, the National Heart, Lung, and Blood Institute, or the National Institutes of Health.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

T.M.L.-M., D.M.H., S.A.A., and M.C.A. conception and design of research; T.M.L.-M. performed experiments; T.M.L.-M. analyzed data; T.M.L.-M., S.A.A., and M.C.A. interpreted results of experiments; T.M.L.-M. and D.M.H. prepared figures; T.M.L.-M., S.A.A., and M.C.A. drafted manuscript; T.M.L.-M., D.M.H., S.A.A., and M.C.A. edited and revised manuscript; T.M.L.-M., D.M.H., S.A.A., and M.C.A. approved final version of manuscript.