TRPV1 enhances cholecystokinin signaling in primary vagal afferent neurons and mediates the central effects on spontaneous glutamate release in the NTS

Abstract

The gut peptide cholecystokinin (CCK) is released during feeding and promotes satiation by increasing excitation of vagal afferent neurons that innervate the upper gastrointestinal tract. Vagal afferent neurons express CCK1 receptors (CCK1Rs) in the periphery and at central terminals in the nucleus of the solitary tract (NTS). While the effects of CCK have been studied for decades, CCK receptor signaling and coupling to membrane ion channels are not entirely understood. Previous findings have implicated L-type voltage-gated calcium channels as well as transient receptor potential (TRP) channels in mediating the effects of CCK, but the lack of selective pharmacology has made determining the contributions of these putative mediators difficult. The nonselective ion channel transient receptor potential vanilloid subtype 1 (TRPV1) is expressed throughout vagal afferent neurons and controls many forms of signaling, including spontaneous glutamate release onto NTS neurons. Here we tested the hypothesis that CCK1Rs couple directly to TRPV1 to mediate vagal signaling using fluorescent calcium imaging and brainstem electrophysiology. We found that CCK signaling at high concentrations (low-affinity binding) was potentiated in TRPV1-containing afferents and that TRPV1 itself mediated the enhanced CCK1R signaling. While competitive antagonism of TRPV1 failed to alter CCK1R signaling, TRPV1 pore blockade or genetic deletion (TRPV1 KO) significantly reduced the CCK response in cultured vagal afferents and eliminated its ability to increase spontaneous glutamate release in the NTS. Together, these results establish that TRPV1 mediates the low-affinity effects of CCK on vagal afferent activation and control of synaptic transmission in the brainstem.

NEW & NOTEWORTHY Cholecystokinin (CCK) signaling via the vagus nerve reduces food intake and produces satiation, yet the signaling cascades mediating these effects remain unknown. Here we report that the capsaicin receptor transient receptor potential vanilloid subtype 1 (TRPV1) potentiates CCK signaling in the vagus and mediates the ability of CCK to control excitatory synaptic transmission in the nucleus of the solitary tract. These results may prove useful in the future development of CCK/TRPV1-based therapeutic interventions.

INTRODUCTION

The gut peptide cholecystokinin (CCK) acutely controls food intake by facilitating satiation and meal termination (1–4). Following ingestion of fats and proteins, enteroendocrine cells of the duodenum respond by secreting CCK into the surrounding epithelial tissue and blood (5–8). CCK acts locally at the peripheral terminals of vagal afferent fibers in the gut to cause calcium influx and depolarization in these neurons (9–11). This increases the rate of action potential firing within vagal afferent fibers (12) that project to the nucleus of the solitary tract (NTS) (13, 14), thereby communicating ingestive information to the brainstem. The CCK1 receptor (CCK1R) is located not only at peripheral terminals (14) but also at central terminals of vagal afferent neurons (15, 16). Within the NTS, endocrine or NTS-derived CCK acts presynaptically on vagal afferent terminals to increase the rate of spontaneous excitatory neurotransmission (17, 18).

While the effects of CCK have been studied for decades, the cellular mechanisms linking CCK1R activation to neuron depolarization remain unclear. In pancreatic tissue, CCK1R signaling has clear high- and low-affinity binding linked to differential signaling cascades (19). However, in vagal afferent neurons the effects of high versus low CCK binding affinity are difficult to resolve, especially given that the signaling cascades and calcium influx pathways are not the same as they are in the pancreas (20). It is widely accepted that CCK1R activation of vagal afferent neurons recruits a ruthenium red-sensitive conductance in the plasma membrane, leading to depolarization and subsequent calcium influx (4, 21). However, the specific pathways mediating the ruthenium red component of the CCK response remain debatable. Ruthenium red blocks both L-type voltage-gated calcium channels and transient receptor potential (TRP) channels, suggesting that either of these channels could mediate CCK1R activation of vagal afferents. Therefore, conclusive evidence distinguishing the responsibility of one pathway over another is still lacking. Vagal afferent neurons abundantly express a range of TRP channels, none more robustly than the capsaicin receptor transient receptor potential vanilloid subtype 1 (TRPV1). TRPV1 is a nonselective cation channel activated by various stimuli including temperature, pH, and signaling downstream of G protein-coupled receptors (GPCR) (22). A long history of research exists connecting vagal afferent CCK signaling to TRPV1.

Numerous studies have shown that CCK acts to control food intake by targeting unmyelinated (C-fibers) and lightly myelinated (Aδ-fibers) vagal afferent fibers that also express TRPV1. In fact, capsaicin-mediated lesioning of TRPV1-expressing (TRPV1+) fibers, either peripherally or centrally, significantly attenuates the reduction in food intake caused by systemically administered CCK (23–26). Additionally, these fibers were shown to control CCK-mediated inhibition of gastric motility and delay of gastric emptying following food intake (26, 27). Unsurprisingly, cellular and histological studies revealed that around 80% of CCK-sensitive vagal afferent neurons coexpress TRPV1 (28, 29). Taken together, these findings suggest that CCK-mediated reduction in food intake is largely controlled by TRPV1+ vagal afferent neurons.

The roles of TRPV1 in somatosensory neurons are generally clearer, as the channel serves as a detector of high heat and other noxious stimuli to convey nociception and chemoreception (among other signals) (30, 31). However, determining the physiological role of TRPV1 in vagal afferent neurons has been more difficult. We have previously determined that one clear role of TRPV1 is to drive quantal neurotransmitter release from vagal afferent terminals onto NTS neurons (32–34); this form of neurotransmission in the NTS is also controlled by CCK. Additional evidence suggests that in cultured vagal afferent neurons, the ability of CCK to induce calcium influx depends on the expression of TRP channels, possibly including TRPV1 (4, 20). Given that vagal afferent neurons express a variety of GPCRs (such as the CCK1R) that bind hormones related to food intake (35), we hypothesize that TRPV1 serves as an important point of cellular integration for hormonal signals and the control of food intake. Experiments in this report use a combination of cellular, pharmacological, and genetic approaches to investigate the relationship between CCK1R signaling and TRPV1 in primary vagal afferent neurons.

MATERIALS AND METHODS

Animals

Adult male Sprague-Dawley (SD) rats (350–500 g) were obtained from Envigo or our in-house colony originally derived from Envigo-sourced rats. Adult (3–5 wk old) C57BL/6J and B6.129X1-Trpv1^tm1Jul/J (TRPV1 knockout) mice were obtained from Jackson Laboratories. Males were used to isolate the contribution of TRPV1 as CCK1 receptor signaling is known to be modulated by estrogen signaling (36). Animals were maintained under a standard 12:12-h light-dark cycle in a temperature-controlled (23 ± 1°C) room with ad libitum access to water and standard pellet chow. All experiments were performed in accordance with procedures approved by the Institutional Animal Use and Care Committee at Washington State University.

Nodose Ganglia Isolation and Primary Neuronal Culture

Nodose ganglia were bilaterally isolated from rats and mice under a deep plane of anesthesia (ketamine, 25 mg/100 g; with xylazine, 2.5 mg/100 g) as previously reported (20). Following a midline incision in the neck, the musculature was retracted, and blunt dissection techniques were used to dissociate the vagal trunk from the common carotid artery. The nodose ganglia were removed and placed in ice-cold Hank’s balanced salt solution (HBSS; Invitrogen). The ganglia were desheathed (only in rats), minced, and then digested with 1 mg/mL of both collagenase (Sigma) and dispase (Roche) in calcium and magnesium-free HBSS (Invitrogen) (90 min at 37°C in 95% air-5% CO₂). Following digestion, the tissue was dissociated by gentle trituration through a siliconized pipette (Sigmacote; Sigma), washed by centrifuging (2 min, 1,040 rpm) two times with Dulbecco’s modified Eagle’s medium (DMEM; Invitrogen) supplemented with 10% fetal bovine serum (Invitrogen) and 1% penicillin-streptomycin (Invitrogen). Cells were then plated onto poly-l-lysine (Sigma) coated coverslips (1–2 h at 37°C in 95% air-5% CO₂). Once adhered, the cell cultures were maintained in DMEM supplemented with fetal bovine serum and penicillin-streptomycin (37°C in 95% air-5% CO₂). Ratiometric calcium imaging was performed within 24 h of isolation.

Ratiometric Calcium Imaging

As previously reported (37, 38), cultured nodose ganglia neurons were loaded with the fluorescent calcium indicator fura-2 AM (1 µg/mL standard bath, 0.1% DMSO; Invitrogen) for 30–60 min. Coverslips containing the neurons were placed into the recording chamber and constantly perfused with room temperature (23°C) physiological saline containing the following (mM): 140 NaCl, 5 KCl, 2 CaCl₂, 1 MgCl₂, 10 HEPES, and 6 dextrose. Saline was brought to a pH of 7.40 using 2 M Tris base. Fluorescence was measured using a Nikon Eclipse Ti inverted microscope (Nikon Instruments Inc., Melville, NY) with a ×40 oil immersion objective and an Andor Zyla sCMOS digital camera (Andor, South Windsor, CT). Neurons loaded with fura-2 AM were alternately excited at 340 nm and 380 nm, and emission was measured at 510 nm. Images were acquired every 6 s and analyzed with Nikon Imaging Software by calculating the ratio of emission under 340/380 nm of light for each region of interest (ROI). ROIs were captured as a circular region around the cell body; we then averaged the fluorescence ratio within the ROI for a single cell. Ratios were converted to the corresponding calcium concentration by using a previously established standard curve with known calcium concentrations. The raw light data and calculated ratios are compared to the standard curve to derive the intracellular calcium concentrations. Neuron viability was confirmed through depolarization with standard bath containing elevated potassium concentration (Hi-K+, 55 mM KCl with equimolar reduction to 90 mM NaCl).

Chemicals and Drugs

Drugs were purchased from retail distributers including sulphated cholecystokinin (CCK; Peptides International, no. PCK-4100-v), capsaicin (CAP; Tocris, no. 0462, Sigma, no. M2028), ruthenium red (RuR; Cayman Chemical; no. 14339; Sigma; no. R2751), SB366791 (SB; Cayman Chemical; no. 11019), and JNJ17203212 (JNJ; Cayman Chemical; no. 30930). These were stored as stock aliquots and diluted in fresh standard bath to their final concentration on the day of recording. The salts used for making bath solutions were purchased from Sigma-Aldrich. All drugs were delivered via bath perfusion at a flow rate of ∼2 mL/min.

Molecular Cloning and COS-7 Cell Transfection

To generate TRPV1-GFP and CCK1R-GFP, the full-length rat coding regions (NCBI accession number: TRPV1 NM_031982.1 and CCK1R NM_012688.3) without stop codons were PCR amplified from nodose ganglion cDNA using the primers below and inserted into the HindIII and EcoRI sites of pEGFP-N1 (Clontech).

Primer sequences were as follows: TRPV1; 5′- AGCTAAGCTTCCACCATGGAACAACGGGCTAGC-3′ and 5′- CCATGAATTCCTTTCTCCCCTGGGACCAT-3′; and CCK1R: 5′- GGGGACAAGTTTGTACAAAAAAGCAGGCTGCCACCATGAGCCATTCACCAGCTCGCCAGC-3′ and 5′- GGGGACCACTTTGTACAAGAAAGCTGGGTCGGGGGGTGGAGCAGAGGTG-3′.

Clonal COS-7 cells (ATCC, Manassa, VA) were plated on glass coverslips at a density of 250,000 cells per well of a 6-well culture plate. Cells were maintained in DMEM (Invitrogen), 10% fetal calf serum (FCS, Atlanta Biologicals), 25 units/mL penicillin (pen), and 25 µg/mL streptomycin (strep; Sigma). Approximately 24 h after plating, media were replaced with DMEM, 10% FCS without pen/strep, and cells were transiently transfected overnight. COS-7 cells were transfected with either 2 mg of CCK1R-GFP, 2 µg of TRPV1-GFP, or both and 2.4 mL lipofectamine 2000 (Invitrogen) per well. Transfected COS-7 cells were subsequently used for calcium imaging over the following 1–2 days.

Horizontal Brainstem Slice Preparation

Brainstem slices were isolated from rats and mice deeply anesthetized with isoflurane as previously described (39). The brainstem was removed from the occipital crest to the first cervical vertebrae and placed in ice-cold artificial cerebrospinal fluid (aCSF) containing the following (mM): 125 NaCl, 3 KCl, 1.2 KH₂PO₄, 1.2 MgSO₄, 25 NaHCO₃, 2 CaCl₂, and 10 dextrose, bubbled with 95% O₂-5% CO₂. aCSF was brought to a pH of 7.40 using 1 M HCl. Once chilled, the tissue was cut to remove the cerebellum, and the tissue block was mounted horizontally to a pedestal with cyanoacrylate glue and submerged in cold aCSF on a vibrating microtome (Leica VT1200S). Approximately 150–200 μm were removed from the dorsal surface, and then a single 250-μm-thick horizontal slice was collected containing the solitary tract (ST) along with the neuronal cell bodies of the medial NTS region. Slices were cut with a sapphire knife (Delaware Diamond Knives, Wilmington, DE) and secured using a fine polyethylene mesh in a perfusion chamber with continuous perfusion of aCSF bubbled with 95% O₂-5% CO₂ at 32°C.

Whole Cell Patch-Clamp Electrophysiology

Recordings were performed on NTS neurons in horizontal brainstem slices using an upright Nikon FN1 microscope with a Nikon DS-Qi1Mc digital camera and NIS-elements AR imaging software. Recording electrodes (3.0–4.5 MΩ) were filled with an intracellular solution containing the following (mM): 10 CsCl, 130 Cs-methanesulfonate, 11 EGTA, 1 CaCl₂, 2 MgCl₂, 10 HEPES, 2 Na₂ATP, and 0.2 Na₂GTP. Excitatory postsynaptic currents (EPSCs) were measured under voltage-clamp conditions with a MultiClamp 700 A amplifier (Molecular Devices, Union City, CA) and held at V_H = −60 mV in whole-cell patch configuration. Only recordings with a series resistance of <20 MΩ were used for experiments to ensure good access and maintenance of voltage clamp. Signals were filtered with a 1-kHz bezel filter and sampled at 20 kHz using Axon pClamp10 software (Molecular Devices).

Data Analysis

For calcium imaging recordings, traces of calcium concentration across time were acquired with Nikon Imaging Software and analyzed in Excel (Microsoft, Redmond, WA). Changes in cytosolic calcium (ΔCa²⁺) were calculated by taking the difference between the peak response to a drug and the average baseline one minute before drug application. A response was considered positive if the peak calcium level was >15 nM above baseline. Area under the curve (AUC) was calculated by taking the cumulative area during the response period multiplied by the 6-s bin size. Time constant (tau) was calculated using the decay fit function with an event detection and analysis program (MiniAnalysis, Synaptosoft, Decatur, GA) as follows: calcium imaging data from Excel were saved as a tab delimited file and converted to an axon binary file (abf) 1.8 integer by using Clampfit 10 (Molecular Devices). Positive events that were previously identified were selected and fit for their decay kinetics using the curve fitting function. Responses were fit from the peak to the end of the steady state with the following parameters in MiniAnalysis: aligned to peak, double exponential decay, 5,000 iterations, 0.001 tolerance, and 10–90% rise time. For brainstem slice recordings the digitized waveforms of synaptic events were analyzed using an event detection and analysis program (MiniAnalysis, Synaptosoft, Decatur, GA) for all quantal synaptic currents. All events >10 pA were counted for frequency values.

General Statistical Analyses

We calculated statistical comparisons using Sigma Stat software (Systat Software Inc., San Jose, CA). The data were tested for normality and equal variance and the appropriate parametric or nonparametric statistics were used, including ANOVA with post hoc Bonferroni analysis, paired and unpaired t test, linear regression, Mann-Whitney rank sum test, χ², and Kruskal-Willis one-way ANOVA on ranks with post hoc Dunn’s method. The specific tests used are primarily indicated in the figure legends, with additional tests shown in results. Comparisons were considered statistically different with an alpha level of P < 0.05.

RESULTS

Primary Vagal Afferent Neurons Containing TRPV1 Have Enhanced Responses to CCK

Using fluorescent calcium imaging, we measured a CCK concentration-response curve in TRPV1-expressing (TRPV1+) and TRPV1-nonexpressing (TRPV1−) cultured nodose neurons (Fig. 1). We found that increasing concentrations of CCK (0.01–100 nM) produced calcium responses that were noticeably larger in TRPV1+ compared to TRPV1− afferents, in particular at the higher concentrations (Fig. 1A). Expression of TRPV1 was confirmed by a >15 nM increase in cytosolic calcium levels in response to the agonist capsaicin (CAP; 1 µM) following the last CCK treatment (not shown). Overall, there was a significant difference in the responses to CCK in TRPV1+ (n = 33 neurons/3 rats) versus TRPV1− neurons (n = 11 neurons/3 rats) (Fig. 1B). Post hoc analysis confirmed that the peak calcium response was nearly twice as large in TRPV1+ neurons at high (10 and 100 nM CCK) but not low (0.01–1 nM CCK) concentrations of CCK when compared to TRPV1− neurons (Fig. 1B). Fitting of the concentration-response curves using the Hill equation confirmed the differences in calcium response to high concentrations of CCK was also reflected in the estimated maximum (E_max) effect, where the E_max was 1.8 times greater in TRPV1+ neurons (Fig. 1C). The EC₅₀ of the CCK response curve was somewhat greater (as a function of the increased maximum) (Fig. 1D), while the Hill coefficient was slightly lower in TRPV1+ neurons (Fig. 1E); however, neither of these measures were statistically different between the two neuronal types.

Figure 1.

Primary vagal afferent neurons containing transient receptor potential vanilloid subtype 1 (TRPV1) have enhanced responses to cholecystokinin (CCK). A: representative calcium imaging traces from dissociated nodose ganglia neurons that were bath exposed to increasing concentrations of CCK. TRPV1 expression was determined by a response to capsaicin (1 µM; response not shown). B: in CCK responsive neurons, the average delta calcium responses (peak − baseline) were fit with a Hill function in TRPV1− (blue circles, n = 11 neurons/3 rats) and TRPV1+ neurons (red circles, n = 33 neurons/3 rats). There was a significant interaction between TRPV1 expression and CCK concentration to increase calcium influx (n = 44 neurons/3 rats, **P = 0.003, two-way mixed design ANOVA). Post hoc analysis confirmed TRPV1+ neurons had significantly greater responses to 10 nM (*P = 0.02, Bonferroni) and 100 nM CCK (**P = 0.002, Bonferroni). C–E: fit analysis from the sigmoidal models in B showing estimated maximum (E_max; ***P < 0.001, unpaired t test; C), EC₅₀ (P = 0.11, unpaired t test; D), and Hill coefficient (P = 0.13, unpaired t test; E). Data are shown as average ± SE.

Activity and Expression of TRPV1, but Not Transient Receptor Potential Ankyrin Subtype 1, Enhances the Response to CCK in Vagal Afferent Neurons

The responsiveness to CCK and CAP was variable, but we found nodose neurons that had large calcium responses to CCK also produced a large response to CAP (Fig. 2A). When plotted across many neurons we observed a significant positive correlation between the CCK responses and CAP activation of TRPV1 (n = 33 neurons/3 rats) (Fig. 2B). This suggests that not only the expression of TRPV1 but also the magnitude of TRPV1 activation is predictive of the response to CCK. We observed that TRPV1+ neurons had twice as much CCK-induced calcium influx than TRPV1– neurons at the 100 nM CCK challenge (Fig. 2C). The population distribution of CCK and CAP sensitivity shows a major overlap between CCK responses and CAP responses (Fig. 2D), as was expected given the majority (76%) of vagal afferent neurons express TRPV1. We confirmed that 35% of all recorded neurons were CCK sensitive at 100 nM and 74% of the CCK-sensitive neurons were also responsive to CAP (40).

Figure 2.

Activity and expression of transient receptor potential vanilloid subtype 1 (TRPV1), but not transient receptor potential ankyrin subtype 1 (TRPA1), enhances the response to cholecystokinin (CCK) in vagal afferent neurons. A: representative calcium imaging traces from the same dissociated nodose neurons as in Fig. 1B showing the responses to CCK (100 nM) and the TRPV1 agonist capsaicin (CAP; 1 µM). B: proportional relationship between the calcium responses to CAP and CCK (n = 33 neurons/3 rats, slope = 0.33, R² = 0.60, P < 0.001, linear regression). Each data point represents a single neuron that responded to both CCK and CAP. C: responses to CCK in TRPV1+ neurons (n = 33 neurons/3 rats) were significantly greater than those in TRPV1− neurons (n = 11 neurons/3 rats, *P = 0.03, Mann-Whitney rank sum). D: population distribution of CCK and CAP responsive neurons (n = 125 neurons/3 rats, P = 0.98, χ²-test). E: dissociated nodose neurons were exposed to CCK (100 nM) and the TRPA1 agonist allyl isothiocyanate (AITC; 300 µM). There was no correlation between the calcium responses to CCK and AITC (n = 13 neurons/4 rats, slope = 0.16, R² = 0.04, P = 0.51, linear regression). Each data point represents a single neuron that responded to both CCK and AITC. F: responses to CCK in TRPA1− (n = 17 neurons/4 rats) and TRPA1+ neurons (n = 13 neurons/4 rats) were not significantly different (P = 0.65, Mann-Whitney rank sum). G: population distribution of CCK and AITC responsive neurons (n = 101 neurons/4 rats, P = 0.94, χ²-test). Data are shown as average ± SE with individual data points plotted as circles.

It is possible that TRP channels other than TRPV1 could enhance CCK-induced calcium influx given the broad expression of TRP channels in vagal afferent neurons. The transient receptor potential ankyrin subtype 1 (TRPA1) ion channel is commonly expressed with TRPV1 (38); therefore, we investigated whether TRPA1 also enhances CCK-induced calcium influx in vagal afferent neurons by using calcium imaging. Unlike in TRPV1+ neurons, there was no correlation between the response to CCK and the response to TRPA1 agonist allyl isothiocyanate (AITC; 300 µM) in CCK-sensitive/TRPA1+ neurons (Fig. 2E). In CCK-sensitive neurons, there was no difference in the calcium response to 100 nM of CCK in TRPA1− (n = 17 neurons/4 rats) versus TRPA1+ neurons (n = 13 neurons/4 rats) (Fig. 2F). Out of all neurons, 46% were TRPA1+ and 43% were CCK-sensitive neurons; of CCK-sensitive neurons, 43% were TRPA1+. Across neurons, CCK responses were not segregated to TRPA1+ neurons (P = 0.94, χ²-test) (Fig. 2G). Therefore, the presence and activity of TRPA1 do not enhance the response to CCK in vagal afferent neurons. Augmentation of CCK-induced calcium influx is characteristic of and specific to TRPV1 signaling in these experiments.

CCK Receptor Desensitization Kinetics Are Independent of TRPV1 Expression

If CCK-mediated calcium influx is greater in TRPV1+ neurons because of differences in the CCK1R itself across these two neuronal types, we would predict that receptor kinetics may be different in TRPV1+ and TRPV1− neurons. We previously demonstrated that acute CCK desensitization occurs as a function of ligand binding concentration at the CCK1R itself, rather than as a result of downstream effectors that enable calcium influx (such as TRP channels) (41). We utilized this analytical approach to determine the role of CCK1R kinetics in contributing to the observed difference in CCK signaling between TRPV1 containing and lacking vagal afferent neurons (Fig. 3). At low concentrations, exposure to CCK results in a modest peak calcium response followed by a period of sustained steady-state signaling, whereas high concentrations of CCK produce rapid and large peak calcium responses with pronounced desensitization to steady state (Fig. 3A). Fitting of the transient peak to steady-state with an exponential decay function provides an estimate of the decay time constant (tau) that is reflective of the rate of CCK1R desensitization (41). At higher concentrations of CCK, receptor desensitization occurs faster as indicated by a smaller tau (Fig. 3, B and C). However, at a given CCK concentration, tau is independent of the magnitude of the peak calcium response (Fig. 3B). If differences in CCK1R desensitization accounted for the potentiated CCK signaling in TRPV1+ afferents, we would expect to see a difference in tau between TRPV1+ and TRPV1− neurons at each given concentration. Instead, we observed no significant difference in the tau between TRPV1+ (n = 33 neurons/3 rats) and TRPV1− (n = 11 neurons/3 rats) neurons at each concentration (Fig. 3C).

Figure 3.

Cholecystokinin (CCK) receptor desensitization kinetics are independent of transient receptor potential vanilloid subtype 1 (TRPV1) expression. A: representative calcium traces in response to two different concentrations of CCK in TRPV1− and TRPV1+ nodose neurons. The red line illustrates the fit of decay curves used in calculating the decay-time constant (tau). B: tau plotted as a function of the calcium response to CCK (black solid line, linear regression; gray dotted lines, 95% confidence interval). There was no correlation between the tau and the magnitude of the CCK response at 1 nM CCK (n = 27 neurons/3 rats, slope = −0.02, adjusted R² < 0.001, P = 0.41, linear regression) and at 100 nM CCK (n = 32 neurons/3 rats, slope = −0.002, adjusted R² = 0.04, P = 0.14, linear regression). C: average tau across CCK concentrations for TRPV1+ (n = 33 neurons/3 rats) and TRPV1− neurons (n = 11 neurons/3 rats). There was a significant main effect of CCK concentration (***P < 0.001, two-way ANOVA) but no significant main effect of TRPV1 expression (P = 0.95) nor an interaction (P = 0.84). Data are shown as average ± SE with individual data points plotted as circles.

Threshold TRPV1 Activation Potentiates the Response to CCK in Vagal Afferent Neurons

To examine whether activating TRPV1 can further enhance the calcium response to CCK, cultured nodose neurons (n = 39 neurons/5 rats) were exposed to 5 min of a sub-maximal concentration of CCK (10 nM), simultaneously exposed to CCK and a threshold concentration of CAP (30 nM), to CCK alone to test for recovery, and to CAP alone (Fig. 4A). While we detected the main effects of CCK and CAP alone, there was no interaction between CCK and CAP on the peak calcium response. Instead, we observed a subadditive effect where the peak response to both drugs together was less than the sum of CCK alone + CAP alone (Fig. 4B). However, when we calculate total calcium influx over time (measured as AUC), we observed a synergistic interaction where CAP treatment enhanced the response to CCK when both drugs were applied simultaneously (Fig. 4C).

Figure 4.

Threshold transient receptor potential vanilloid subtype 1 (TRPV1) activation potentiates the response to cholecystokinin (CCK) in vagal afferent neurons. A: calcium influx was measured in TRPV1+ dissociated nodose neurons that were exposed to CCK alone (10 nM), CCK and capsaicin (CAP) together, and CAP alone (30 nM). B: addition of CAP doubled the peak calcium response to CCK in TRPV1+ neurons; there was a main effect of CAP [***P < 0.001, two-way repeated measures (RM) ANOVA] and of CCK (***P < 0.001, two-way RM ANOVA) compared to controls, but no synergistic interaction on the peak response with both CCK and CAP (n = 39 neurons/5 rats, P = 0.09, two-way RM ANOVA). C: addition of CAP more than doubled the area under the curve (AUC) compared to CCK alone; there was a significant synergistic interaction when CCK and CAP were combined (n = 39 neurons/5 rats, ***P < 0.001, two-way RM ANOVA). B and C plot the first exposure to CCK (CCK¹) or CAP. Data are shown as average ± SE with individual data points plotted as circles.

CCK-Induced Calcium Influx Occurs Independent of Vanilloid-Specific TRPV1 Antagonism in Vagal Afferent Neurons

We next investigated whether the potentiated response to CCK in TRPV1+ afferents was dependent on the vanilloid binding site of TRPV1 by using competitive antagonists (42, 43). We hypothesized that downstream G-protein signaling, following CCK1R activation, acted on the vanilloid binding site of TRPV1. Cultured nodose neurons were treated with CCK (100 nM) to determine their initial response level, pretreated with TRPV1-selective vanilloid site antagonists SB366791 (SB; 10 µM) or JNJ17203212 (JNJ; 10 µM), and then treated with both the antagonist and CCK (Fig. 5, A and D). We found pretreatment with either the SB or JNJ compound did not change the average calcium response to CCK in TRPV1+ nodose neurons (SB: n = 41 neurons/4 rats; JNJ: n = 33 neurons/3 rats) (Fig. 5, B and E, right). Surprisingly, however, the SB and JNJ antagonists did significantly reduce the response to CCK in TRPV1− neurons, suggesting potential off-target effects (SB: n = 12 neurons/4 rats; JNJ: n = 10 neurons/2 rats) (Fig. 5, B and E, left). To verify that SB and JNJ were effective at the concentration used, we inhibited CAP activation of TRPV1 using both compounds (Fig. 5, C and F), confirming their ability to antagonize the vanilloid binding site of TRPV1. From these results, we conclude that if CCK1R interacts with TRPV1, it does so via a mechanism other than the vanilloid binding site as determined by the antagonists used here.

Figure 5.

Cholecystokinin (CCK)-induced calcium influx occurs independent of vanilloid-specific transient receptor potential vanilloid subtype 1 (TRPV1) antagonism in vagal afferent neurons. A: representative calcium trace showing responses to CCK alone (100 nM) or CCK following pretreatment with the TRPV1 antagonist SB366791 (SB; 10 µM). B: SB reduced the response to CCK in TRPV1− neurons (n = 12 neurons/4 rats, **P = 0.005, paired t test) but not in TRPV1+ neurons (n = 41 neurons/4 rats, P = 0.29, Wilcoxon signed rank test). C: neurons were treated with capsaicin (CAP; 30 nM) in the absence or presence of SB to confirm antagonism of the vanilloid binding site of TRPV1 (n = 59 neurons/1 rat, ***P < 0.001, Wilcoxon signed rank test). D: neurons were exposed to CCK alone (100 nM) or following pretreatment with TRPV1 antagonist JNJ17203212 (JNJ; 10 µM). E: JNJ reduced the response to CCK in TRPV1− neurons (n = 10 neurons/2 rats, **P = 0.01, Wilcoxon signed rank test) but not in TRPV1+ neurons (n = 33 neurons/3 rats, P = 0.10, Wilcoxon signed rank test). F: neurons were treated with CAP (30 nM) in the absence or presence of JNJ to confirm antagonism of the vanilloid binding site of TRPV1 (n = 12 neurons/1 rat, ***P < 0.001, paired t test). B and E plot the first exposure to CCK. Data are shown as average ± SE with individual data points plotted as circles.

Ruthenium Red Reduces the Response to CCK in Vagal Afferent Neurons and COS-7 Cells Transfected with CCK1R and TRPV1

Since vanilloid-specific TRPV1 antagonists did not reduce the CCK response in TRPV1+ neurons, we then used a broad TRP channel pore blocker, ruthenium red (RuR), to block TRP channel calcium influx. Cultured nodose neurons were treated with CCK (100 nM), pretreated with RuR (1 µM), and then treated with both RuR + CCK to determine if the CCK-mediated calcium influx was RuR dependent (Fig. 6A). TRPV1 expression was determined by a positive response to CAP (not shown). Pretreatment with RuR reduced the CCK response by 38 ± 12% in TRPV1− neurons (n = 19 neurons/4 rats), and by 28 ± 7% in TRPV1+ neurons (n = 41 neurons/4 rats) (Fig. 6B). This suggests that TRPV1 in addition to other ion channels targeted by RuR mediate CCK-induced calcium influx.

Figure 6.

Ruthenium red reduces the response to cholecystokinin (CCK) in vagal afferent neurons and COS-7 cells transfected with CCK1R and transient receptor potential vanilloid subtype 1 (TRPV1). A: representative calcium trace showing the CCK (100 nM) response alone or following pretreatment with the broad TRP channel pore blocker ruthenium red (RuR; 1 µM). TRPV1 expression was determined by the response to capsaicin (CAP; 1 µM). B: RuR significantly reduced the response to CCK in both TRPV1− (n = 19 neurons/4 rats, **P = 0.004, paired t test) and TRPV1+ nodose neurons (n = 41 neurons/4 rats, ***P < 0.001, Wilcoxon signed rank test). C: COS-7 cells were transfected with the CCK1R alone (n = 91 cells/7 cultures) or in combination with TRPV1 (n = 40 cells/4 cultures). Average calcium traces from both transfections following exposure to CCK (100 nM). D: integrated calcium influx over time was measured by the area under the curve (AUC) and was significantly different across the 3 treatments (***P < 0.001, Kruskal-Willis one-way ANOVA on ranks). The presence of TRPV1 tripled the CCK-induced calcium influx (*P < 0.05, post hoc multiple comparison with Dunn’s method). The potentiating effect of TRPV1 was eliminated by RuR (n = 48 cells/9 cultures, *P < 0.05, post hoc multiple comparison with Dunn’s method). Data are shown as average ± SE with individual data points plotted as circles.

To more directly investigate our hypothesis that TRPV1 enhances CCK-induced calcium influx, we transfected COS-7 cells with the CCK1R alone or in combination with TRPV1 and treated them with CCK (100 nM) (Fig. 6C). This controlled system allowed us to isolate the role of a single TRP channel in the CCK response. Although inserting TRPV1 did not change the peak calcium response to CCK, TRPV1 prolonged cytosolic calcium levels resulting in an AUC that was three times greater in TRPV1+ COS-7 cells (n = 40 cells/4 cultures) compared to TRPV1− COS-7 cells (n = 91 cells/7 cultures) (Fig. 6D). Furthermore, blocking TRPV1 with RuR (1 µM) (n = 48 cells/9 cultures) eliminated the TRPV1-driven component of the AUC and returned total calcium influx to that of the TRPV1− COS-7 cells (Fig. 6D). This provides further evidence that TRPV1 enhances CCK-induced calcium influx.

Genetic Deletion of TRPV1 Attenuates CCK Signaling in Vagal Afferent Neurons

To definitively test the role of TRPV1 on CCK signaling, we next used a genetic approach with TRPV1 knockout (KO) mice (Fig. 7). Cultured nodose neurons were taken from C57BL/6J control mice and TRPV1 KO mice, and soma were imaged with ratiometric calcium imaging. Neurons from control mice were categorized as TRPV1+ by a positive response to CAP (1 µM), whereas the absence of TRPV1 in the TRPV1 KO was confirmed by showing no response to CAP exposure. We treated cultures with increasing concentrations of CCK and found in control mice, as was true in afferents taken from rats, that TRPV1+ neurons (n = 57 neurons/4 mice) had larger CCK responses compared to TRPV1− neurons (n = 5 neurons/2 mice) (Fig. 7A, top traces). In TRPV1 KO mice, the CCK response curve was very similar to the TRPV1− neurons from the control mice except at the highest concentration tested (Fig. 7A, bottom traces). On average across recordings the concentration-dependent effect of CCK on calcium influx was dependent on the cell type, with TRPV1+ neurons having the largest responses to CCK, TRPV1− the lowest, and in TRPV1 KO neurons (n = 54 neurons/3 mice) the CCK responses were indistinguishable from the TRPV1− neurons, except at the highest concentration tested (Fig. 7B). The response to 100 nM CCK was significantly different across the three cell types, with TRPV1+ neurons having the highest response, followed by TRPV1 KO neurons, and TRPV1− neurons (Fig. 7C). We confirmed the neurons from TRPV1 KO animals did not express TRPV1 by treating them with CAP (1 µM) and observing no CAP-induced calcium influx (Fig. 7D). We conclude that TRPV1 mediates a large portion of the CCK induced calcium influx in cultured vagal afferent neurons.

Figure 7.

Genetic deletion of transient receptor potential vanilloid subtype 1 (TRPV1) attenuates cholecystokinin (CCK) signaling in vagal afferent neurons. A: representative calcium imaging traces of dissociated nodose neurons exposed to increasing concentrations of CCK. Responses were measured from TRPV1+ neurons (n = 57 neurons/4 mice) or TRPV1− neurons in C57BL/6J mice (n = 5 neurons/2 mice), and from TRPV1 knockout (KO) mice (n = 54 neurons/3 mice). B: calcium responses (average ± SD) to increasing CCK concentrations across vagal afferent neurons taken from C57BL/6J mice (TRPV1+ or TRPV1− neurons) and TRPV1 KO mice. There was a significant interaction between cell type and CCK concentration to increase calcium influx (***P < 0.001, two-way mixed design ANOVA). There was a significant difference in the response between cell types at 100 nM CCK (***P < 0.001, multiple comparisons with Bonferroni correction). C: the response to 100 nM CCK was significantly different across the three cell types (P < 0.001, Kruskal-Willis one-way ANOVA on ranks), as confirmed by post hoc multiple comparison testing (*P < 0.05, Dunn’s method). D: neurons were treated with capsaicin (CAP; 1 µM) to verify TRPV1 expression. TRPV1+ neurons were significantly activated by CAP while TRPV1− and TRPV1 KO neurons were not (P < 0.001, Kruskal-Willis one-way ANOVA on ranks; *P < 0.05, post hoc multiple comparison with Dunn’s method). Data are shown as average ± SE with individual data points plotted as circles.

CCK Signaling Requires TRPV1 to Control the Frequency of Spontaneous Glutamate Release in the NTS

Finally, we examined the effect of TRPV1 on spontaneous excitatory postsynaptic currents (sEPSCs) in the NTS. The NTS receives direct excitatory synaptic innervation from vagal afferent neurons. TRPV1 and the CCK1R are expressed on the central terminals of vagal afferent neurons and agonists of these receptors are known to increase the frequency of spontaneous glutamate release onto NTS neurons (17, 18, 34). We measured sEPSCs using whole cell patch-clamp electrophysiology on an acute horizontal brainstem slice of the NTS. Currents were recorded from TRPV1+ synapses in SD rats (n = 16 neurons/11 rats), TRPV1− synapses in SD rats (n = 3 neurons/3 rats), TRPV1+ synapses in C57BL/6J mice (n = 11 neurons/6 mice), and TRPV1 KO mice (n = 9 neurons/3 mice) under control conditions and in response to CCK (100 nM) (Fig. 8A). We treated with 1 µM CAP to identify TRPV1− synapses, which had an average sEPSC frequency (±SD) of 1 ± 1 Hz during the control period and CAP treatment. We also confirmed that TRPV1 KO mice were not responsive to CAP (control: 4 ± 5 Hz, CAP 1 µM: 3 ± 4 Hz; P = 0.20, Wilcoxon signed rank test) (Fig. 8B). Consistent with the calcium imaging data (seen in Fig. 2A), there was a trend toward a correlation between the frequency of sEPSCs in response to 100 nM CCK versus 1 µM CAP, as measured by percent of control frequency (Fig. 8B). On average, CCK doubled the frequency of sEPSCs in TRPV1+ rats and mice (Fig. 8B). In contrast, TRPV1− rats and TRPV1 KO mice showed no change in sEPSC frequency in response to CCK (Fig. 8C). It is possible that the synapses recorded from TRPV1 KO mice were CCK insensitive by chance; however, this is unlikely considering that 69% of synapses recorded from C57BL/6J mice were CCK-sensitive. This suggests that the presence of TRPV1 in vagal afferent neurons is responsible for mediating the central effects of CCK at the vagal afferent to NTS synapse.

Figure 8.

Cholecystokinin (CCK) signaling requires transient receptor potential vanilloid subtype 1 (TRPV1) to control the frequency of spontaneous glutamate release in the nucleus of the solitary tract (NTS). Spontaneous glutamatergic excitatory postsynaptic currents (sEPSCs) were recorded using patch-clamp electrophysiology from acute brainstem slices containing the NTS and central vagal afferent terminals that express CCK1Rs. A: representative current traces recorded from TRPV1+ neurons in SD rats, TRPV1− neurons in SD rats, TRPV1+ neurons in C57BL/6J mice, and TRPV1 knockout (KO) mice under control conditions and with bath application of CCK (100 nM). B: trending toward a proportional relationship between the percent change in sEPSC frequency in response to capsaicin (CAP) and CCK (n = 13 neurons/10 rats, slope = 0.07, R² = 0.29, P = 0.06, linear regression). C: CCK significantly increased the sEPSC frequency in TRPV1+ rats (n = 16 neurons/11 rats, ***P = <0.001, paired t test) and TRPV1+ mice (n = 11 neurons/6 mice, ***P < 0.001, paired t test) but not in TRPV1− rats (n = 3 neurons/3 rats, P = 0.12, paired t test) nor TRPV1 KO mice (n = 9 neurons/3 mice, P = 0.09, paired t test). Data are shown as average ± SE with individual data points plotted as circles.

DISCUSSION

The effects of CCK on food intake and satiety have been well established; however, specific receptor-effector coupling of CCK signaling remains unknown. Here we report that the ion channel TRPV1 enhances CCK-induced calcium influx in vagal afferent neurons and mediates the central effects of the CCK1R to control spontaneous glutamate release in the NTS. Using a combination of fluorescent calcium imaging and patch-clamp electrophysiology coupled with pharmacological and genetic tools, we demonstrated that low-affinity binding of CCK recruits TRPV1 to potentiate the magnitude of the CCK response in vagal afferent cell bodies. At the central terminals, the ability of CCK to control synaptic transmission required TRPV1 and was eliminated in TRPV1 KO mice. Taken together, these results provide a mechanism by which the ion channel TRPV1 transduces downstream GPCR signaling from the CCK1R into neurophysiological signals that control vagal afferent activity. This novel coupling expands our understanding of CCK1R signaling under physiological conditions and provides a plausible point of modulation under conditions known to impact vagal CCK signaling such as a high-fat diet (44).

TRPV1 Causally Contributes to CCK1R Signaling

While it has been known for some time that the C-fiber afferents, known to contain TRPV1, are critical in relaying the satiety effects of CCK (23, 45), the observations that CCK efficacy was potentiated in this population (Fig. 1) and proportionate to the CAP response (Fig. 2) was suggestive of a direct role of TRPV1 itself. However, establishing this causal role of TRPV1 was more complicated than suspected. We observed a strong correlation between the responses to a high dose of CAP and CCK. We interpreted these results to be consistent with an interaction between CCK and TRPV1 signaling, although the correlation may be a coincidence and not specific to TRPV1 expression. To address this possibility, we also looked at the effect of TRPA1 signaling on the response to CCK in vagal afferent neurons. We choose TRPA1 as it is an abundant TRP channel in the vagal afferents and is commonly coexpressed with TRPV1 (38). We found the response to CCK was the same in both TRPA1+ and TRPA1− neurons (Fig. 2). We therefore concluded the coupling to CCK does not occur with all TRP channels and may be unique to TRPV1.

If the CCK1R was coupled to TRPV1, it seemed logical that using a TRPV1 antagonist should disconnect the signaling, especially given the link between GPCR activation and phosphoinositide generation (see Ref. 46 for review). Surprisingly, we found that antagonizing the vanilloid site with the specific TRPV1 antagonists SB366791 (SB) and JNJ17203212 (JNJ) had no effect on CCK-mediated calcium influx (Fig. 5). Although, curiously, they did reduce the CCK response in the TRPV1− neurons suggesting a degree of off-target inhibition. While this shows that CCK signaling does not couple via the vanilloid binding site, it does not eliminate the possibility that CCK couples to TRPV1 via an alternative pathway. Phospholipids that are downstream of CCK1R activation along with potential G_q signaling can act at a different site on TRPV1 than the vanilloid site (47), so it is possible that CCK was still able to signal in the presence of TRPV1 antagonists due to allosteric interactions with TRPV1. Alternatively, the CCK1R may be coupling via yet another mechanism such as β-arrestin activation/scaffolding or a previously unreported pathway (48).

The specific signal transduction pathway that links CCK1R and TRPV1 activity remains an interesting and important question; yet, it has proven to be very difficult to determine experimentally. Cellular calcium signaling can result from many signaling transduction pathways, which presents a challenge in determining specific signal cascades involved, especially when only measuring intracellular calcium levels. Previous research has attempted to establish the pathways that enable CCK-induced calcium influx in vagal afferent neurons; however, even blocking all characterized signaling cascades does not eliminate the CCK response (4, 29). We predict future work, perhaps using techniques with pathway-level resolution, may make more progress investigating this question. Nonetheless, based on previous work and our current findings, we hypothesize CCK1R coupling to TRPV1 requires aspects of phosphoinositide signaling and likely β-arrestin; however, controlled experiments still need to directly test these possibilities.

Since the competitive capsaicin selective TRPV1 antagonists failed to prevent CCK-mediated calcium influx, we used the pore blocker ruthenium red (RuR) to ensure that TRPV1 was physically blocked and could not pass ions, independent of how the CCK1R may be coupling to the channel (Fig. 6). We report that RuR markedly reduced the CCK response in TRPV1+ and TRPV1− neurons, confirming that ∼30–40% of the CCK response is mediated by a RuR-sensitive calcium influx pathway (29). Ruthenium red is a broad-spectrum blocker and targets a variety of TRP and other voltage-gated calcium channels, such as L-type calcium channels (49, 50). The role of L-type calcium channels in the CCK response has been previously assessed using nicardipine (21). However, nicardipine also activates TRPA1 (51), which is widely expressed in vagal afferent neurons and confounds the results from these experiments (21). While it is possible that L-type calcium channels are involved in the CCK response, here we aimed to isolate the contribution of TRPV1. As there are no TRPV1 selective pore blockers, the definitive evidence we provide that causally links TRPV1 to CCK signaling includes the direct transfection of CCK1R and TRPV1 into a clonal cell line and using TRPV1 KO mice.

We transfected clonal COS-7 cells with the CCK1R alone or in combination with TRPV1 (Fig. 6). We found that the addition of TRPV1 significantly potentiated the CCK-induced calcium influx; this effect was eliminated by pretreatment with RuR. This approach served to isolate the signaling of the CCK1R to TRPV1 independent of other ion channels that may potentially couple. However, to directly test the role of TRPV1 in vagal afferent neurons, we assayed CCK signaling in vagal afferent neurons taken from TRPV1 KO mice (Figs. 7 and 8). Given the well-documented ability of TRP channels to show functional redundancy in knockout mice (52), we were surprised how effective genetic deletion of TRPV1 was at reducing the efficacy of CCK. In the cultured vagal afferents, the CCK concentration-response curve from the TRPV1 KO neurons looked nearly identical to that of the TRPV1− neurons at all concentrations except the highest (Fig. 7). At the highest CCK concentration, the response in the TRPV1 KO was markedly lower than in the TRPV1+ neurons taken from control mice. Nevertheless, the response in the TRPV1 KO was higher than in the TRPV1− neurons even though neither of them expressed TRPV1. Although the KO mice do not express TRPV1, likely 70–80% of these vagal afferent neurons retain characteristics of C-type fibers: they are small, unmyelinated, and normally express TRPV1 in wild-type animals. We and others have found that additional TRP channels often coexpress with TRPV1, including TRPA1 (29), TRP canonical type 1 (TRPC1) (29), TRP melastatin type 3 (TRPM3) (53), and others. While TRPV1 is not expressed in the knockout animals, these additional TRP channels that are abundant in C-fibers likely provide alternative calcium influx pathways in the TRPV1 KO neurons at high concentrations of CCK.

Perhaps most surprising of all was the complete inability of CCK to increase spontaneous glutamate release in brainstem slices taken from TRPV1 KO mice (Fig. 8). Meanwhile, in the vagal afferent cell bodies the response to CCK was reduced in TRPV1− neurons but still present. There are several explanations for this observation. First, it is possible that culturing the afferent neurons results in CCK1Rs and ion channel effectors (including others in addition to TRPV1) being more concentrated in the cell body than they are in the terminals. This would likely generate larger responsiveness to CCK in the cultured neurons than in the brainstem slice and may explain the differences in responsiveness. Second, a threshold of calcium influx must be reached within proximity to synaptic vesicle release mechanisms for spontaneous vesicles to fuse and be released. We suspect, that even with other TRP and ion channels available to couple, the absence of TRPV1 in the KO mice prevents CCK1R activation from producing enough calcium influx to cross threshold and increase the rate of spontaneous release onto NTS neurons. With these considerations, we posit that this finding strongly supports the exclusive role of TRPV1 to couple to the CCK1R in the central vagal terminals with the spontaneous glutamate release pathway, even while CCK1Rs couple via several effectors in the periphery (including TRPV1).

It is important to point out that in the current studies we only examined the role of TRPV1 in mediating the effects of CCK on spontaneous glutamate release. Yet, CCK has several neurophysiological effects in the NTS and is also known to modestly increase action potential evoked synchronous release (18) and may, although no one has yet explored its signaling, impact asynchronous glutamate release (32). Future studies will need to focus on the contribution of TRPV1 to other forms of CCK-mediated release.

Low-Affinity Coupling of CCK1R to TRPV1

Responses to CCK were not statistically different between TRPV1+ and TRPV1− neurons at lower concentrations of CCK (0.01–1 nM). However, at higher concentrations (10–100 nM CCK), peak responses were significantly greater in TRPV1+ neurons (Fig. 1). Previous work, mostly performed in pancreatic tissue, has characterized that the CCK1R maintains two binding affinity states (41). High-affinity CCK binding (at low concentrations, K_d ∼ 0.1 nM) results in an oscillatory calcium signal with little or no desensitization over time. In contrast, low-affinity CCK binding (at higher concentrations, K_d > 1 nM) results in a larger phasic response with a transient peak in calcium levels which decay to a steady-state level. We observed that the presence of TRPV1 increased CCK-mediated calcium influx only under high concentrations of CCK, suggesting that the CCK1R couples to TRPV1 when it is in the low-affinity state. If true, this concentration-dependent coupling could explain differences in peripheral versus central terminal CCK1R signaling, predict differences in high-concentration paracrine CCK effects (peripheral and central) compared to lower concentration CCK signaling, and may help resolve the signaling cascade paradox for the CCK1R in vagal afferents neurons.

Over a decade of work in vagal afferent neurons has pursued the identification of the signaling pathways downstream of CCK1R G protein-coupled receptor (GPCR) activation that couple the receptor to the membrane effectors (now including TRPV1) (41, 54). Yet, disruption and blocking of all characterized pathways, either alone or in combination, has failed to eliminate the CCK response. Most of the previous work targeting the CCK1R signal transduction pathway was conducted with relatively high concentrations of CCK, which we now predict couple largely via TRPV1 and in a seemingly non-canonical way (Fig. 5). A potential candidate for this low-affinity CCK1R coupling to TRPV1 is β-arrestin. β-Arrestin is best characterized for its role in GPCR internalization but is now increasingly recognized to have broader functions, including scaffolding receptor/effector complexes to facilitate signaling. Extensive work with A-kinase anchoring proteins has demonstrated intracellular coordination of protein complexes including GPCRs, TRP channels, and voltage-gated calcium channels to spatially maximize their interactions and signaling (55, 56). The potential for β-arrestin to be activated at low-affinity CCK binding to the CCK1R introduces another “inducible” component, which may facilitate the CCK1R to TRPV1 signaling. While experimentally difficult, more work is needed to pursue this possibility.

Renewed Therapeutic Potential of CCK1R with TRPV1 Coupling?

With the long history of CCK research, it is important to remember why this “satiety” hormone is not used clinically. While inducing satiety is an aspect of the CCK repertoire, it is not the only effect. Additional roles of CCK include the control of gastric emptying and pancreatic secretion. Early work aimed at the clinical use of this compound found relatively quickly that chronic activation of the CCK1R rapidly loses its effectiveness at reducing food intake, and thus treating obesity, in humans (57). Further, continued CCK use often leads to pancreatitis (58). Neither of these outcomes is compatible with clinical use. However, while CCK1Rs are expressed in many different tissues, the presence of CCK1R and TRPV1 are only colocalized to the vagal afferent neurons (as far as it has been determined so far). Thus could these unwanted side effects that arise via general targeting of CCK1R be mitigated by selectively targeting CCK1Rs in TRPV1-expressing vagal afferent neurons? By targeting treatments specifically to vagal afferent neurons or to signaling downstream of low-affinity CCK1R activation that couples to TRPV1, we could theoretically isolate the therapeutic actions of CCK to the vagal afferents and reduce unwanted side effects. Our results certainly justify more work in this direction.

Conclusions

From these findings, we have begun to generate a model for how the CCK1R couples to TRPV1 in vagal afferent neurons. We posit that TRPV1 acts as a potentiating ion channel effector that is recruited with low-affinity binding of CCK at the CCK1R. This coupling of the CCK1R to TRPV1 is present in the vagal afferent cell bodies, as are additional signaling pathways downstream of CCK1R activation. Surprisingly, at the central vagal afferent terminals in the NTS, TRPV1 is essential for mediating the CCK control of spontaneous glutamate release. Collectively this positions the TRPV1 ion channel as a key mediator of CCK signaling in vagal afferent neurons and may provide a novel point of therapeutic intervention.

DATA AVAILABILITY

Data will be made available upon reasonable request.

GRANTS

This work was supported by a grant from the National Institutes of Health (DK092651 to J.H.P.).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

R.A.A., D.K.F., and J.H.P. conceived and designed research; R.A.A., D.K.F., and J.H.P. performed experiments; R.A.A., D.K.F., and J.H.P. analyzed data; R.A.A., D.K.F., and J.H.P. interpreted results of experiments; R.A.A. and J.H.P. prepared figures; R.A.A. and J.H.P. drafted manuscript; R.A.A. and J.H.P. edited and revised manuscript; R.A.A., D.K.F., and J.H.P. approved final version of manuscript.