Capsaicin augments synaptic transmission in the rat medial preoptic nucleus

Brain Research 1043 (2005) 1 – 11 www.elsevier.com/locate/brainres

Research report

Capsaicin augments synaptic transmission in the rat medial preoptic nucleus Urban Karlssona,T, Anna K. Sundgren-Anderssona, Staffan Johanssonb, Johannes J. Kruppa a AstraZeneca R&D So¨derta¨lje, S-151 85 So¨derta¨lje, Sweden Department of Integrative Medical Biology, Section for Physiology, Umea˚ University, S-901 87 Umea˚, Sweden

b

Accepted 30 October 2004 Available online 31 March 2005

Abstract The medial preoptic nucleus (MPN) is the major nucleus of the preoptic area (POA), a hypothalamic area involved in the regulation of body-temperature. Injection of capsaicin into this area causes hypothermia in vivo. Capsaicin also causes glutamate release from hypothalamic slices. However, no data are available on the effect of capsaicin on synaptic transmission within the MPN. Here, we have studied the effect of exogenously applied capsaicin on spontaneous synaptic activity in hypothalamic slices of the rat. Whole-cell patchclamp recordings were made from visually identified neurons located in the MPN. In a subset of the studied neurons, capsaicin enhanced the frequency of spontaneous glutamatergic EPSCs. Remarkably, capsaicin also increased the frequency of GABAergic IPSCs, an effect that was sensitive to removal of extracellular calcium, but insensitive to tetrodotoxin. This suggests an action of capsaicin at presynaptic GABAergic terminals. In contrast to capsaicin, the TRPV4 agonist 4a-PDD did not affect GABAergic IPSCs. Our results show that capsaicin directly affects synaptic transmission in the MPN, likely through actions at presynaptic terminals as well as on projecting neurons. Our data add to the growing evidence that capsaicin receptors are not only expressed in primary afferent neurons, but also contribute to synaptic processing in some CNS regions. D 2004 Elsevier B.V. All rights reserved. Theme: Excitable membranes and synaptic transmission Topic: Presynaptic and postsynaptic mechanisms Keywords: Vanilloid receptor; Postsynaptic currents; Slice

1. Introduction The medial preoptic nucleus (MPN) is the major nucleus of the preoptic area (POA). This hypothalamic area is, among other functions, involved in the regulation of body temperature by controlling heat-loss and heat-retention mechanisms. POA neurons can be classified into temperature-sensitive (~30%) and temperature-insensitive neurons. The temperature-sensitive POA neurons can be further grouped into warm-sensitive and cold-sensitive. There is evidence that at least some of the temperature-sensitivity is

T Corresponding author. Fax: +46 8 553 255 81. E-mail address: [email protected] (U. Karlsson). 0006-8993/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.brainres.2004.10.064

inherent to these neurons (for review, see [3]). The most direct evidence for this is that warming causes activation of a yet unidentified, non-selective cation channel in warmsensitive neurons [14]. However, synaptic inputs from other brain regions, the periphery, as well as from local networks within the POA [28] may also contribute to the thermosensitivity of POA neurons. For example, it has been suggested that cold-sensitive neurons lack an inherent temperature sensitivity altogether, but derive their temperature sensitivity through synaptic inhibition by warmsensitive neurons [3]. The molecular basis for inherent as well as synaptically mediated temperature sensitivity of POA neurons is unknown. However, several lines of arguments support the idea that ion channels of the Transient Receptor Potential

2

U. Karlsson et al. / Brain Research 1043 (2005) 1–11

(TRP) family may be involved in this phenomenon. Morphological studies have shown the presence of such receptors in the POA. Resiniferatoxin, an analogue of the TRPV1 (formerly called VR1) agonist capsaicin, specifically binds in the POA in rat [1] and monkey [30]. In addition, TRPV1 mRNA [26] as well as protein [23] is present in several hypothalamic nuclei of the rat. Recent evidence also indicates the presence of TRPV4 protein, another member of the TRP channel family, in the medial POA of the rat hypothalamus [10]. Further, capsaicin causes hypothermia only when injected into the POA, but not into other regions of the hypothalamus [16]. Likewise, capsaicin applied by microelectrophoresis increases the firing rate of warm-sensitive units and decreases that of cold-sensitive units [14]. In hypothalamic slices, capsaicin also evokes calcium-dependent glutamate release, an effect that can be blocked by capsazepine, a competitive antagonist at TRPV1 [26]. Despite these strong indications that TRP channels may be involved in the thermosensitivity of POA neurons, only little is known about how activation of these receptors affects synaptic transmission within this region. Here, we have studied the effects of exogenously applied capsaicin on spontaneous synaptic activity in MPN neurons in hypothalamic slices of the rat. Our results show that in the MPN capsaicin increases glutamatergic as well as GABAergic synaptic transmission. The effect on both transmitter systems can be observed in the presence of TTX. To our knowledge, the present study is the first to describe in detail a direct enhancement of GABAergic synaptic transmission by the TRPV1 agonist capsaicin. This uncommon effect was not evoked when applying 4aPDD, an agonist of TRPV4, the only other TRPV channel described in the POA.

Pella, Redding, CA, USA). After cutting, slices were allowed to recover for at least 1 h in artificial cereQ brospinal fluid (aCSF, see below) at room temperature (21–23 8C). 2.2. Acute dissociation of MPN neurons The procedure for dissociation of MPN neurons followed the details previously described [19]. 2.3. Recording solutions

2. Materials and methods

The aCSF used for slice recordings contained (in mM): NaCl 124, KCl 3.0, CaCl2 2.4, MgSO4 1.3, NaH2PO4 1.4, NaOH 18, HEPES 5.0, glucose 11, pH 7.4 (95% O2, 5% CO2). Tetrodotoxin (TTX), 6-nitro-7-sulphomoylbenzo[f]quinoxaline-2,3-dione (NBQX), or bicuculline methiodide were added to the aCSF without further adjustments, as indicated in the text. In some experiments, a Ca2+-free extracellular solution of the following composition was used (in mM): NaCl 124, KCl 3.0, MgSO4 10, NaH2PO4 1.4, NaOH 18, HEPES 5.0, glucose 11, pH 7.4 (95% O2, 5% CO2). IPSCs were recorded with an intracellular solution containing (in mM): KCl 140, NaCl 3.0, MgCl2 1.2, HEPES 10, EGTA 1.0, Mg-ATP 4, pH 7.2 (KOH). The intracellular solution used to record EPSCs contained (in mM): Csgluconate 140, NaCl 3.0, MgCl2 1.2, HEPES 10, EGTA 10, Mg-ATP 4.0, pH 7.2 (CsOH). Recordings from acutely dissociated neurons were made with an extracellular solution of composition (in mM): NaCl 137, KCl 5.0, CaCl2 1.0, MgCl2 1.2, HEPES 10, glucose 10, pH 7.4 (NaOH). The intracellular solution used for these experiments was the same as the one used for IPSC recordings in slices (see above). All ATP-containing intracellular solutions were stored on ice until used in the experiments.

2.1. Slice preparation

2.4. Electrophysiological recordings

All animal experiments were approved by the regional ethics committee for animal research (S123/01). Fourteen days pregnant female Sprague–Dawley rats were purchased from Mflleg3rd (Skensved, Denmark), and housed at 20 F 0.5 8C with a 12-h light/dark cycle and free access to food and water. Offspring of both sexes were used at ages from 26 to 34 days postnatal. For recording, a rat was lightly anaesthetized with enflurane and decapitated. The brain was rapidly removed and placed throughout the entire slicing procedure in preoxygenated ice-cold (4 8C) slice preparation solution containing (in mM): Sucrose 225, KCl 5, CaCl2 2.5, MgCl2 1.5, NaH2PO4 1.2, NaHCO3 25, glucose 10, pH 7.4 (95% O2, 5% CO2). A block of tissue containing the anterior hypothalamus was used to cut 200–250 Am thick coronal slices using a vibratome (Vibratome 100 plus, Ted

The recording chamber used for slice recordings (Warner Instrument, Hamden, CT, USA) had a volume of 180 Al. It was perfused at a rate of ~1.5 ml/min with continuously bubbled (95% O2, 5% CO2) aCSF, that was preheated before entering the recording chamber using an automatic temperature controller TC-324B with inline heater SH-27B (Warner Instrument Corp., Hamden, CT, USA). The temperature was measured in the recording chamber, between the solution inlet and the slice, and was adjusted to 36 F 2 8C. For recording, a slice was fixed with a slice anchor grid with parallel threads (Warner Instrument, Hamden, CT, USA). Slices were observed using a Zeiss Axioskop (Carl Zeiss, Gfttingen, Germany) mounted with an infrared CCD camera C7500 with camera controller C2741 (Hamamatsu, Japan). Whole-cell recordings were made from visually identified MPN neurons, using the

U. Karlsson et al. / Brain Research 1043 (2005) 1–11

3

anterior commissure, optic chiasm and the third ventricle as landmarks [29]. The recording chamber used for experiments on acutely dissociated neurons had a volume of 400 Al. It was positioned on the stage of an inverted Nikon Diaphot 200 microscope (Nikon, Japan) and continuously superfused at room temperature. Recording electrodes for all experiments were pulled from borosilicate glass (GC150, Harvard Apparatus, Edenbridge, Kent, UK), and had a resistance of 3–4 MV when filled with gluconate-containing intracellular solutions. Signals from both recording set-ups were recorded using an Axopatch 200A amplifier, a Digidata 1200 interface and pClamp 8 software (all from Axon Instruments, Union City, CA). Experiments were done in whole-cell voltage-clamp mode with the holding potential clamped to 70 mV. For the gluconate-containing intracellular solutions, a liquidjunction potential of approximately 14 mV was corrected for. Signals were sampled at 2–5 kHz after low-pass filtering at 1–2 kHz ( 3 dB). The series resistance of the recordings of MPN neurons was 15.0 F 1.5 MV (n = 15) and was not compensated.

apparent event frequency prior to capsaicin application. For each cell cumulative inter-event interval histograms in control solution and during drug application were compared using the Kolmogorov–Smirnoff test, with significance set at P b 0.001. As described further below, the effects of capsaicin and 4a-PDD on synaptic events showed considerable cell-to-cell variability with clear increases in the frequency of events in some cells, while other cells showed no obvious response to capsaicin or 4a-PDD. We thus separated responding from non-responding cells by using an increase in frequency of synaptic events to N150% of control during the application of capsaicin or 4a-PDD as criterion. A similar, albeit less stringent threshold has recently been used to group capsaicin-responsive and non-responsive neurons in dorsal horn recordings [24]. Subsequent statistical analysis was done using paired Student’s t test (2-sample equal variances, two-tailed), with significance set at P b 0.05. The decay of postsynaptic currents was fitted with a monoexponential function of the form I = I max*exp( t/s) + I Leak. All data are given as mean F SEM.

2.5. Drugs and drug application

3. Results

In slice recordings, all drugs, including capsaicin and 4aphorbol 12,13-didecanoate (4a-PDD), were applied via the bath perfusion. In control experiments, the onset of glycineinduced (3.0 mM) whole-cell responses [18] was between 30 and 90 s after glycine entered the bath (not shown). The time from onset to peak was below 20 s. In experiments using acutely dissociated neurons, all compounds were applied using a DAD-12 superfusion system (ALA Scientific, Westbury, NY, USA). Capsaicin, 4a-PDD, strychnine, bicuculline methiodide, iodo-resiniferatoxin (all from Sigma) and capsazepine (Tocris) were dissolved in dimethyl sulfoxide at a concentration of 10 mM, then diluted in aCSF to the final concentration. NBQX (Tocris) was dissolved in aCSF at 10 AM. TTX (Sigma) was dissolved in water at 200 AM and diluted in aCSF to a final concentration of 1.0 AM.

3.1. General characteristics of MPN neurons and effects of capsaicin

2.6. Data analysis Postsynaptic events were analyzed off-line using MiniAnalysis (Synaptosoft, Decatur, GA, USA). Interevent intervals were grouped for each cell into a control (before application of capsaicin or 4a-PDD) and a test group. The length of the control trace and the test trace was 21 s each for EPSCs and 52 s each for IPSCs. Because the activity induced by capsaicin had a variable temporal onset and occasionally showed desensitization at higher concentrations, we defined the start of the test trace for each cell as the time-point of maximal synaptic activity observed within a 3min period after capsaicin or 4a-PDD entered the bath. Control traces were chosen from the period with the highest

All recorded neurons were located in the medial preoptic nucleus and typically had oval somata of approximately 10 by 15 Am. Usually, two or three prominent dendrites extending from the cell body were visible with translucent illumination. The studied MPN neurons had an input resistance of 604 F 64 MV and a membrane capacitance of 17.4 F 1.4 pF (n = 15). When recorded in voltage-clamp mode, bath application of capsaicin (up to 4.0 AM) to MPN neurons did not induce whole-cell currents in any of the neurons tested (n = 85, from 33 animals). As described further below, these concentrations of capsaicin affected synaptic currents recorded from the MPN neurons, indicating that an effective capsaicin concentration was reached. Because a failure of capsaicin to evoke whole-cell currents could possibly be caused by desensitization of responses due to slow perfusion in the slice, we also tested whether capsaicin induced wholecell currents in dissociated MPN neurons. This enabled a quicker change of solution (within about 15 ms) and thus quicker application of capsaicin. However, similar to the results for MPN neurons in situ, capsaicin (2.0–10 AM) did not induce whole-cell currents in any of the dissociated MPN neurons tested (n = 4 from 2 animals) (data not shown). At a temperature of 36 F 2 8C, spontaneous synaptic currents were observed in all MPN neurons recorded in the slice preparation (n = 105; Figs. 1A, B). At 70 mV, they had amplitudes of 6 to 210 pA and a mean frequency of

4

U. Karlsson et al. / Brain Research 1043 (2005) 1–11

Fig. 1. Characteristics of spontaneous synaptic activity in MPN neurons. (A) Spontaneous synaptic activity was completely and reversibly abolished by coapplication of 10 AM NBQX and 20 AM bicuculline. (B) In a subset of MPN neurons, application of capsaicin, but not control vehicle, augmented the frequency of postsynaptic currents.

15.1 F 1.3 Hz (770 events from n = 4 neurons). Bath application of the GABAA receptor antagonist bicuculline (20 AM) together with the AMPA-receptor antagonist NBQX (10 AM) completely and reversibly blocked all events in the five neurons tested (Fig. 1A). Whereas application of vehicle control had no effect on spontaneous synaptic events in all neurons tested, application of capsaicin (at concentrations N1.0 AM) increased the frequency of synaptic events in a subset of neurons (Fig. 1B), but had no detectable effect in other neurons. Consequently, there was a large cell-to-cell variability across all recorded neurons when comparing the frequencies of synaptic events upon capsaicin application with their respective pre-capsaicin controls. For example, in the five neurons tested in aCSF without bicuculline or NBQX, the frequencies of synaptic events upon capsaicin application ranged from 87 to 5942% of control. A similar large cell-to-cell variability in response to capsaicin was observed when either NBQX or bicuculline was added to the bath. This variability is consistent with previous findings that MPN neurons are heterogeneous [3] and that different subsets of MPN neurons receive input from different neurotransmitter systems [27]. Similar to the methods described by Pan and Pan [24], we separated responding from non-responding neurons in all following experiments along a threshold of N150% event frequency during capsaicin as compared to control (see Materials and methods). 3.2. Capsaicin enhances excitatory synaptic transmission Glutamatergic EPSCs were studied after addition of 20 AM bicuculline to the bath. At 36 F 2 8C spontaneous EPSCs (sEPSCs) had a 10–90% rise time of 0.5 F 0.1 ms. The decay was well fitted by a mono-exponential function with a time constant, s EPSC, of 2.4 F 0.3 ms. Spontaneous EPSCs were small, 11 F 4 pA at 70 mV, and occurred

with a frequency of 6.9 F 2.9 Hz (567 events from n = 8 neurons). As shown in Fig. 2, bath application of 4.0 AM capsaicin increased the frequency of sEPSCs in 5 of 8 neurons from 4.7 F 1.3 Hz to 12.3 F 2.6 Hz (paired t test; P b 0.05), but did not affect their amplitude ( 12 F 2 pA before and 11 F 1 pA during capsaicin). In four of the five responsive neurons, the sEPSC inter-event interval distribution was significantly shifted to shorter intervals in the presence of capsaicin as compared to control (Kolmogorov–Smirnoff test, P b 0.001; Fig. 2), whereas none of the three nonresponsive neurons showed such a shift. At a capsaicin concentration of 1 AM, the sEPSC frequency increased in two of five neurons tested, but 500 nM capsaicin did not affect the sEPSC frequency (n = 3). Consistent with the data obtained with application of 4 AM capsaicin, there was no effect on sEPSC amplitude for any of the groups. To test whether the effect of capsaicin was action potential dependent, we studied miniature EPSCs (mEPSCs) in the presence of 1.0 AM TTX and 20 AM bicuculline. Miniature EPSCs were similar in amplitude ( 10 F 1 pA) and kinetics (10–90% rise time: 0.5 F 0.1 ms; s mEPSC = 2.4 F 0.2 ms; 150 events from 7 neurons) compared with sEPSCs. They occurred with a frequency of 1.8 F 0.5 Hz. Bath application of 4.0 AM capsaicin increased the mEPSC frequency in 2 of 7 neurons above the 150% threshold from 1.1 to 2.3 Hz and from 1.2 to 2.0 Hz, respectively. However, the distribution of the inter-event intervals was not significantly altered in any of the neurons. This is probably due to the fact that in both neurons capsaicin responses desensitized quickly, allowing only for a small number of events to be recorded. As for the sEPSCs, capsaicin did not affect the amplitude or kinetics of mEPSCs. Thus, capsaicin increases the frequency of EPSCs in a subpopulation of MPN neurons in an apparent concentration-dependent manner. This effect is, at least partially, insensitive to TTX.

U. Karlsson et al. / Brain Research 1043 (2005) 1–11

5

Fig. 2. Capsaicin augments the sEPSC frequency in the presence of 20 AM bicuculline in a subset of MPN neurons. (A) Example of a neuron in which bath perfusion of 4.0 AM capsaicin induced an increase in the sEPSC frequency. The continuous frequency (middle) and running average amplitude (bottom) histograms for the recording are aligned to the original recording shown on top (0.5 s bin size for both histograms). Note that the delayed, but sharp increase in the frequency of synaptic events upon application of capsaicin is not matched by a similar increase in the amplitude of events (same time-axis for original recording and both histograms). (B) Expanded records from the recording shown in A at the time-points indicated by the arrows and numbers. (C) Cumulative probability plot of the inter-event intervals for control (thin line) and during capsaicin perfusion (bold line). (D) Cumulative probability plot of the sEPSC amplitudes for control (thin line) and during capsaicin perfusion (bold line). All data are from the same neuron.

3.3. Capsaicin also enhances inhibitory synaptic transmission Spontaneous GABAergic IPSCs were studied after addition of 10 AM NBQX to the bath. The pharmacologically isolated IPSCs had a 10–90% rise time of 0.6 F 0.1 ms, and their decay was well fitted by a monoexponential function with a time constant, s IPSC, of 4.6 F 0.4 ms. Their amplitude was 103 F 11 pA at 70 mV and the frequency was 4.3 F 1.2 Hz (1493 events from n = 16 neurons). As shown in Fig. 3, bath application of 4.0 AM capsaicin increased the frequency of spontaneous IPSCs (sIPSCs) in

10 of 16 neurons from 2.2 F 0.8 to 12.3 F 3.3 Hz (paired t test; P b 0.05). The cumulative sIPSC inter-event interval distribution was significantly shifted to shorter intervals in the presence of capsaicin, in all responsive neurons (Kolmogorov–Smirnoff test; P b 0.001), but in none of the six non-responsive neurons. Although capsaicin slightly increased the amplitude of sIPSCs in the 10 responsive neurons ( 96 F 10 pA before versus 140 F 20 pA during capsaicin) this effect was not significant. Capsaicin did not significantly affect the kinetics of sIPSCs (10–90% rise time: 0.6 F 0.1 ms; s IPSC = 5.5 F 0.5 ms, 1351 events from 10 neurons).

6

U. Karlsson et al. / Brain Research 1043 (2005) 1–11

Fig. 3. Capsaicin augments the sIPSC frequency in the presence of 10 AM NBQX in a subset of MPN neurons. (A) Example of a neuron in which bath perfusion of 4.0 AM capsaicin induced an increase in the sIPSC frequency. The continuous frequency (middle) and running average amplitude (bottom) histograms for the recording are aligned to the original recording shown on top (1 s bin size for both histograms). Note that the delayed, but sharp increase in the frequency of synaptic events upon application of capsaicin is not matched by a similar increase in the amplitude of events (same time-axis for original recording and both histograms). (B) Expanded records from the recording shown in A at the time-points indicated by the arrows and numbers. (C) Cumulative probability plot of the inter-event times for control (thin line) and during capsaicin perfusion (bold line). (D) Cumulative probability plot of the sIPSC amplitudes for control (thin line) and during capsaicin perfusion (bold line). All data are from the same neuron.

The capsaicin-induced increase in sIPSC frequency was concentration-dependent. Thus, 2.0 AM capsaicin increased the sIPSC frequency in 6 of 12 neurons above the threshold level of 150% of the pre-capsaicin frequency. The inter-

event interval distribution for capsaicin was indeed significantly shifted to shorter intervals in all six neurons, but in none of the non-responsive neurons. However, in contrast to the results obtained with 4 AM capsaicin, the pooled

U. Karlsson et al. / Brain Research 1043 (2005) 1–11

increase in sIPSC frequency in the six responsive neurons from 4.1 F 2.0 to 8.4 F 3.8 Hz was not statistically significant. Also indicative of a concentration-dependent effect by capsaicin was the finding that none of the seven neurons tested at 1.0 AM capsaicin could be classified as responder. In two neurons, however, capsaicin caused a slight increase in sIPSC frequency below the 150% threshold (127% and 132%, respectively), accompanied by a change in the pattern of synaptic activity, which switched from regular to burst-like. This caused a significant shift in the inter-event histogram to shorter intervals in both neurons. Capsaicin did not significantly affect the sIPSC amplitude in any of the groups. To test for specificity we used two structurally nonrelated TRPV1-antagonist, capsazepine and iodo-resiniferatoxin. Bath application of capsazepine (40 AM) reduced the frequency of sIPSCs from 5.1 F 1.4 Hz to 2.6 F 0.6 Hz (n = 9). Upon subsequent application of capsaicin (4.0 AM), only 1 of 9 neurons showed an increase in the frequency of sIPSCs above the 150% threshold (from 1.5 to 5.9 Hz). Similar results were obtained for iodo-resiniferatoxin. Bath application of iodo-resiniferatoxin (2 AM) reduced the frequency of sIPSCs from 7.8 F 3.8 to 4.0 F 1.4 Hz (n = 6). In none of the six cells subsequent application of capsaicin (4.0 AM) increased the frequency of events above the 150% threshold. Thus, as for the EPSCs, capsaicin increases, the frequency of sIPSCs in a subpopulation of MPN neurons in a concentration-dependent manner. Because the enhancement of inhibitory synaptic transmission by capsaicin is an unusual finding, the rest of the study is focused on this phenomenon.

7

3.4. The TRPV4 agonist 4a-PDD has no clear effect on inhibitory synaptic transmission A recent study showed that, in addition to TRPV1, TRPV4 is present in the preoptic region [10]. To test whether the TRPV4 protein in this region results in functional TRPV4 receptors, we applied the TRPV4 agonist 4a-PDD [34] to cells in which sIPSCs had been isolated pharmacologically by 10 AM NBQX. At a concentration of 1.0 AM – about 6 times higher than the EC50 value for TRPV4 [34] – 4a-PDD did neither affect sIPSC frequency, amplitude, inter-event interval distribution, nor the whole-cell holding current in any of the MPN neurons tested (Fig. 4; n = 5). Because TRPV4 receptors may already be active at bath temperatures of 36 F 2 8C [35], we also tested 1 AM 4a-PDD at room temperature (22 F 1 8C). Still, in none of the neurons tested (n = 6) were sIPSCs or the whole-cell holding current affected. Likewise, there was no effect on the whole-cell holding current in any of the 9 neurons tested at 36 F 2 8C with an even higher concentration of 4a-PDD (4.0 AM). In one of these 9 neurons, 4.0 AM 4a-PDD increased the frequency of sIPSCs above the 150% threshold from 0.7 to 3.5 Hz, and also significantly shifted the inter-event interval distribution to shorter intervals. 3.5. Capsaicin increases inhibitory synaptic transmission through a TTX-insensitive mechanism To clarify whether the effects of capsaicin on IPSCs were dependent on action potentials, we added 1.0 AM TTX (+10 AM NBQX) to the bath. Miniature IPSCs (mIPSCs) had

Fig. 4. The TRPV4 agonist 4a-PDD does not affect sIPSCs in MPN neurons. (A) Example of a neuron in which bath perfusion of 1.0 AM 4a-PDD in the presence of 10 AM NBQX did not affect sIPSCs. (B) Cumulative probability plot of the inter-event times for control (thin line) and during 4a-PDD perfusion (bold line). (C) Cumulative probability plot of the sIPSC amplitudes for control (thin line) and during 4a-PDD perfusion (bold line). All data are from the same neuron.

8

U. Karlsson et al. / Brain Research 1043 (2005) 1–11

kinetic characteristics (10–90% rise time 0.7 F 0.1 ms, s mIPSC = 6.6 F 0.7 ms; 594 events from 13 neurons) that were similar to those of sIPSCs. However, mIPSCs were smaller ( 69 F 9 pA) and less frequent (1.2 F 0.3 Hz; n = 13) than sIPSCs. Bath application of 4.0 AM capsaicin increased the mIPSC frequency in 6 of 13 neurons from 0.9 F 0.2 to 13.7 F 5.7 Hz (Fig. 5A). The large variation in the presence of capsaicin was due to two neurons, in which capsaicin caused a massive increase in mIPSC frequency (3096% and 7689% of control, respectively). Due to this large variation, the pooled increase in mIPSC frequency was statistically not significant. Capsaicin significantly shifted the inter-event interval distribution to shorter intervals in four of the six responsive neurons, but in none of the non-responsive neurons. In the responsive neurons,

the mean mIPSC amplitude was 71 F 12 pA before and 66 F 8 pA during application of capsaicin (paired t test, P =0.72). The data indicate that capsaicin may enhance inhibitory synaptic transmission in a subset of MPN neurons through a mechanism independent of presynaptic impulse generation. To further clarify the mechanism of capsaicin action, we tested whether the effect of capsaicin was calciumdependent. As expected, sIPSCs occurred at a lower frequency in Ca2+-free extracellular solution (0.6 F 0.1 Hz, 147 events from 5 neurons) compared to the standard recordings with 2.4 mM extracellular Ca2+. Capsaicin (4.0 AM) did neither increase the sIPSC frequency in nominally Ca2+-free conditions, nor was the inter-event interval distribution significantly changed in any of the 5 neurons tested (Fig. 5B).

Fig. 5. The effect of capsaicin on the sIPSC frequency is not blocked by TTX, but is dependent on extracellular Ca2+. (A) Effect of capsaicin on mIPSCs. (a) In the presence of 1.0 AM TTX, bath perfusion of capsaicin induced an increase in the mIPSC frequency. (b) Expanded records from the recording shown in a at the time-points indicated by the arrows and numbers. (c) Cumulative probability plot of the inter-event times for control (thin line) and during capsaicin perfusion (bold line). (d) Cumulative probability plot of the mIPSC amplitudes for control (thin line) and during capsaicin perfusion (bold line). All data are from the same neuron. (B) In Ca2+-free extracellular solution, bath perfusion of 4.0 AM capsaicin was without effect on sIPSCs.

U. Karlsson et al. / Brain Research 1043 (2005) 1–11

5. Discussion In the present work, we studied the effects of capsaicin and 4a-PDD on synaptic transmission to MPN neurons in hypothalamic slices. In agreement with previous studies in this preparation, our data show that MPN neurons receive glutamatergic as well as GABAergic synaptic input [13]. Our results indicate that capsaicin can enhance both transmitter systems in the MPN. Because capsaicin effects were observed in the presence of TTX, they are not likely to depend on action potential generation. The capsaicin effect was, however, Ca2+ dependent, suggesting that the effect involves a direct modulation of transmitter release from presynaptic terminals. The results are consistent with the idea that in the MPN capsaicin directly affects GABA release from presynaptic terminals by increasing the Ca2+ concentration in the terminals. In contrast to capsaicin, 4aPDD neither affected GABAergic IPSCs nor did it induce postsynaptic currents in the recorded MPN neurons. We conclude that capsaicin acts on receptors located at presynaptic terminals in the MPN and possibly also on neurons projecting to the MPN from nearby regions in the slice preparation (see below). 5.1. Comparison with earlier results Although capsaicin and the classical capsaicin receptor, TRPV1 (formerly called VR1), are typically discussed in relation to noxious stimuli, there is also a broad body of evidence that capsaicin can directly modulate synaptic transmission. At synapses in substantia gelatinosa of the spinal cord, capsaicin facilitates excitatory, but not inhibitory, synaptic transmission [24,37,38]. Likewise, capsaicin facilitates excitatory synaptic transmission by a presynaptic action at primary afferent synapses in various brainstem nuclei [8,17]. Capsaicin also affects synaptic transmission at synapses that are not formed by primary afferents. For example, through presynaptic actions, capsaicin increases the frequency of mEPSCs in locus coeruleus neurons [21] and in dopaminergic neurons of the substantia nigra [22]. Finally, capsaicin also affects synaptic transmission in the hippocampus [2]. Our results add to these data by directly demonstrating a similar increase in synaptic transmission onto MPN neurons, in central synapses for which such an effect was expected on basis of previous, more indirect studies [16,26]. Similar to the situation at most studied synapses, capsaicin increased glutamatergic excitatory transmission onto MPN neurons. In some neurons, this effect was not sensitive to blockade of action potential generation, a finding that fits to an immediate presynaptic site of action, as reported for other glutamatergic synapses [21,22]. However, there was a marked difference between the percentages of capsaicin-responsive neurons in the presence vs. absence of TTX, suggesting that the observed enhancement of sEPSC frequency may in some cases not

9

have been caused by an immediate presynaptic action. Thus, capsaicin may have acted presynaptically on another neuron contained within the hypothalamic slice, with the latter neuron projecting to the studied MPN neuron. This scenario is conceivable because MPN neurons receive synaptic inputs from all major nuclei of the hypothalamus [28]. Some of these nuclei, like the anteroventral periventricular nucleus and the bed nucleus of the stria terminalis are present in the coronal slice preparation used in this study. We show that in the MPN, capsaicin also increases the frequency of GABAergic IPSCs. The enhancement of inhibitory synaptic transmission by capsaicin is an uncommon finding. At most synapses studied, capsaicin affects excitatory transmission only [22,37]. One notable exception from this is the hippocampus, where capsaicin increases paired-pulse depression of population spikes in the CA1 region, apparently through an increase in GABAergic synaptic transmission [2]. To our knowledge, the present study is the first to directly study a capsaicin-mediated modulation of GABAergic transmission at a synapse. We find that capsaicin only affected the frequency of IPSCs, but was without significant effect on the amplitude of sIPSCs or mIPSCs. Together with the fact that the effect of capsaicin was TTX-insensitive and dependent on extracellular Ca2+, our results strongly argue for a direct presynaptic modulation by capsaicin of GABAergic transmission onto MPN neurons. In the present study, high concentrations (more than 1.0 AM) of capsaicin were required for effects. Such high concentrations may cause unspecific effects [31] that in principle could explain the findings described in our study. However, the increase in sIPSC or sEPSC frequency was not observed in every MPN neuron, which makes this explanation unlikely. Furthermore, the effect of capsaicin was attenuated by two structurally non-related TRPV1antagonists, capsazepine and iodo-resiniferatoxin. In addition, we found that the capsaicin-induced increase in the frequency of spontaneous synaptic activity was concentration dependent. These findings are not reconcilable with an unspecific effect, and indicate that capsaicin acted through activation of a defined molecular identity. Of note, in a previous study in the hypothalamus [26], 3.0 AM capsaicin was used to evoke glutamate release. Micromolar concentrations of capsaicin have also been used to study effects in dopaminergic neurons of the substantia nigra [22]. The permeability of TRP receptors typically shows a high Ca2+/Na+ ratio [5,36] and a presynaptic localization of the postulated capsaicin receptor would favor a direct effect on synaptic release by the inflowing Ca2+. However, Ca2+dependent transmitter release onto MPN neurons has been shown to be a complex process [9] and an indirect mechanism where activation of the capsaicin receptor would depolarize the presynaptic terminal, and thereby induce Ca2+ influx through voltage-dependent Ca2+ channels, is also possible. A similar mechanism has recently been

10

U. Karlsson et al. / Brain Research 1043 (2005) 1–11

described in another study on MPN neurons, where presynaptic GABAA receptors were shown to influence spontaneous GABA release via depolarization of the presynaptic membrane [11]. We also tested for functional TRPV4 receptors in the MPN because this protein is expressed in the hypothalamus [10]. The presence of TRPV4 in this region is of interest because recent studies have shown that TRPV4, like TRPV1, can function as a multimodal integrator of diverse stimuli, including temperature and osmolarity [10,35]. However, even when using a high concentration of 4aPDD (4.0 AM), we observed synaptic effects in only one of 9 cells tested. Micromolar concentrations of 4a-PDD affect numerous receptors, including Ca2+ channels [12,25], and the observed increase in sIPSC frequency in this one cell may be unspecific. We thus did not find conclusive evidence for the presence of functional homomeric TRPV4 receptors. One possible explanation for this finding may be the existence of heteromeric receptors with altered pharmacology. However, other scenarios are also possible. 5.2. Physiological considerations We did not find evidence for the presence of postsynaptic capsaicin-activated receptors in the MPN. This is similar to findings by Marinelli et al. [21] who reported that locus coeruleus neurons also do not respond directly to capsaicin, although their glutamatergic inputs are modulated by capsaicin. In the hypothalamus, this implies that the capsaicin receptor is unlikely to account for the intrinsic thermo-sensitivity of warm-sensitive neurons [3]. Thus, our results raise the question as to the molecular identity of the receptor causing their intrinsic thermo-sensitivity. One candidate for this role is the TREK-1 potassium channel. This channel is expressed in the hypothalamus and is opened by heat [20]. However, there are caveats with this scenario: Opening of a K+ channel is likely to hyperpolarize the neuron, making it less likely to fire, and thereby transform the neuron into a cold-sensing unit [20]. Further, Hori et al. [15] showed that warming activated, with a threshold of ~36 8C, a non-selective cation channel in warm-sensitive neurons in the POA. The induced whole-cell current was large enough to affect the firing-rate of the neurons. Clearly, further studies are required to pinpoint the receptor responsible for the intrinsic thermo-sensitivity of warm-sensitive neurons. Our results do not rule out important effects for TRP channels on the overall thermo-sensitivity within the POA. For example, most of the cold-sensitive neurons within the POA lose their thermo-sensitivity upon blockade of synaptic transmission, suggesting that cold-sensitive neurons are inhibited by nearby warm-sensitive neurons [3]. If warmsensitive neurons, besides their intrinsic thermo-sensitivity, would be endowed with capsaicin- and thermo-sensitive TRPV channels on their GABAergic terminals, the overall inhibitory synaptic drive from warm-sensitive neurons onto

cold-sensitive neurons would underlie a dual regulation by temperature-sensitive mechanisms. An additional consideration is that TRP channels act as integrators for multiple signals. For example, homomeric TRPV1 channels integrate noxious heat, acidification, as well as multiple intracellular signaling cascades [6,32]. Bradykinin has been shown to lower the temperature threshold for activation of TRPV1 channels by releasing the channel from phosphatidylinositol-4,5-biphosphatemediated inhibition [6]. Because there is good evidence for the involvement of bradykinin as a causal factor in the generation of fever [7,33], it seems plausible that during fever presynaptic TRPV1 receptors in the POA have a lowered temperature threshold and are involved in synaptic signaling during hyperthermia. If one assumes that TRPV channels are involved in thermo-sensation in the POA, our findings suggest that a significant degree of the temperature sensitivity within this area is generated at the presynaptic terminal. In a model proposed by Jancso-Gabor et al. [16], hypothalamic capsaicin receptors act as detectors of core body warmth. Hypothermia, induced by injection of capsaicin into the POA, is in this model explained by receptor activation beyond the level set by core body temperature. This would lead to activation of compensatory mechanisms, resulting in hypothermia. Our results indicate that this action of capsaicin may be through increased transmitter release, adding to the growing evidence that in various CNS regions capsaicin receptors contribute to synaptic processing [4].

Acknowledgments This work was supported by the Swedish Research Council (Project No. 11202) and Ume3 University. We thank Dr. Mark Connor for helpful discussions.

References [1] G. Acs, M. Palkovits, P.M. Blumberg, Specific binding of [3H]resiniferatoxin by human and rat preoptic area, locus coeruleus, medial hypothalamus, reticular formation and ventral thalamus membrane preparations, Life Sci. 59 (1996) 1899 – 1908. [2] A. Al-Hayani, K.N. Wease, R.A. Ross, R.G. Pertwee, S.N. Davies, The endogenous cannabinoid anandamide activates vanilloid receptors in the rat hippocampal slice, Neuropharmacology 41 (2001) 1000 – 1005. [3] J.A. Boulant, Hypothalamic neurons. Mechanisms of sensitivity to temperature, Ann. N. Y. Acad. Sci. 856 (1998) 108 – 115. [4] M.J. Caterina, Vanilloid receptors take a TRP beyond the sensory afferent, Pain 105 (2003) 5 – 9. [5] M.J. Caterina, M.A. Schumacher, M. Tominaga, T.A. Rosen, J.D. Levine, D. Julius, The capsaicin receptor: a heat-activated ion channel in the pain pathway, Nature 389 (1997) 816 – 824. [6] H.-H. Chuang, E.D. Prescott, H. Kong, S. Shields, S.E. Jordt, A.I. Basbaum, M.V. Chao, D. Julius, Bradykinin and nerve growth factor release the capsaicin receptor from PtdINs(4,5)P2-mediated inhibition, Nature 411 (2001) 957 – 962.

U. Karlsson et al. / Brain Research 1043 (2005) 1–11 [7] M.M. Coelho, C.R. Oliveira, G.P. Pajolla, J.B. Calixto, I.R. Pela´, Central involvement of kinin B1 and B2 receptors in the febrile response induced by endotoxin in rats, Br. J. Pharmacol. 121 (1997) 296 – 302. [8] M.W. Doyle, T.W. Bailey, Y.H. Jin, M.C. Andresen, Vanilloid receptors presynaptically modulate cranial visceral afferent synaptic transmission in nucleus tractus solitarius, J. Neurosci. 22 (2002) 8222 – 8229. [9] M. Druzin, D. Haage, E. Malinina, S. Johansson, Dual and opposing roles of presynaptic Ca2+ influx for spontaneous GABA release from rat medial preoptic nerve terminals, J. Physiol. 542 (2002) 131 – 146. [10] A.D. Guler, H. Lee, T. Iida, I. Shimizu, M. Tominaga, M. Caterina, Heat-evoked activation of the ion channel, TRPV4, J. Neurosci. 22 (2002) 6408 – 6414. [11] D. Haage, M. Druzin, S. Johansson, Allopregnanolone modulates spontaneous GABA release via presynaptic Cl permeability in rat preoptic nerve terminals, Brain Res. 958 (2002) 405 – 413. [12] P. Hockberger, M. Toselli, D. Swandulla, H.D. Lux, A diacylglycerol analogue reduces neuronal calcium currents independently of protein kinase C activation, Nature 338 (1989) 340 – 342. [13] N.W. Hoffman, J.P. Wuarin, F.E. Dudek, Whole-cell recordings of spontaneous synaptic currents in medial preoptic neurons from rat hypothalamic slices: mediation by amino acid neurotransmitters, Brain Res. 660 (1994) 349 – 352. [14] T. Hori, M. Shibata, T. Kiyohara, T. Nakashima, A. Asamai, Responses of anterior hypothalamic-preoptic thermosensitive neurons to locally applied capsaicin, Neuropharmacology 27 (1988) 135 – 142. [15] A. Hori, K. Minato, S. Kobayashi, Warming-activated channels of warm-sensitive neurons in rat hypothalamic slices, Neurosci. Lett. 275 (1999) 93 – 96. [16] A. Jancso-Gabor, J. Szolcsanyi, N. Jancso, Stimulation and desensitization of the hypothalamic heat-sensitive structures by capsaicin in rats, J. Physiol. 208 (1970) 449 – 459. [17] E.A. Jennings, C.W. Vaughan, L.A. Roberts, M.J. Christie, The actions of anandamide on rat superficial medullary dorsal horn neurons in vitro, J. Physiol. 548 (2003) 121 – 129. [18] U. Karlsson, D. Haage, S. Johansson, Currents evoked by GABA and glycine in acutely dissociated neurons from the rat medial preoptic nucleus, Brain Res. 770 (1997) 256 – 260. [19] U. Karlsson, A.K. Sundgren, J. Nasstrom, S. Johansson, Glutamateevoked currents in acutely dissociated neurons from the rat medial preoptic nucleus, Brain Res. 759 (1997) 270 – 276. [20] F. Maingret, I. Lauritzen, A.J. Patel, C. Heurteaux, R. Reyes, F. Lesage, M. Lazdunski, E. Honore, TREK-1 is a heat-activated background K+ channel, EMBO J. 19 (2000) 2483 – 2491. [21] S. Marinelli, C.W. Vaughan, M.J. Christie, M. Conno, Capsaicin activation of glutamatergic synaptic transmission in the rat locus coeruleus in vitro, J. Physiol. 543 (2002) 531 – 540. [22] S. Marinelli, V. Di Marzo, N. Berretta, I. Matias, M. Maccarrone, G. Bernardi, N.B. Mercuri, Presynaptic facilitation of glutamatergic synapses to dopaminergic neurons of the rat substantia nigra by endogenous stimulation of vanilloid receptors, J. Neurosci. 23 (2003) 3136 – 3144. [23] E. Mezey, Z.E. Toth, D.N. Cortright, M.K. Arzubi, J.E. Krause, R. Elde, A. Guo, P.M. Blumberg, A. Szallasi, Distribution of mRNA for

[24]

[25]

[26]

[27]

[28]

[29] [30]

[31] [32]

[33]

[34]

[35]

[36]

[37]

[38]

11

vanilloid receptor subtype 1 (VR1), and VR1-like immunoreactivity, in the central nervous system of the rat and human, Proc. Natl. Acad. Sci. U. S. A. 97 (2000) 3655 – 3660. Y.Z. Pan, H.L. Pan, primary afferent stimulation differentially potentiates excitatory and inhibitory inputs to spinal lamina II outer and inner neurons, J. Neurophysiol. 91 (2004) 2413 – 2421. H.L. Reeve, P.F. Vaughan, C. Peers, Enhancement of Ca2+ channel currents in human neuroblastoma (SH-SY5Y) cells by phorbol esters with and without activation of protein kinase C, Pflugers Arch.- Eur. J. Phys. 429 (1995) 729 – 737. T. Sasamura, M. Sasaki, C. Tohda, Y. Kuraishi, Existence of capsaicin-sensitive glutamatergic terminals in rat hypothalamus, NeuroReport 9 (1998) 2045 – 2048. R.B. Simerly, R.A. Gorski, L.W. Swanson, Neurotransmitter specificity of cells and fibers in the medial preoptic nucleus: an immunohistochemical study in the rat, J. Comp. Neurol. 246 (1986) 343 – 363. R.B. Simerly, L.W. Swanson, The organization of neural inputs to the medial preoptic nucleus of the rat, J. Comp. Neurol. 246 (1986) 312 – 342. L.W. Swanson, Brain Maps: Structure of the Rat Brain, Elsevier, Amsterdam; 1992. T. Szabo, T. Biro, A.F. Gonzalez, M. Palkovits, P.M. Blumberg, Pharmacological characterization of vanilloid receptor located in the brain, Mol. Brain Res. 98 (2002) 51 – 57. A. Szallasi, P.M. Blumberg, Vanilloid (Capsaicin) receptors and mechanisms, Pharmacol. Rev. 51 (1999) 159 – 211. M. Tominaga, M.J. Caterina, A.B. Malmberg, T.A. Rosen, H. Gilbert, K. Skinner, B.E. Raumann, A.I. Basbaum, D. Julius, The cloned capsaicin receptor integrates multiple pain-producing stimuli, Neuron 21 (1998) 531 – 543. K. Walker, A. Dray, M. Perkins, Development of hyperthermia and hyperalgesia following intracerebroventricular administration of endotoxin in the rat: effect of kinin B1 and B2 receptor antagonist, Immunopharmacology 33 (1996) 264 – 269. H. Watanabe, J.B. Davis, D. Smart, J.C. Jerman, G.D. Smith, P. Hayes, J. Vriens, W. Cairns, U. Wissenbach, J. Prenen, V. Flockerzi, G. Droogmans, C.D. Benham, B. Nilius, Activation of TRPV4 channels (hVRL-2/mTRP12) by phorbol derivatives, J. Biol. Chem. 277 (2002) 13569 – 13577. H. Watanabe, J. Vriens, S.H. Suh, C.D. Benham, G. Droogmans, B. Nilius, Heat-evoked activation of TRPV4 channels in a HEK293 cell expression system and in native mouse aorta endothelial cells, J. Biol. Chem. 277 (2002) 47044 – 47051. J.M. Welch, S.A. Simon, P.H. Reinhart, The activation mechanism of rat vanilloid receptor 1 by capsaicin involves the pore domain and differs from the activation by either acid or heat, Proc. Natl. Acad. Sci. U. S. A. 97 (2000) 13889 – 13894. K. Yang, E. Kumamoto, H. Furue, M. Yoshimura, Capsaicin facilitates excitatory but not inhibitory synaptic transmission in substantia gelatinosa of the rat spinal cord, Neurosci. Lett. 255 (1998) 135 – 138. K. Yang, E. Kumamoto, H. Furue, Y.Q. Li, M. Yoshimura, Capsaicin induces a slow inward current which is not mediated by substance P in substantia gelatinosa neurons of the rat spinal cord, Neuropharmacology 39 (2000) 2185 – 2194.