[1] 22w Cholinergic and glutamatergic transmission at synapses between pedunculopotine tegmental nucleus axonal terminals and A7 catecholamine cell group noradrenergic neurons in the rat

[1] 219w We characterized transmission from the pedunculopotine tegmental nucleus (PPTg), which contains cholinergic and glutamatergic neurons, at synapses with noradrenergic (NAergic) A7 neurons. Injection of an anterograde neuronal tracer, biotinylated-dextran amine, into the PPTg resulted in labeling of axonal terminals making synaptic connection with NAergic A7 neurons. Consistent with this, extracellular stimulation using a train of 10 pulses at 100 Hz evoked both fast and slow excitatory synaptic currents (EPSCs) that were blocked, respectively, by DNQX, a non-N-methyl-D-aspartate receptor blocker, or atropine, a cholinergic muscarinic receptor (mAChR) blocker. Interestingly, many spontaneous-like, but stimulation-dependent, EPSCs, were seen for up to one second after the end of stimulation and were blocked by DNQX and decreased by EGTA-AM, a membrane permeable form of EGTA, showing they are glutamatergic EPSCs causing by asynchronous release of vesicular quanta. Moreover, application of atropine or carbachol, an mAChR agonist, caused, respectively, an increase in the number of asynchronous EPSCs or a decrease in the frequency of miniature EPSCs, showing that mAChRs mediated presynaptic inhibition of glutamatergic transmission of the PPTg onto NAergic A7 neurons. In conclusion, our data show direct synaptic transmission of PPTg afferents onto pontine NAergic neurons that involves cooperation of cholinergic and glutamatergic transmission. This dual-transmitter transmission drives the firing rate of NAergic neurons, which may correlate with axonal and somatic/dendritic release of NA.

[1] 240w Acetylcholine (ACh) and noradrenaline (NA) are two important neurotransmitters in the peripheral and central nervous systems (CNS). These two transmitter systems share some structural features in that their axon terminals are widely distributed in the brain and arise from a small population of neurons located in the basal forebrain and pons. Cholinergic fibers mainly originate from two distinct sources, the basal forebrain and mesopontine area (Butcher and Woolf, 2003). The basal forebrain contains complex cholinergic neurons including the medial septal nucleus, the ventral and horizontal diagonal bands nuclei, and the basal nucleus. These cholinergic neurons project to the entire cerebral cortex, hippocampus, and amygdala. The mesopontine area contains cholinergic neurons located in the laterodorsal tegmental nucleus and the pedunculopotine tegmental (PPTg) nucleus. They provide intense innervation to the brainstem, spinal cord, thalamus, hypothalamus, and medial limbic cortex (Jones, 1993), and this cholinergic system is known for its ability to activate the cerebral cortex and is therefore also referred to as the ascending reticular activating system (Jones, 1993;Inglis et al., 2001). As regards the noradrenergic (NAergic) system, the A5, A6 (also known as the locus coeruleus, LC), and A7 catecholamine cell groups in the dorsal pontine area are the major suppliers of NAergic projections. The A6 cell group (LC) projects to the forebrain, the brainstem and spinal cord; while A5, A6 and A7 cell groups mainly projects to the brainstem and spinal cord (Szabadi, 2013;Howorth et al., 2009;Bruinstroop et al., 2012).

[2] 289w Cholinergic and NAergic neurons have widespread projections to the entire CNS and, accordingly, many brain functions, such as arousal and attention, learning and memory, the sleep-wakefulness cycle, and pain (Castro-Alamancos and Gulati, 2014;Favero et al., 2012;Kimura et al., 2012;Rowley et al., 2005), are simultaneously regulated by these two transmitter systems. Given the overlap of their functional roles, it is expected that coordination and interaction will exist between the cholinergic and NAergic systems. Taking the regulation of pain as an example, it is well known that stimulation of the cholinergic PPTg nucleus can produce an anti-nociceptive effect (Iwamoto, 1991;Carlson et al., 2004;Iwamoto and Marion, 1993;Dias et al., 2009;Pertovaara, 2013) and that this effect is largely attenuated by intra-thecal administration of a-adrenoceptor blockers (Iwamoto and Marion, 1993;Dias et al., 2009). This result shows that activation of cholinergic neurons in the PPTg nucleus can produce analgesia by activating the descending NAergic system, implying a functional connection between the PPTg nucleus and the pontine NAergic system. However, connections between the PPTg nucleus and pontine NAergic neurons have not been proved and the mechanism by which PPTg nucleus inputs drive pontine NAergic neurons has not been examined. In this study, we addressed these issues using the PPTg-NAergic A7 connection as a model. By using a brainstem slice preparation (Min et al., 2008) that allowed us to make whole-cell recordings from NAergic A7 neurons and apply extracellular stimulation to the PPTg nucleus, we characterized synaptic transmission from the axonal terminals of PPTg nucleus neurons to NAergic A7 neurons. We found cholinergic and glutamatergic transmission working together to produce long lasting depolarization of the membrane potential in NAergic A7 neurons that provides an effective means for the PPTg nucleus to drive the firing of NAergic neurons.

[1] 136w In this serial of experiments, the precise BDA injection site within the PPTg nucleus was identified in 3 animals; one example is shown in Fig. 1. Fig. 1A shows a set of 3 consecutive sections that were stained for BDA alone (Fig. 1A1), for cholinergic neurons using anti-ChAT antibodies alone (Fig. 1A2), or for BDA and NAergic neurons using anti-DBH antibodies (Fig. 1A3). Camera lucida drawings for 2 consecutive sets of sections were then superimposed. Sections stained for BDA and ChAT showed that BDA was delivered to a small spot ~300 mm in diameter located immediately rostral to the superior cerebellar peduncle (scp) (see Fig. 1A, B). The BDA injection site was within a cluster of ChATimmunoreactive (-ir) neurons (Fig. 1A, B), the location of which was consistent with that described by Paxinos and Watson (1998).

[2] 113w In sections subjected to combined BDA and anti-DBH staining (Fig. 1A3, C), DBH-ir elements were stained brown-red and could be easily differentiated from the BDA-labeled axonal terminals (darkblue, see arrows and arrowheads in Fig. 1C). The DBH-ir neurons were found beneath the spc and a region ~500 mm rostral to the trigeminal motor nucleus (Mo5) (see Fig. 1A3), and were presumed to be NAergic neurons of the A7 catecholamine cell group (Min et al., 2008). At high magnification, synaptic connections were observed between BDA-labeled axonal terminals and either DBH-ir dendrites (Fig. 1C1, C3) or DBH-ir soma (Fig. 1C2). Together, these results show that PPTg axonal terminals make synaptic connections with NAergic A7 neurons.

[3] 162w If these synaptic connections between the BDA-labeled axons and the DBH-ir dendrites and cell bodies were cholinergic and functional, the NAergic A7 neurons would be expected to express cholinergic receptors and electrical stimulation of the PPTg nucleus should evoke synaptic currents mediated by cholinergic receptors in NAergic A7 neurons. Expression of cholinergic receptors was tested in the studies shown in Figs. 2 and 3, while the effect of stimulation of the PPTg nucleus was tested in the studies shown in Figs. 4e9. The recordings in Figs. 2 and 3 were made in the presence of 100 mM Ptx, 1 mM strychnine, and 5 mM kynurenic acid to block fast synaptic transmission, those in Figs. 4e6 and 8 were made in the presence of 100 mM Ptx and 1 mM strychnine for isolation of EPSCs, and those in Fig. 9 were made in same condition as for isolation of EPSCs except that 1 mM TTX was added for isolation of miniature EPSCs (mEPSCs).

[4] 121w Expression of cholinergic receptors on NAergic A7 neurons was tested by examining the effect of bath application of a cholinergic muscarinic receptor (mAChR) agonist on the activity of NAergic A7 neurons, as activation of mAChRs has been reported to increase the firing activity of NAergic LC neurons (Berridge and Foote, 1991). To confirm that the recording was made from NAergic A7 neurons, previously described physiological criteria and post hoc IHC with anti-DBH antibodies were used (Min et al., 2008). Fig. 2A shows an example of a bright field image of a brainstem slice (A1) stained with anti-DBH antibodies, a recorded neuron filled with biocytin (A2) and showing DBH-ir in the slice (A3) and the merged images in A2 and A3 (A4).

[5] 808w Under recording conditions in which 100 mM Ptx, 1 mM strychnine, and 5 mM kynurenic acid were added to the bath medium to block fast synaptic transmission, application of 25 mM carbachol (CCh) for 100 s increased the spontaneous firing rate of NAergic A7 neurons from 2.3 ± 1 Hz to 5.5 ± 06 Hz (n ¼ 3 cells, p < 0.02 paired t-test; Fig. 2B). Consistent with this observation, CCh application for 100 s induced an inward current (Fig. 3A), referred to as the I CCh , in voltage-clamp recordings with the Vm clamped at À70 mV. To explore the cellular and ionic mechanisms of I CCh , we compared I CCh induced in control and in ACSF containing a test drug targeting on different types of mAChR and ion channel condition in a same cell. We first examined the response of a NAergic A7 neuron to repeated application of CCh, each for 100 s, and observed a slight decline of I CCh amplitude, which reached to a stable level since the second CCh application. Accordingly, to avoid possible interference of the results with the observed desensitization of mAChR signaling, we applied CCh 3 times as in Fig. 3A and applied the test drug between the second and third CCh application, starting after full recovery of the whole-cell current produced by the second CCh application. The amplitudes of the I CCh induced by the second and third CCh application were measured and the ratio of third to second application, I CCh ratio, was calculated. Under control conditions, namely, no test drug was applied, the I CCh ratio estimated from 6 cells was 1.08 ± 0.11 (Fig. 3B, top trace and 3F), showing no significant attenuation of the I CCh between the second and third CCh application. When 1.5 mM atropine, an mAChR antagonist, was applied between the second and third CCh applications, the I CCh ratio was markedly reduced to 0.20 ± 0.05 (Fig. 3B, bottom trace and 3F) (n ¼ 7 cells; p < 0.01 compared to control, Mann-Whitney U test). As shown in Fig. 3C and F, when another mAChR antagonist, himbacine, was used, no significant effect was seen on the I CCh ratio (0.95 ± 0.17) at a concentration of 1 mM (top trace; n ¼ 5 cells; p ¼ 0.749 compared to control, Mann-Whitney U test), but the I CCh ratio was significantly reduced to 0.51 ± 0.06 at 2 mM himbacine (bottom trace, n ¼ 5 cells; p < 0.01 compared to control, Mann-Whitney U test). Since the IC 50 of himbachine for M1 receptor (M1R) and M3 receptor (M3R) are 220 and 900 nM, respectively, while those for M2 receptor (M2R) and M4 receptor (M4R) are 15 nM and 46 nM, respectively (Waelbrock et al., 1990;Miller et al., 1992) and only high concentration of himbachine could reduced I CCh amplitude, M1R and M3R were likely mediating the I CCh . To test this argument, 4-DAMP, a more selective antagonist for M3R, was used. Application of 100 nM 4-DAMP resulted a significant reduction of I CCh ratio to 0.35 ± 0.10 (n ¼ 4 cells, p < 0.01 compared to control, Mann-Whitney U test) (top trace in Fig 3D and F). Coapplication of 4-DAMP with 1 mM Vu0225035, a selective M1R antagonist, further reduced I CCh ratio to 0.09 ± 0.03 (n ¼ 5 cells, p < 0.05 compared to 4-DAMP application, Mann-Whitney U test) (Fig. 3 F). Application of Vu0225035 also significantly reduced I CCh ratio to 0.75 ± 0.03 (n ¼ 5 cells, p < 0.05 compared to control, Mann-Whitney U test) (middle trace in Fig. 3D and F), though the amount was less then that of 4-DAMP application (p < 0.05 compared to 4-DAMP, Mann-Whitney U test). In contrast, application of AF-Dx116, a selective M2R antagonist, did not have a significant effect on I CCh ratio (0.96 ± 0.16; n ¼ 5 cells, p ¼ 0.812 compared to control, Mann-Whitney U test) (bottom trace in Fig. 3D and F). Replacing extracellular Na þ with an equimolar concentration of NMDG markedly reduced the I CCh ratio to 0.11 ± 0.05 (n ¼ 7 cells; p < 0.01 compared to control, Mann-Whitney U test) (upper trace in Fig. 3E and F), while 2-APB (100 mM), a cation channel blocker, also significantly reduced the I CCh ratio to 0.57 ± 0.08 (lower trace in Fig. 3E and F; n ¼ 7 cells), and 2 other cation channel blockers, ruthenium red (50 mM) and SKF96365 (100 mM), also caused, respectively, a reduction to 0.32 ± 0.07 (n ¼ 5 cells) or 0.38 ± 0.12 (n ¼ 4 cells) (Fig. 3F) (all p < 0.01 compared to control, Mann-Whitney U test). Together, these results show that NAergic A7 neurons express M1R and M3R, and activation of which produces depolarization through with opening cationic channels.

[6] 44w We next examined the effect of electrical stimulation of the PPTg nucleus on NAergic A7 neurons. Fig. 4A shows the arrangement of the bipolar electrode for electrical stimulation of the PPTg nucleus and recording of a NAergic A7 neuron in a sagittal brainstem slice.

[7] 315w To isolate EPSCs, 100 mM Ptx and 1 mM strychnine were added to the bath to block fast inhibitory synaptic transmission. Under these conditions, electrical stimulation of the PPTg nucleus with a single constant current-pulse (0.5 ms duration) evoked fast EPSCs that were blocked by application of 10 mM DNQX, a non-NMDA receptor antagonist, showing that the EPSCs were glutamatergic (Fig. 4B). Stimulation with a train of 10 pulses at 100 Hz evoked large and long-lasting EPSCs, the amplitude of which was partially, but markedly, reduced by application of 10 mM DNQX, leaving residual slow EPSCs (two superimposed traces in Fig. 4Ci). Of 13 cells tested, dramatic attenuation of the residual EPSCs left after DNQX application was observed in 10 when DNQX and atropine were coapplied (two superimposed traces in Fig. 4Cii), while no effect was observed in the remaining 3 (Fig. 4D). Pooling of the data from all 13 cells over the period of 10e12 min showed that atropine application significantly reduced the residual EPSCs to 56 ± 9% compared to the 3 min before atropine application (Fig. 4E, diamonds; p < 0.01, paired t-test). To compare this result at 10e12 min with results for the same time period in the absence of atropine, an additional 5 experiments were performed as above, but in the absence of atropine (see Fig. 4E, filled circles). Although the activity showed a significant rundown to 78 ± 5% of the control level at 10e12 min (dotted rectangle in Fig. 4E), the extent of EPSCs reduction was significantly different in the presence or absence of atropine (Fig. 4E; p < 0.05, Mann-Whiney U test). These results show that, although rundown of the cholinergic activity occurred under control conditions in some axonal terminals, possibly due to depletion of cholinergic vesicles by stimulation at 100 Hz, stimulation of the PPTg nucleus evoked atropine-sensitive (i.e. mAChRsmediated) slow EPSCs in NAergic A7 neurons.

[8] 119w To confirm the slow cholinergic EPSCs were specific to PPTg stimulation, we evoked synaptic activity with stimulating electrode placed within Mo5 or in area rostroventral to Mo5 (RVMo5), with the distance to A7 of the both stimulation positions being similar to that of the PPTg to A7 (see Fig. 4A). In both cases, EPSCs were evoked with a stimulating train consisting of 10 pulses at 100 Hz and application of DNQX markedly reduced the activity (Fig. 4F, Mo5 stimulation) (Fig. 4G, RVMo5 stimulation); subsequent application of atropine did not have a further effect on the activity (Fig. 4FeH). These results show that the slow cholinergic EPSCs could only be evoked by stimulation of PPTg but not other adjacent areas.

[9] 533w A very impressive feature of the response of NAergic A7 neurons to a stimulation train of 10 pulses at 100 Hz was the occurrence of spontaneous-like EPSC activity, some of which occurred even as late as one second after the end of delivery of the stimulation train. Fig. 5A1 shows a sweep (top black trace) of a representative recording from a NAergic neuron in the presence of 100 mM Ptx and 1 mM strychnine. The episodes before (horizontal bars 1), after (horizontal bars 3 and 4) and during stimulation (horizontal bars 2) of the sweep are shown at a faster and larger scale (see green traces 1e6). As can be seen, each pulse in the stimulation train synchronously evoked EPSCs (see green traces 2), followed by many spontaneous-like events (green traces 3 and 4). These events were electrical stimulation-evoked, rather than real spontaneous synaptic activity, as their incidence was much lower before stimulation (compare green traces 1 and 3, 4). Bath application of 10 mM DNQX blocked these the spontaneous-like events (Fig. 52A; compare green traces 3, 4 in Fig. 5A1 and green traces 3 and 4 in Fig. 5A2). These results show the evoked spontaneous-like events were glutamatergic. Although their amount varied, these evoked spontaneous-like EPSCs were observed in every recorded NAergic A7 neurons (also see arrow in Fig. 4Ci). We speculated that the activity might result from the asynchronous release of glutamatergic vesicles triggered by the stimulation train (referred to as asynchronous EPSCs, aEPSCs). If this argument were true, manipulations that alter the presynaptic calcium profile produced by the stimulation train would be expected to have an affect on this activity. We therefore tested the effect of two manipulations, the number of pulses in the stimulation train and the concentration of extracellular calcium. To quantify the aEPSCs, we measured the charge transfer over a 1 s window starting 100 ms after delivery of the stimulation train (see shaded gray area in Fig. 5A1). As shown in Fig. 5B, the charge transfer increased in a pulse number-dependent manner (n ¼ 5 cells) and, as shown in Fig. 5C, could be increased or decreased by, respectively, increasing or decreasing the extracellular Ca 2þ concentration. Lowering the extracellular Ca 2þ Fig. 7. PPTg projections to NAergic A7 neurons contain glutamatergic fibers. A. Photographs show BDA injection site (see asterisk) using dark Ni-DAB reaction method. Mo5, trigeminal motor nucleus; scp, superior cerebellar peduncle. B. Merged images of immunofluorescent stain using anti-VGluT2 (red) and anti-DBH (blue) antibodies. C. Merged images of B and BDA-fluorescent stain (green). D. Enlargement of dashed-circle in B. Arrowheads and arrow mark 3 BDA-labeled terminals showing VGluT2-ir with one (see arrow in D) making synaptic connection on a DBH-ir neuron. E. Merged images of BDA-fluorescent stain (green) and stain of anti-DBH antibodies. Arrows mark a BDA-labeled fiber of which a large bouton makes connection on dendrite of a DBH-ir neuron (see dashed circle). F. Enlargement of dashed-circle in E show merged images of stains of anti-VGLUT2 (red) and anti-DBH (blue) antibodies. G. Merged images of F and BDA-fluorescent stain (green). Arrow marks a BDA-labeled bouton (see dashed circle in E) that makes contact on dendrite of a DBH-ir neuron but is non-VGLUT2-ir.

[10] 201w concentration from 2.5 mM to 0.5 mM reduced the charge transfer to 58.8 ± 9.6% of the control level (n ¼ 6; p < 0.01, paired t-test), and a subsequent increase to 4 mM increased it to 133.2 ± 11.7% of the control level (n ¼ 5; p < 0.05, paired t-test). Finally, we examined the effect of EGTA-tetra-acetoxymethyl ester (EGTA-AM) on the activity. EGTA-AM is a Ca 2þ chelator with slow chelating dynamics that can be loaded into the cell by bath application and can therefore block the slow rise in Ca 2þ required to trigger asynchronous release of glutamate, but has no effect on synchronous release, which requires a rapid increase in Ca 2þ upon arrival of the action potential at axonal terminals (Ali and Todorova, 2010;Hefft and Jonas, 2005;Iremonger and Bains, 2007). As shown in Fig. 6A, bath application of 50 mM EGTA-AM markedly reduced the aEPSCs to 58.9 ± 8.1% of the control level (n ¼ 6, p < 0.005, paired t-test), but only reduced the synchronous EPSCs, measured as charge transfer produced by the stimulation train, to 83.4.9 ± 7.7% of the control level, which was not significant (n ¼ 6, p ¼ 0.09, paired t-test).

[11] 225w In current clamp recordings, the stimulation train (10 pulses at 100 Hz) could averagely evoke 5 action potentials, and this was followed by additional 2e5 action potentials of spontaneous discharge (upper traces in Fig. 6B1). Consistent with the voltageclamp recording observations, application of EGTA-AM markedly reduced the number of action potentials generated after, but not during, the stimulation train (see lower traces in Fig. 6B1). Before EGTA-AM application, the averaged number of action potentials generated during and after stimulation were, respectively, 4.8 ± 0.8 and 3 ± 0.7; in the presence of EGTA-AM, the corresponding values were 4.0 ± 0.7 and 0.5 ± 0.5 (n ¼ 4 cells) (Fig. 6B2). The reduction by EGTA-AM of action potential number generating after the stimulation train was significant (p < 0.005, paired t-test); it was, however, not significant for action potential number generated during the stimulation train (p ¼ 0.318, paired t-test). Similarly, the area of the aEPSP integrated over a 1 s window starting from the end of the stimulation train was reduced to 42 ± 9% of control levels by EGTA-AM (Fig. 6B3, n ¼ 4 cells, p ¼ 0.07, paired t-test). Together, these results show that PPTg terminals can synchronously and asynchronously release glutamate vesicles, which results in a longlasting voltage depolarization and an increase in the spiking window of the postsynaptic NAergic A7 neurons.

[12] 190w Since it was possible that the fast EPSCs, including both synchronous EPSCs and aEPSCs, was due to stimulating glutamatergic fibers of passage of PPTg rather than neurons in PPTg, we investigated whether PPTg projection to NAergic A7 neurons contained glutamatergic fibers. Again, 10% BDA was injected into PPTg and sections comprising A7 area were stained either with dark Ni-DAB method for identification of the injection site or with fluorescent markers for visualizing BDA-labeled fibers (see green color in Fig. 7), DBH-ir neurons (see blue color in Fig. 7) and presynaptic elements showing immunoreactive to vesicular glutamate transporter 2 (VGluT2), a marker for glutamatergic terminals (see red color in Fig. 7). An example of the results is shown in Fig. 7; as can be seen, the BDA injection site was again confirmed within the PPTg nucleus (Fig. 7A), and in the consecutives section stained with fluorescent method there were BDA-labeled terminals making synaptic connections on a DBH-ir soma (Fig. 7BeD) or dendrite (Fig. 7EeG). These DAB-labeled fibers were either VGluT2-ir (Fig. 7BeD) or non-VGluT2-ir (Fig. 7EeG). These observations supported existence of glutamatergic neurons in PPTg that project to NAergic A7 neurons.

[13] 577w Interestingly, addition of 5 mM atropine resulted in an increase in both the amplitude and frequency of the aEPSCs, but had no effect on the synchronous EPSC activity (Fig. 8). Application of atropine significantly increased the amount of charge transfer during aEPSC activity to 136 ± 23% of the control level (Fig. 8A2, light gray bars; n ¼ 6 cells, p < 0.05, paired t-test), while the synchronous EPSCs remained unchanged at 97 ± 11% of the control level (Fig. 8A2, dark gray bars, p ¼ 0.489, paired t-test). One interpretation of this observation is that there were mAChRs on the axonal terminals of PPTg glutamatergic neurons and that ACh released from these terminals by delivery of the stimulation train might act on these receptors to provide presynaptic inhibition of glutamatergic transmission onto NAergic A7 neurons. To test these two ideas, the effect of atropine on EPSCs evoked by the 10th pulse of a stimulation train of 10 pulses at 10 Hz was examined. The hypothesis here is that delivery of the stimulation train would markedly increase the extracellular ACh concentration and, if MRs are present on the glutamatergic PPTg terminals and exert presynaptic inhibition, the EPSCs evoked by the 10th pulse would be markedly lowered due to the activation of presynaptic mAChRs by ACh released from stimulated cholinergic PPTg terminals during delivery of the stimulation train and this inhibition would be blocked by atropine, resulting in an increase in amplitude of the 10th EPSC. In support of this argument, as shown in the example in Fig. 8B1 and the summarized results in Fig. 8B2, the amplitude of the 10th EPSC was increased to 167 ± 28% of control upon atropine application (n ¼ 9 cells, p < 0.05, paired t-test). To confirm presynaptic modulation of EPSCs by mAChRs, we also examined the effect of CCh on miniature EPSCs (mEPSCs). Application of CCh markedly increased inter-event interval of the mEPSCs to 293 ± 78% of control levels (n ¼ 5 cells, p < 0.05, paired t-test), but only caused a non-significant increase in the amplitude to 127 ± 12% of control levels (p ¼ 0.0781, paired t-test) (Fig. 9A, E). Similar results were also observed with addition of AF-DX116 into ACSF, showing M2R was not involved in mediating presynaptic inhibition. In the presence of AF-DX116, application of CCh still markedly increased interevent interval of the mEPSCs to 293 ± 78% of control levels (n ¼ 5 cells, p < 0.05, paired t-test), and had no significant effect on the amplitude (127 ± 12% of control levels; p ¼ 0.0781, paired ttest) (Fig. 9 B, F). In contrast, the effect of CCh on inert-event interval of the mEPSCs was strongly inhibited by either addition of 4-DAMP (n ¼ 3 cell) (Fig. 9C and black lines in Fig. 9G) or Vu0225035 (n ¼ 3 cell) (Fig. 9D and gray lines in Fig. 9G). Application of 4-DAMP or Vu0225035 resulted in a change of, respectively, 91 ± 36% or 144 ± 33% of control for intereevent interval and 103 ± 3% or 95 ± 8% of control for amplitude. Pooling together the data of 4-DAMP and Vu0225035 application showed a change of 101 ± 17% (p ¼ 0.476, paired t-test) and 99 ± 4% (p ¼ 0.693, paired t-test) of control for inter-event interval and amplitude, respectively. Together, the above results show blockade of the effect of CCh application on mEPSCs by M1R and M3R antagonists but not M2R antagonist.

[1] 107w In this study, we provided morphological and physiological evidence for synaptic connections between the PPTg nucleus and NAergic A7 neurons. We found that inputs from the PPTg to A7 area consisted of cholinergic and glutamatergic fibers, which together produced a long-lasting depolarization to increase the time window for spiking of NAergic neurons. The long-lasting depolarization was achieved through activation of M1Rs and M3Rs by cholinergic inputs and asynchronous release of a large number of glutamate vesicles from glutamatergic inputs. In addition, presynaptic M1Rs and M3Rs were shown to be present on glutamatergic terminals and to provide inhibitory feedback regulation of PPTg nucleus inputs onto NAergic A7 neurons.

[2] 276w We demonstrated that deposition of BDA in the PPTg nucleus labeled axonal terminals in A7 area and some of them were VGluT2-ir that together with non-VGluT2-ir of the BAD-labeled terminals made synaptic connections on NAergic A7 neurons. These connections were functional, as extracellular stimulation of the PPTg nucleus evoked EPSCs in NAergic A7 neurons in brain slice preparations. The EPSCs have both cholinergic and glutamatergic components, as, in most of the NAergic neurons examined, the activity was only completely blocked if DNQX and atropine were co-applied. These observations are consistent with results in a previous study using in situ hybridization and IHC that showed that the PPTg nucleus contains glutamatergic, cholinergic, and GABAergic neurons (Wang and Morales, 2009). However, whether ACh and glutamate were released from separate axonal terminals or co-released from the same terminals remains a matter of debate and this could not be resolved by the morphological or the physiological approaches used in the present study. Wang and Morales (2009) argued that co-release of ACh and glutamate is unlikely to be a major factor in the interactions between ACh and glutamate at the postsynaptic site, since they found that the vast majority of PPTg cholinergic neurons lack transcripts encoding essential proteins for the vesicular glutamate transporter. However, co-localization of cholinergic and glutamatergic markers has been suggested in other brain areas. In the basal forebrain, cholinergic neurons were reported to express glutamatergic marker proteins in a study using IHC and in situ hybridization approaches (Nickerson Poulin et al., 2006), while, in the hebanula, an epithalamic area, optogenetic activation of cholinergic neurons causes both cholinergic and glutamatergic transmission at the postsynaptic neurons (Ren et al., 2011).

[3] 215w Our pharmacological study showed that, in NAergic A7 neurons, I CCh was mediated by M1Rs and M3Rs through with opening cationic channels. Since 2-APB, SKF96365 and ruthidium red are not selective to specific type of ion channel and have multiple other effects, the molecular identity of the cationic channel gated by M1Rs and M3Rs in NAergic A7 neurons remains to be determined. Previous studies have shown that activation of M3Rs produced an excitatory effect through the gating of the canonical type of transient-receptor-potential channels (TRPCs) in gastrointestinal smooth muscle (Zholos, 2006) or through activating the Na þ /Ca2 þ exchanger in neurons in the sublaterodosal nucleus in the pons (Weng et al., 2014). TRPCs are known to be gated by the Gq-PLC (phospholipase C) signaling pathway of G-protein-coupled receptors and have been shown to be involved in regulating the membrane potential in CNS neurons (Clapham, 2003;Bollimuntha et al., 2011). We have previously reported that substance P acting on neurokinine 1, a G-protein-coupled receptor, gates the TRPC6 channels via the Gq-PLC signaling pathway and increases neuronal excitability in NAergic A7 neurons, (Min et al., 2009). Since M1R and M3R also active Gq-PLC signaling pathway, it is likely that TRPCs could potentially be the molecules acting downstream of M1R and M3R receptor activation in NAergic A7 neurons.

[4] 211w An impressed feature of the glutamatergic synaptic transmission from the PPTg onto NAergic A7 neurons was the asynchronous release of a large number of synaptic vesicles by stimulation using a train of 5e10 pulses at 100 Hz. Asynchronous release, defined as the release of neurotransmitter-filled vesicles several hundred milliseconds after presynaptic stimulation, has been reported in both excitatory and inhibitory synaptic transmission in the CNS (Atluri and Regehr, 1998;Lu and Trussell, 2000;Otsu and Murphy, 2004;Hefft and Jonas, 2005;Iremonger and Bains, 2007) and is proposed to amplify presynaptic activity, which, in turn, results in a prolonged spiking of postsynaptic neurons. This form of transmission may be functionally important in translating synaptic signals into a specific spiking pattern of postsynaptic neurons (Iremonger and Bains, 2008). In the paraventricular and supraoptic nuclei of the hypothalamus, the secretory neurons releasing vasopressin and oxytocin have been shown to receive glutamatergic inputs that release glutamate asynchronously over several hundreds of milliseconds (Iremonger and Bains, 2007). The integration of these synaptic events provides sufficiently long-lasting depolarization of the membrane voltage to sustain high frequency discharge of the postsynaptic neurons, an important factor in the regulation of hormone release from these hypothalamic neurons, the rate of which is roughly proportional to the neuronal firing rate (Iremonger and Bains, 2008).

[5] 199w The regulation of NA release may be similar to that of secretory neurons in hypothalamus. The rate of NA release from NAergic terminals in the neocortex or the anteroventral thalamic nucleus has been shown to have, respectively, a linear or non-liner correlation with the discharge rate of NAergic LC neurons, ranging from a basal discharge level of approximately 1.5 Hze5 Hz (Brun et al., 1993;Berridge and Abercrombie, 1999). In addition to release from axonal terminals, NAergic LC neurons also undergo somatic/ dendritic NA release, and a physiological role for this kind of NA release has been suggested to be negative feedback autoregulation of NAergic LC neurons via a-2A adrenoreceptors (Sevensson et al., 1975;Callado and Stamford, 2000). Likewise, while little NA is released at low frequency discharge, a large somatic/dendritic release of NA is recorded in the LC when the action potential frequency of NAergic LC neurons is increased by an excitatory input (Huang et al., 2007). As in hypothalamic neurons, the transmission of PPTg glutamatergic neuron input to NAergic neurons with asynchronous release of quanta may provide an efficient means of regulating the rate, and of shaping the pattern, of NAergic neuron firing and thus the NA release profile.

[6] 315w In addition to the asynchronous release of quantal transmitter from glutamatergic terminals, our results showed that inputs from the PPTg nucleus could also produce long-lasting depolarization of the membrane voltage in NAergic A7 neurons via release of ACh acting at M1Rs and M3Rs. The threshold for recruiting cholinergic transmission appears to be higher than that for recruiting glutamatergic input, as slow EPSCs mediated by M1Rs and M3Rs were only observed when a train, and not a single pulse, of stimulation was used. Interestingly, extracellular stimulation with a train of 10 pulses at 100 Hz not only evoked slow mAChR-mediated EPSCs, but also the asynchronous release of large number of glutamatergic quanta. These dual components of transmission may provide more flexible regulation of both long-lasting membrane depolarization and inhibitory presynaptic feedback control of PPTg transmission to NAergic A7 neurons. The mAChR-mediated inhibitory presynaptic feedback control of glutamatergic input is clearly seen in the results showing that bath application of atropine significantly increased the incidence of aEPSCs and the amplitude EPSCs that were evoked by a single pulse of stimulation preceded with a stimulation train of 9 pulses at 10 Hz. Moreover, application of CCh significantly decreased the frequency of glutamatergic mEPSCs and M1R and M3R antagonist but not M2R antagonist attenuated this effect. Consistent with the present observations, presynaptic modulation of excitatory and inhibitory synaptic transmission by M1Rs and M3Rs have been reported in many brain regions, such as basolateral amygdalar nucleus (Muller et al., 2013), hippocampus (de Vin et al., 2015), and globus pallidus (Hern andez-Martínez et al., 2015). Surprisingly, in our study, atropine application did not have a significant effect on synchronous EPSCs. One explanation could be that the increased charge transfer of glutamatergic EPSCs during the stimulation train due to the blockade of mAChR-mediated inhibitory presynaptic feedback might be compensated by the blockade of mAChR-mediated slow EPSCs that occurred during the stimulation train.

[7] 116w The A7 catecholamine cell group consists of NAergic neurons that have intense axonal projections to the spinal cord, including the dorsal and ventral horns (Clark and Proudfit, 1991;Howorth et al., 2009;Bruinstroop et al., 2012). Furthermore, many previous morphological and behavior studies have shown that NAergic A7 neurons are involved in modulating nociceptive signaling at the spinal cord level and form a part of the descending analgesia system (e.g. Burnett and Gebhart, 1991;Yeomans et al., 1992;Holden et al., 1999). Our results therefore suggest possible synaptic mechanisms underlying the previously reported antinociceptive effect caused by activation of the PPTg nucleus cholinergic neuron system that involves recruitment of the descending NAergic system (Iwamoto and Marion, 1993;Dias et al., 2009).

[8] 237w In conclusion, we report direct synaptic transmission from the PPTg nucleus onto pontine NAergic neurons that involves cooperation of cholinergic and glutamatergic transmission. This form of transmission with dual transmitters provides not only long-lasting depolarization of the membrane potential to drive the firing rate of NAergic neurons, which may correlate with axonal and somatic/ dendritic release of NA (Brun et al., 1993;Berridge and Abercrombie, 1999;Huang et al., 2007), but also mAChR-mediated inhibitory presynaptic feedback control of transmitter release. Our brain slice preparations only contained the PPTg and NAergic A7 neurons and so our investigation was limited to PPTg-Nergic A7 neuron transmission. There are some similarities in physiological properties between different pontine NAergic neurons, for example, NAergic LC and A7 neurons can both fire action potentials spontaneously at a similar frequency in a slice preparation (Min et al., 2008;Wang et al., 2015), both have an A-type potassium current that regulates their firing rate (Ishimatsu and Williams, 1996;Min et al., 2010;Wang et al., 2015), and both are subject to tonic inhibition mediated by GABA B receptors (Wu et al., 2011;Wang et al., 2015). It is therefore likely that, in addition to NAergic A7 neurons, this kind of synaptic transmission with a dual transmitter system, namely cholinergic transmission with mAChR activation and glutamatergic transmission with asynchronous release of quanta, may also occurs in synaptic transmission between PPTg inputs onto other pontine NAregic neurons, such as LC and A5 NAergic neurons.

[1] 223w The use of animals in this study was approved by the institutional (National Taiwan University and Chung-Shan Medical University) Ethics Committee for Animal Research, the guidelines of which comply with the European Communities Council Directive (24 November 1986). Male Sprague-Dawley rats weighting 300e400 g were anesthetized with 5% isoflurane in oxygen and placed on a stereotaxic apparatus to allow the plane of the skull to be adjusted to a horizontal position, then a small craniotomy was performed and the dura reflected. A glass pipette loaded with 10% biotinylated-dextran-amine 1000 (BDA; Invitrogen, Carlsbad, CA) in 0.01 M phosphate buffer (PB), pH 7.4, tilted at 28 to the vertical was placed into the PPTg nucleus of the left brainstem at the coordinates of 3.30 mm posterior to the bregma, 1.9 mm lateral to the midline, and 8.2 mm ventral to the surface of the cerebellar cortex (Paxinos and Watson, 1998). The tracer was injected by iontophoresis using a 5 mA DC current (50% duty, cycle length 14 s) for 10 min, then the glass pipette was left in place for a further 10 min to minimize diffusion of the tracer along the pipette shank, after which the pipette was removed, the scalp sutured, and the rats returned to their home cages for a period of 5e7 days to allow uptake and transportation of the BDA.

[2] 135w Animals of both sexes aged 8e10 days were anesthetized with 5% isoflurane in O 2 and decapitated, then their brains were rapidly exposed and chilled with ice-cold artificial cerebrospinal fluid (ACSF) consisting of (in mM): 119 NaCl, 2.5 KCl, 1.3 MgSO 4 , 26.2 NaHCO 3 , 1 NaH 2 PO 4 , 2.5 CaCl 2 , and 11 glucose, oxygenated with 95% O 2 and 5% CO 2 , pH 7.4. Sagittal brainstem slices (300 mm) containing the PPTg nucleus and A7 were prepared using a vibroslicer (D.S.K. Super Microslicer Zero 1, Dosaka EM, Kyoto, Japan) and maintained in a moist air-ACSF interface chamber and allowed to recover for at least 90 min before being transferred to an immersion-type chamber mounted on an upright microscope (BX51WI, Olympus Optical Co., Ltd., Tokyo, Japan) for recordings.

[3] 347w Neurons were viewed using Nomarski optics. Patch pipettes were pulled from borosilicate glass tubing (1.5 mm outer diameter, 0.32 mm wall thickness; Warner Instruments Corp., Hamden, CT, USA) and had a resistance of about 5e8 MU when filled with the pipette solution consisting of (in mM) 131 K-gluconate, 20 KCl, 10 HEPES, 2 EGTA, 8 NaCl, 2 ATP, and 0.3 GTP, pH adjusted to 7.2 with KOH. Recordings were made in the whole-cell configuration with a patch amplifier (Multiclamp 700B; Molecular Device, Sunnyvale, CA, USA) at room temperature (~25 C). For current-clamp recording, the bridge was balanced and recordings were only accepted if the recorded neuron had a membrane potential (Vm) of at least À45 mV without applying a holding current and if the action potential was able to overshoot 0 mV. For voltage-clamp recording, neurons were clamped at À70 mV unless specified otherwise. The serial resistance was monitored throughout recording and the data discarded if the values varied by more than 20% of the original value, which was less than 20 MU. Extracellular stimulation was performed by delivering a stimulating train through a bipolar stainless steel electrode (FHC, Bowdoinham, ME 04008 USA) positioned in the PPTg nucleus. The stimulating train consisted of 10 constant-current pulses (50e250 mA; 100 ms) at 100 Hz and was delivered every 30 s. Signals were low-pass filtered at a corner frequency of 2 kHz and digitized at 10 kHz using a Micro 1401 interface running Signal software for episode-based capture or Spike2 software for continuous recording (Cambridge Electronic Design, Cambridge, UK). Miniature excitatory postsynaptic currents (mEPSCs) were recorded in the presence of 1 mM tetrodotoxin (TTX) and were measured and analyzed using the Mini-Analysis program (Synaptosoft Inc., NJ, USA). All data are presented as the mean ± standard error of the mean (SEM) and were compared using the paired t-test and non-parametric Mann-Whitney U test. The Kolmogrov-Smirnov test was used for comparison of interevent interval and amplitude of mEPSCs before and after application of a test drug in an individual experiment. The criterion for significance was a p value < 0.05.

[4] 49w Paraformaldehyde and all chemicals used to prepare the ACSF and pipette solution were from Merck (Frankfurt, Germany), atropine, baclofen, biocytin, N-methyl-D-glucose (NMDG), picrotoxin (Ptx), and strychnine were from Sigma (St. Louis, USA), and AF-Dx116, 2-APB, 4-DAMP, 6,7-dinitroquinoxaline-2,3-dione (DNQX), EGTA-AM, himbacine, ruthenium red, SKF96365, Vu0225035 were from Tocris-Cookson (Bristol, UK).

[1] 55w The animals were deeply anesthetized with sodium pentobarbitone and perfused via the cardiovascular system with normal saline, then with fixative consisting of 4% paraformaldehyde (Merck, Frankfurt, Germany) in 0.1 M PB. The brains were then rapidly removed, placed in the same fixative at 4 C for 3e4 h, stored overnight in cold (4 C) 0.

[2] 115w 1 M PB, and transferred to 30% sucrose in 0.01 M PB for cryoprotection. Serial sagittal brainstem Abbreviations: A7 A7 catecholamine cell groups ACh Acetylcholine ACSF Artificial cerebrospinal fluid BDA Biotinylated-dextran-amine 100T carbachol CCh ChAT O-acetyltransferase CNS Central nervous systems DAB Diaminobenzidine DBH Dopamine-b-hydroxylase DNQX 6,7-dinitroquinoxaline-2,3-dione EGTA-AM EGTA-tetra-Acetoxymethyl Ester EPSCs Excitatory postsynaptic currents aEPSC asynchronous EPSCs mEPSCs miniature EPSCs IHC immunohistochemistry -ir -immunoreactive LC Locus coeruleus mAChR Cholinergic muscarinic receptor Mo5 Trigeminal motor nucleus RVMo5 Brain area rostroventral to Mo5 NA Noradrenaline NAergic noradrenergic NMDG N-methyl-D-glucose PB Phosphate buffer PPTg Pedunculopotine tegmental Ptx Picrotoxin scp superior cerebellar peduncle TPBS 0.3% Triton X-100 in phosphate-buffered saline TRPCs Transient-receptor-potential channels TTX Tetrodotoxin VGluT2 Vesicular glutamate transporter 2

[3] 335w sections (50 mm thick) containing the A7 area and PPTg nucleus were cut using a frozen sectioning technique and divided into 3 groups, each consisting of every third section. Unless specified otherwise, all reactions and incubations were performed at room temperature (~25 C) and sections were washed 3 times (10 min each) with 0.3% Triton X-100 in phosphate-buffered saline (TPBS) between the different steps. The first group of sections was processed so as to visualize the BDA injection site as described previously (Liu et al., 2015). Briefly, the sections were incubated overnight at 4 C with avidin-biotin-peroxidase complex (ABC) reagent (Vectastain ABC peroxidase kit, Vector Labs) and rinsed sequentially in TPBS and in PB, then the BDA injection site was visualized using the dark nickel (Ni)-diaminobenzidine (DAB; Sigma) reaction by incubation for 10 min with 0.1 M PB containing 0.05% DAB, 0.01% H 2 O 2 , and 0.04% Ni ammonium sulfate, followed by extensive washes in PB. The second group of sections was processed to identify cholinergic neurons using antibodies against choline O-acetyltransferase (ChAT), an enzyme involved in the synthesis of ACh. The sections were incubated for 1 h with 2% bovine serum albumin and 10% normal goat serum in TPBS, overnight at 4 C with mouse anti-ChAT antibodies (Merck Millipore, Darmstadt, Germany; 1:250 dilution in TPBS), and for 1 h with biotinylated goat anti-mouse IgG antibodies (Vector Labs, Burlingame, CA, USA; 1:200 dilution in TPBS). The sections were then incubated with ABC reagent for 1 h, then were subjected to the dark Ni-DAB reaction described above to visualize anti-ChAT immunoreactivity. The third group of sections was processed for combined BDA detection and immunohistochemistry using antibodies against dopamine-b-hydroxylase (DBH), an enzyme involved in the synthesis of NA. The sections were first processed for visualization of BDA by the dark Ni-DAB reaction, then for identification of noradrenergic neurons using mouse anti-DBH antibodies (Merck Millipore; 1:1300 dilution in PBST) as primary antibody and detecting DBH immunoreactivity by incubation with Nova Red (Vector Labs) for 1e2 min.

[4] 170w In another serial of experiments designed for determining whether PPTg projection to A7 area contained glutamatergic fibers, BDA injection combined with double-immunofluorescent methods were used. The procedures of BDA injection and the cut of frozen sections comprising A7 and the PPTg nucleus were same as described above. The sections were divided into 2 groups, each consisting of every second section. The first group of sections was processed for visualizing BDA injection site as described above; the second group of sections were stained for visualization of BDAlabeled and glutamatergic fibers and DBH-immunoreactive (-ir) elements. The sections were incubated for 1 h with 2% bovine serum albumin and 10% normal goat serum in TPBS, overnight at 4 C with mouse anti-DBH and guinea pig anti-vesicular glutamate transporter 2 (VGluT2, 1:2000; Merck, Millipore), and for 120 min with Alexa 405 anti-mouse IgG (1:200; Jackson Laboratories), Alexa 568 anti-guinea pig IgG (1:200; Jackson Laboratories) and Alexa 488 streptavidin (1:200; Jackson Laboratories). The sections were then examined with confocal microscope (Leica TCS SP5, Hamburger, Germany).

[5] 152w In some experiments, 6.7 mM biocytin was included in the internal solution used to fill the recorded neurons. The detailed procedures for viewing the biocytin-filled neurons and for post hoc immunohistochemistry (IHC) for cell type identification have been described previously (Min et al., 2008). After recording, the slices were fixed overnight at 4 C in 4% paraformaldehyde in 0.1 M PB, then were subjected to biocytin histochemistry and IHC procedures without further sectioning. The slices were incubated for 1 h at room temperature in TPBS containing 2% bovine serum albumin and 10% normal goat serum, then were incubated overnight at 4 C in PBST containing anti-DBH antibodies and avidin-AMCA, (Vector Lab; 1:200 dilution in PBST). After PBST rinses, they were incubated for 2 h with tetramethylrhodamine isothiocyanate-conjugated goat anti-rabbit IgG antibodies (Jackson ImmunoResearch, West Grove, PA, USA; 1/50 dilution in PBST), then observed under a fluorescence microscope (Aioplan 2, Zeiss, Oberkochen, Germany).