Logo of jphysiolThe Journal of Physiology SiteMembershipSubmissionJ Physiol
J Physiol. 1999 Jul 15; 518(Pt 2): 551–559.
PMCID: PMC2269434

Mechanism underlying increased neuronal activity in the rat ventrolateral periaqueductal grey by a μ-opioid


  1. The overall effect of the μ-opioid receptor agonist DAMGO (Tyr-D-Ala-Gly-MePhe-Gly-ol) on ventrolateral periaqueductal grey (PAG) neurons in brain slices was studied using the whole-cell patch-clamp recording technique.
  2. Under current-clamp conditions, DAMGO (1 μM) increased cell firing in many PAG neurons even though the opioid induced hyperpolarization and inhibited excitatory postsynaptic potentials (EPSPs) in these cells.
  3. The increase in cell activity by DAMGO was observed in both transverse and horizontal slices. The increase persisted when the membrane potential was re-depolarized to the control level. Thus, different planes of sections or the removal of Na+ channel inactivation could not account for the observation.
  4. The GABA antagonist bicuculline caused cell firing, mimicking the excitatory effect of DAMGO. Unlike DAMGO, however, bicuculline depolarized PAG cells.
  5. Under voltage-clamp conditions, with the same driving force, the evoked inhibitory postsynaptic currents (IPSCs) in these neurons were 2·3 times larger than the evoked excitatory postsynaptic currents (EPSCs). Furthermore, DAMGO inhibited IPSCs by 60·7 % while it inhibited EPSCs by 35·3 %.
  6. We propose that the overall effect of an opioid depends on the dynamic balance of its excitatory and inhibitory actions. In the PAG, the blockade of the inhibitory drive of GABAergic inputs by DAMGO is large. It overcomes the DAMGO-induced hyperpolarization and inhibition of EPSCs and thus results in the excitation of these neurons.

Opioids are thought to exert mostly inhibitory actions on neurons. They activate K+ channels (William et al. 1982) and inhibit Ca2+ channels (Schroeder et al. 1991; Moises et al. 1994), thus producing hyperpolarization in cells. Micro-injection of morphine into the ventrolateral periaqueductal grey (PAG), a brain region that participates in descending antinociceptive control (Besson et al. 1991; Morgan, 1991), inhibits pain behaviour in animals (Yaksh et al. 1988; Behbehhani, 1995). However, morphine is believed to excite PAG output neurons projecting to the medullary nucleus raphe magnus (NRM). This results in activation of NRM neurons that project to the spinal dorsal horn. They in turn inhibit nociceptive transmission in the spinal cord. This idea is based largely on the observation that many PAG neurons project to the NRM but few of them project directly into the spinal cord (Kuypers & Maisky, 1975; Castiglioni et al. 1978; Mantyh & Peschanski, 1982).

To explain the excitatory effect of opioids, a mechanism of disinhibition has been proposed (Yaksh et al. 1976; Basbaum & Fields, 1984; Reichling et al. 1988; Fields et al. 1991). It is suggested that PAG-NRM projection neurons are tonically inhibited by GABAergic interneurons in the PAG. Opioids block the activity of interneurons, thus disinhibiting PAG-NRM neurons. The evidence supporting the mechanism includes the following: (1) PAG-NRM neurons frequently receive GABAergic innervation (Reichling & Basbaum, 1990; Williams & Beitz, 1990b); (2) μ-opioid receptors are often located on GABAergic PAG neurons (Kalyuzhny & Wessendorf, 1998); (3) GABA antagonists injected into the PAG potentiate opioid-induced antinociception and GABA agonists reduce the antinociceptive effect of opioids (Depaulis et al. 1987); and (4) opioids reduce evoked IPSCs (Vaughan & Christie, 1997) and GABA antagonists increase the activity of PAG cells (Behbehani et al. 1990b) in brain slices. Not all the data, however, are consistent with the disinhibition mechanism. For example, some PAG-NRM projection neurons receive opioidergic inputs (Kalyuzhny et al. 1996; Williams & Beitz, 1990a). Opioids hyperpolarize a small but significant percentage (14 %) of PAG-NRM projection neurons in the caudal ventrolateral PAG (Osborne et al. 1996). Besides inhibition of GABA-mediated neurotransmission, opioids also inhibit EPSCs (Vaughan & Christie, 1997). To complicate the matter further, many PAG neurons receive GABAergic and glutamatergic inputs, both of which are affected by opioids (Vaughan & Christie, 1997). We therefore studied the overall effect of opioids on the activity of PAG neurons and asked the question: can an opioid excite PAG neurons when it increases K+ currents and decreases both IPSCs and EPSCs? If it can, what is the mechanism underlying the excitation?


Long-Evans or Wistar rats (11-18 days old) were anaesthetized with the inhalant methoxyflurane (dose, 0·2 mg in a 730 ml glass jar) and killed by decapitation. The brain was then removed and the midbrain containing the PAG was isolated. Coronal slices (400 μm thick) between the level of the superior and inferior colliculus, or horizontal slices containing the ventral half of the PAG (Beitz, 1985) were cut with a vibratome in ice-cold artificial cerebrospinal fluid (ACSF). The ACSF solution contained (mM): NaCl, 117; KCl, 4; CaCl2, 2·5; MgCl2, 1·2; dextrose, 11·4; NaH2PO4, 1·2; and NaHCO3, 25 (pH 7·4; osmolarity, 290-295 mosmol l−1) and was gassed with 95 % O2 and 5 % CO2. One hour after cutting, the slices were transferred to a recording chamber, submerged and perfused with ACSF at 2-4 ml min−1. Most experiments were conducted at 30°C. The experimental procedures were approved by the Animal Care and Use Committee (ACUC) at the University of Texas Medical Branch.

Neurons located in the nuclei lateralis and medialis in the ventral portion of the PAG in coronal slices, and neurons located in the caudal one-third of the lateral PAG in horizontal slices were selected for the study. Membrane currents and potentials were recorded from these ventrolateral PAG neurons using the blind patch-clamp recording technique (Blanton et al. 1989). The composition of the pipette solution was (mM): potassium gluconate, 125; KCl, 4; CaCl2, 0·5; MgCl2, 2·4; BAPTA, 5·0; Hepes, 10; Na2ATP, 5; and GTP-tris salt, 0·33 (pH 7·3; osmolarity, 275-280 mosmol l−1). Synaptic current or potentials were evoked by 100 μs, 2-10 V square pulses delivered at 0·03 Hz with a bipolar concentric stimulating electrode (100 μm in diameter), which was placed 200-800 μm away from the recording site. The exact location of the stimulating electrode was not critical for the determination of opioid action. To prevent K+ and Na+ channel activation in voltage-clamp experiments, the pipette solution contained (mM): Cs2SO4, 110; CaCl2, 0·5; MgCl2, 2·4; BAPTA, 5·0; Hepes, 10; Na2ATP, 5; GTP-tris salt, 0·33; the Na+ channel blocker lidocaine N-ethyl bromide (QX314), 10; and the K+ channel blocker TEA-Cl, 5 (pH 7·3; osmolarity, 275-280 mosmol l−1). The GABAA receptor antagonist bicuculline, the N-methyl-D-aspartate (NMDA) receptor antagonist 2-amino-5-phosphonopentanoic acid (APV), the non-NMDA receptor antagonist 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX), the μ-opioid receptor agonist Tyr-D-Ala-Gly-MePhe-Gly-ol (DAMGO), the non-selective opioid receptor antagonist naloxone and TEA-Cl were purchased from Sigma or RBI. QX314 was a gift from Astra Pharmaceuticals (Westboro, MA, USA). CNQX and bicuculline stock solutions were dissolved in DMSO.

An Axopatch 200A amplifier (Axon Instruments) was used for recordings. Data were digitized at 2-10 kHz with an ITC-16 computer interface (Instrutech, Greatneck, NY, USA) and acquired with the Pulse Control program (kindly supplied by R. J. Bookman, University of Miami, Miami, FL, USA). IGOR (Wavemetrics, Lake Oswego, OR, USA) software was used for data analysis and graphing. A liquid junction potential of -15 mV (for potassium gluconate pipette solution) or of -11 mV (for Cs2SO4 solution) was corrected during off-line analyses. Data are expressed as means ±s.e.m. Student's two-tailed t test was used to evaluate the significance of changes.


DAMGO increases the firing activity of PAG cells

We studied the effect of the μ-opioid receptor agonist DAMGO on synaptic transmission in the PAG in horizontal slices. Under current-clamp conditions, postsynaptic potentials (PSPs, i.e. the sum of EPSPs and IPSPs) and the action potential activity of neurons were monitored in 11 cells before and after superfusion of DAMGO. In the control solution, all cells were quiescent and PSPs were evoked when presynaptic neurons were stimulated (hereafter referred to as presynaptic stimulation). One to three action potentials were elicited when PSPs reached threshold (Fig. 1A).

Figure 1
DAMGO increases cell firing of PAG neurons in horizontal slices

Application of DAMGO (1 μM) hyperpolarized membrane potentials in 8 of the 11 cells. The hyperpolarization ranged from 3 to 15 mV; the mean hyperpolarization was 8·3 ± 1·0 mV (n= 8 cells). In the remaining three cells, no change in resting membrane potential or in synaptic response was observed. Following DAMGO treatment, the PSPs in one cell became subthreshold and the action potential was suppressed. In the other seven cells, PSPs in DAMGO were 7·2 ± 1·7 mV more depolarized than those in the control. In 3 out of 7 cells, the PSPs were suprathreshold and cell firing was increased. In the cell shown (Fig. 1), the resting potential was at -60 mV in the control. Presynaptic stimuli evoked one spike. After application of DAMGO, the membrane resting potential was hyperpolarized to -66 mV. A burst of action potentials was generated in response to the same stimuli.

We repeated the same experiments on ventrolateral PAG neurons in transverse slices. Of the 14 cells tested, only one cell fired spontaneously and the rest were quiescent under control conditions. PSPs were evoked in all the quiescent cells in response to presynaptic stimulation. DAMGO (1 μM) produced hyperpolarization (7·3 ± 0·9 mV, n= 11) in 11 of the 14 cells. Evoked PSPs were more depolarized in 10 of the 11 cells; cell firing was enhanced in seven of them (Figs 2 and and3).3). In the one spontaneously active cell, although the evoked PSP had a larger amplitude, the firing activity was inhibited by DAMGO as the result of hyperpolarization (data not shown).

Figure 2
DAMGO also increases the cell activity in transverse slices
Figure 3
DAMGO caused repetitive firing of a PAG neuron in a transverse slice

The cell illustrated in Fig. 2 had a resting potential of -52 mV in the control. Focal stimulation did not elicit any action potentials. In DAMGO, the resting potential hyperpolarized to -65 mV. Despite the hyperpolarization, stimuli of the same strength triggered action potentials. The firing was inhibited when both DAMGO and the non-selective opioid receptor antagonist naloxone were applied to the cell. In another example (Fig. 3), the cell had a resting potential of -61 mV in control conditions. Synaptic stimulation evoked an initial spike and two spikes at a later time (Fig. 3A). After DAMGO treatment, the cell was hyperpolarized to -64 mV (Fig. 3B). The same stimulus elicited a burst of action potentials. The firing threshold was raised from -41 to -37 mV in this cell. However, the mean threshold change by DAMGO in all responding cells was not significant (< 1 mV). Thus, in both horizontal and transverse slices, DAMGO facilitated synaptic transmission and increased the evoked action potential activity in many PAG cells despite the fact that the opioid produced hyperpolarization in these cells.

Mechanism underlying the excitatory action of DAMGO

One possible mechanism for the increased neuronal activity is a removal of Na+ channel inactivation as a result of DAMGO-induced hyperpolarization. To test this, we determined the effects of hyperpolarization on cell excitability with and without DAMGO. In the control solution, cells were hyperpolarized 3-10 mV from their resting potentials. None of the cells (n= 4) responded with an increase in firing during presynaptic stimulation (data not shown). We then compared the effect of DAMGO on cell firing activity evoked by the same strength stimuli with and without hyperpolarization (Figs 3B and C). After recording the DAMGO effects, the membrane potential was re-depolarized to the control level (Fig. 3C). Fewer action potentials were evoked. Nevertheless, compared with the number of action potentials in the control, the firing activity was still enhanced, i.e. the action potential number increased from three to five. The same experimental protocol was repeated in another four cells. In two cells, results similar to those shown in Fig. 3C were observed. In the other two cells, membrane re-depolarization did not reduce cell firing in the presence of DAMGO; rather it further increased firing activity. Thus, removal of Na+ channel inactivation may enhance cell activity in some cases, but it cannot account for the full magnitude of the excitation produced by DAMGO.

Another possible mechanism for the excitatory effect of DAMGO is a block of GABAergic inputs. To test this, we examined the effect of the GABAA receptor antagonist bicuculline in these cells. Bicuculline (20 μM) caused repetitive cell firing during presynaptic stimulation, mimicking the action of DAMGO (Fig. 3D). Since free base bicuculline, not the N-methyl derivative of bicuculline, was used in these experiments, the increase in cell firing could not have resulted from a bicuculline-induced block of apamine-sensitive Ca2+-activated K+ channels (Johnson & Seutin, 1997). In contrast to the hyperpolarization produced by DAMGO, membrane potentials were depolarized in bicuculline. Thus, PAG neurons appear to be tonically blocked by GABAergic inputs in our slices. Bicuculline relieves the block, causing an increase in cell firing.

In addition to reducing IPSPs and causing hyperpolarization, the μ-opioid also inhibited EPSPs in PAG cells. To understand the role of opioid inhibition of EPSPs in cell activity, we studied the effects of DAMGO on EPSCs and IPSCs under voltage-clamp conditions. The total synaptic currents evoked by focal stimulation were examined first at different holding potentials (Fig. 4). In the control solution, the evoked synaptic currents were inward at hyperpolarized membrane potentials. The currents reversed their direction and became outward when the potentials were held at values more positive than -50 mV (Fig. 4A, left). To determine the contribution of EPSCs and IPSCs to the total synaptic currents, we treated the cell first with the non-NMDA receptor antagonist CNQX (10 μM). The inward synaptic currents were reduced and the outward currents increased (Fig. 4A, middle). When both CNQX and bicuculline (20 μM) were subsequently applied to the same cell, all synaptic currents were blocked (Fig. 4A, right). The contribution of NMDA receptor-mediated currents to the total synaptic current was minimal in this cell. Thus, the currents following CNQX treatment (Fig. 4A, middle) were going through bicuculline-sensitive GABAA receptor channels. The currents going through non-NMDA receptor channels, i.e. the CNQX-sensitive currents, were obtained by subtracting currents in the presence of CNQX from currents in the control solution. GABA receptor-mediated IPSCs were observed in 96 % of cells tested, while non-NMDA receptor-mediated EPSCs were in 84 % of cells.

Figure 4
Evoked synaptic responses in a ventrolateral PAG neuron under voltage-clamp conditions

The voltage dependence of the CNQX-sensitive and bicuculline-sensitive currents is shown in Fig. 4B. The CNQX-sensitive currents, i.e. EPSCs, for this cell reversed at -10 mV. The bicuculline-sensitive currents, i.e. IPSCs, reversed at -70 mV. Similar reversal potentials were obtained in four other cells for EPSCs and in six other cells for IPSCs. In cells with both EPSCs and IPSCs present, like the one shown here, the total synaptic currents reversed at approximately -55 mV (Fig. 4B). The contribution of the EPSCs to the total synaptic current was larger at membrane potentials more hyperpolarized than -55 mV. The EPSCs and IPSCs at -55 mV had approximately equal magnitudes. The contribution of the IPSCs started to dominate as the membrane potential became more depolarized than -55 mV.

In 87 % of cells, slow EPSCs, in addition to fast EPSCs, could be seen following bicuculline treatment (data not shown). The currents could be blocked by APV. In six cells where the I-V relationships were examined, the amplitude of the slow EPSCs depended on the membrane potential (11 ± 2 pA at -70 mV, 35 ± 3 pA at -20 mV). The I-V curves had a characteristic negative-conductance region, suggesting that the currents were NMDA receptor mediated.

The relative contributions of EPSCs and IPSCs to synaptic responses were determined next. To isolate EPSCs and IPSCs and keep the same driving force for the synaptic currents, we evoked EPSCs at the reversal potential of typical IPSCs (-70 mV) and evoked IPSCs at the reversal potential of typical EPSCs (-10 mV). In the cell shown in Fig. 5, both EPSCs and IPSCs were evoked. At -10 mV, electrical stimuli evoked an IPSC with a peak amplitude of 440 pA (Fig. 5A). By contrast, the peak amplitude of the EPSC evoked at -70 mV was 100 pA (Fig. 5B), which is much smaller than that of the IPSC. Both EPSCs and IPSCs could be evoked in 14 of 39 cells studied. The mean IPSC evoked at -10 mV (191 ± 32 pA, n= 14) was 2·3 times larger than the mean EPSC evoked at -70 mV (85 ± 15 pA, n= 14) (P < 0·01, paired t test). When we included the other 25 cells where only IPSCs or only EPSCs could be evoked, the mean amplitude of IPSCs at -10 mV was 186 ± 23 pA (n= 26) and the mean amplitude of EPSCs at -70 mV was 81 ± 8 pA (n= 27). Thus, GABAergic inputs exert a larger influence on the synaptic responses of ventrolateral PAG neurons than glutamatergic inputs.

Figure 5
Differential effects of DAMGO on EPSCs and IPSCs

We then compared the effects of DAMGO on EPSCs and IPSCs. DAMGO inhibited both evoked IPSCs and evoked EPSCs. In the presence of DAMGO, the IPSC was considerably reduced (66 % inhibition) (Fig. 5Aa). Some stimulus trials failed to evoke any substantial IPSC, as though the firing of GABAergic neurons synapsing onto the recorded cell was blocked (Fig. 5Ab). In contrast, DAMGO inhibited the EPSCs by only a moderate amount (17 %) (Fig. 5B). The effects of DAMGO on both EPSCs and IPSCs could be reversed by naloxone (1 μM). We studied the DAMGO effects on IPSCs and EPSCs in detail in a total of 14 cells. In two cells, DAMGO inhibited IPSCs, but increased EPSCs. In the other 12 cells, DAMGO inhibited both IPSCs and EPSCs. In 11 of the 12 cells, DAMGO exerted a larger inhibitory effect on IPSCs than on EPSCs. If we considered only cells in which DAMGO reduced both IPSCs and EPSCs, the mean percentage inhibition of IPSCs was 60·7 ± 4·0 % (n= 12) and of EPSCs was 35·3 ± 4·2 % (n= 12). Thus, DAMGO inhibits IPSCs to a greater extent than it inhibits EPSCs (P < 0·01, paired t test). The effect of DAMGO did not depend on membrane potential. That is, the same differential block of IPSCs and EPSCs by DAMGO was found when the membrane potential was held at rest (i.e. -52 to -75 mV) (data not shown). Thus, DAMGO indeed exerts a stronger effect on IPSCs than on EPSCs.


We show here that DAMGO evoked cell firing in some PAG cells despite the fact that DAMGO depresses EPSCs and activates K+ currents. Since the excitatory effect of DAMGO was observed in both transverse and horizontal slices, the effect is not likely to arise from biased fibre inputs in different planes of sections. Furthermore, removal of Na+ channel inactivation as a result of opioid-induced hyperpolarization cannot explain the increase in cell activity because re-depolarizing the membrane potential near rest did not block cell firing.

Under voltage-clamp conditions and with the same driving force, the evoked IPSCs were more than double the size of the evoked EPSCs (Fig. 5). Furthermore, DAMGO exerted a larger influence on IPSCs than on EPSCs. The frequent failure of evoked IPSCs in the presence of DAMGO (Fig. 5A) and the prevalence of somatic and dendritic opioid receptors on GABAergic PAG neurons (Kalyuzhny & Wessendorf, 1998) favour the idea that the opioid hyperpolarizes GABAergic neurons presynaptic to the recorded cells. In addition, PAG cells in this region are tonically inhibited by their GABAergic inputs (Fig. 3) (Behbehani et al. 1990b). We suggest that the stronger inhibition of the large IPSCs by DAMGO relieves the inhibitory drive of GABAergic inputs and overcomes the DAMGO-induced hyperpolarization and inhibition of EPSCs. This brings about an increase in the neuronal activity of PAG cells.

Our findings agree partly with the results of Vaughan & Christie (1997). These investigators have studied the effect of μ-opioids on caudal PAG neurons in thin (< 250 μm) horizontal slices obtained from Sprague-Dawley rats. With CsCl (140 mM) as the major component of their pipette solution, the reversal potential of IPSCs (~0 mV) was close to that of EPSCs. Measuring synaptic currents at -74 mV, they found that the mean amplitude of IPSCs was 2·5 times larger than that of non-NMDA receptor-mediated EPSCs, a result consistent with ours. In addition, they found that DAMGO reduced the amplitude of IPSCs by 60 %, and reduced non-NMDA receptor-mediated EPSCs by 62 % and NMDA receptor-mediated EPSCs by 43 %. We observed a similar inhibition (60·7 %) of IPSCs by the μ-opioid, but a smaller (35·3 %) inhibition of EPSCs. The reason for the difference in the opioid inhibition of EPSCs has yet to be identified. Aside from the reduction of synaptic inputs in thin slices and variations in rat strain that might contribute to the difference, the experimental conditions used in the two studies were not identical. To isolate IPSPs and EPSCs, Vaughan & Christie (1997) added antagonists, i.e. bicuculline methiodide, APV or CNQX, to the bath solution. We, on the other hand, separated IPSCs and EPSCs by using their respective reversal potentials in lieu of the antagonists. The EPSCs we measured, therefore, contained both non-NMDA and NMDA components. Although the NMDA receptor-mediated EPSCs measured at -70 mV and normal concentrations of Mg2+ were small, they were not negligible. Considering the large difference between the opioid inhibition of non-NMDA and of NMDA EPSCs in Vaughan & Christie's data, the discrepancy between the two studies may partly be attributed to the variation in the contribution of NMDA EPSCs to overall EPSCs. Under current-clamp conditions, the differential effects of the μ-opioid on EPSPs and IPSPs would be more pronounced as more NMDA receptors activated at depolarizing potentials.

The major advantages of our approach were that the effects of μ-opioids were examined under more physiological conditions and non-specific effects of the antagonists on cell excitability and synaptic responses were avoided. The disadvantage of our approach was that the separation of EPSCs and IPSCs might not be complete. The synaptic currents at either -10 or -70 mV could contain a mixture of IPSCs and EPSCs. Nevertheless, the EPSCs at -10 mV would not introduce significant errors in the measurement of IPSCs because of the fewer glutamatergic than GABAergic inputs in our cells. The contribution of IPSCs to the measured EPSCs at -70 mV, if it existed, would cause an overestimation of the opioid block of EPSCs. Thus, the complication is unlikely to change the conclusion that the μ-opioid exerts differential effects on EPSCs and IPSCs.

The basis for the differential opioid block has yet to be determined. The location of opioid receptors and their coupling to voltage-dependent channels are likely to be important. Anatomical studies clearly show that enkephalinergic axons make axo-dendritic and axo-somatic synaptic contacts with GABAergic neurons in the PAG (Wang et al. 1994). μ-Opioid receptors are frequently located on the somata or dendrites of GABAergic PAG neurons (Kalyuzhny & Wessendorf, 1998). It remains to be determined whether μ-opioid receptors on glutamatergic neurons are as prevalent. In addition, the influence of opioid receptors at the axon terminals should not be ignored (Cohen et al. 1992; Huang, 1995). Although axo-axonic synapses between opioidergic and GABAergic cells in the PAG are rarely observed (Williams & Beitz, 1990b), μ-opioids have been found to reduce GABA and glutamate release (Vaughan & Christie, 1997). The reduction of GABA release is believed to result from the coupling of μ-opioid receptors to a presynaptic K+ channel through a phospholipase A2 pathway (Vaughan et al. 1997). The mechanism for the reduction of glutamate release by opioids has not been determined. Since 4-aminopyridine blocks the opioid inhibition of GABA receptor-mediated miniature IPSCs (mIPSCs), but not the inhibition of glutamate receptor-mediated mEPSCs (Vaughan et al. 1997), presynaptic opioid receptors in glutamatergic neurons appear to couple to different channels. If opioid receptors at presynaptic terminals play a role in determining neuronal activity, the weak opioid inhibition of EPSCs in the ventrolateral PAG could result, in part, from a paucity of presynaptic opioid receptors or a lack of presynaptic ion channels that couple with opioid receptors in the axon terminals of glutamatergic neurons.

We do not know if our recorded cells are output neurons that project to the NRM. Behbehani et al. (1990b) recorded from a large number of cells extracellularly in caudal PAG slices. These cells were mostly spontaneously active and were inhibited by enkephalin. Only a few of the cells that we recorded from had these properties. Most of our cells were quiescent and their firing was enhanced by DAMGO. If the spontaneously active cells are inhibitory interneurons as suggested (Behbehani et al. 1990a), they were seldom sampled in our study. Interneurons are usually small. It is conceivable that a smaller cell size contributes to our sample bias. It would be of interest to determine the percentage of our cells that project to the NRM using retrograde labelling techniques (Huang, 1989; Osborne et al. 1996). Irrespective of the answer, our results emphasize the fact that a direct hyperpolarization or a reduction of EPSPs by an opioid does not necessarily imply that the opioid inhibits cell activity. The overall effect of an opioid depends critically on the dynamic balance of its excitatory and inhibitory actions.

We and others have shown that opioid receptors are not located solely on inhibitory neurons (Fig. 5) (Vaughan & Christie, 1997). The inhibition of EPSCs and IPSCs by opioids in the same cell seems paradoxical because these actions affect cell firing in opposite directions. However, the observation may not be as puzzling if one considers the possibility that an opioid does not necessarily exert its inhibitory effects on GABAergic and glutamatergic neurons in concert in vivo. If this is the case, the effects of opioid modulation would depend on the activity of the cells. During quiescence or a low cell activity period when GABAergic inputs exert dominant influences, opioids would assist the disinhibition process. During high activity conditions when glutamatergic inputs dominate, opioids would reduce the excitation of cells. The results presented here show that an opioid can excite PAG cells even though opioids modulate inhibitory and excitatory neurons simultaneously. Under these conditions, the inhibition of glutamatergic responses could serve as a brake to curb cells from overexcitation. Irrespective of whether or not opioids affect GABAergic and glutamaterigic cells simultaneously, the dual effects on cell firing provide an opioid with wide and multi-level control over the activity of PAG neurons. The flexibility allows opioid actions to change with the stimulating conditions, or with the state of the neurons, thus providing a dynamic control of PAG function. The same principle of opioid actions may well be used in other regions of the central nervous system.


The authors thank Dr W. D. Willis for his comments on the manuscript. This work was supported by NIH grants (NS30045 and NS23061) and the Human Frontier Science Program (RG0073) to L.-Y. M. Huang. L. C. Chiou was supported by a grant for overseas study and a research grant (NSC 87-2314-B002-314) from National Science Council, ROC.


  • Basbaum AI, Fields HL. Endogenous pain control systems: brainstem spinal pathways and endorphin circuitry. Annual Review of Neuroscience. 1984;7:309–338. [PubMed]
  • Behbehhani MM. Functional characteristics of the midbrain periaqueductual gray. Progress in Neurobiology. 1995;46:575–605. [PubMed]
  • Behbehani MM, Jiang M, Chandler SD. The effect of [met]enkephalin on the periaqueductal gray neurons of the rat: an in vitro study. Neuroscience. 1990a;38:373–380. [PubMed]
  • Behbehani MM, Jiang M, Chandler SD, Ennis M. The effect of GABA and its antagonists on midbrain periaqueductal gray neurons in the rat. Pain. 1990b;40:195–204. [PubMed]
  • Beitz A. The midbrain periaqueductal gray in the rat. I. Nuclear volume, cell number, density, orientation and regional subdivisions. Journal of Comparative Neurology. 1985;237:445–459. [PubMed]
  • Besson J-M, Fardin V, Olivéras J-L. The nociceptive modulatory effects of periaqueductal gray activation are mediated by two neuronal classes in the rostral ventromedial medulla. In: Depaulis A, Bandler R, editors. The Midbrain Periaqueductal Gray Matter. New York: Plenum Press; 1991. pp. 121–138.
  • Blanton MG, Lo Turco JJ, Kriegstein AR. Whole cell recording from neurons in slices of reptilian and mammalian cerebral cortex. Journal of Neuroscience Methods. 1989;30:203–210. [PubMed]
  • Castiglioni AJ, Gallaway MC, Coulter JD. Spinal projections from the midbrain in monkey. Journal of Comparative Neurology. 1978;178:329–346. [PubMed]
  • Cohen GA, Doze VA, Madison DV. Opioid inhibition of GABA release from presynaptic terminals of rat hippocampal interneurons. Neuron. 1992;9:325–335. [PubMed]
  • Depaulis A, Morgan MM, Liebeskind JC. GABAergic modulation of the analgesic effects of morphine microinjected in the ventral periaqueductal gray matter of the rat. Brain Research. 1987;436:223–228. [PubMed]
  • Fields HL, Heinricher MM, Mason P. Neurotransmitters in nociceptive modulatory circuits. Annual Review of Neuroscience. 1991;14:219–245. [PubMed]
  • Huang L-Y M. Calcium channels in isolated rat dorsal horn neurones including labelled spinothalamic and trigeminothalamic cells. The Journal of Physiology. 1989;411:161–177. [PMC free article] [PubMed]
  • Huang L-Y M. Cellular mechanisms of excitatory and inhibitory actions of opioids. In: Tseng LF, editor. The Pharmacology of Opioid Peptides. Amsterdam: Hardwood Academic Publishers GmbH; 1995. pp. 131–149.
  • Johnson SW, Seutin V. Bicuculline methiodide potentiates NMDA-dependent burst firing in rat dopamine neurons by blocking apamin-sensitive Ca2+-activated K+ currents. Neuroscience Letters. 1997;231:13–16. [PubMed]
  • Kalyuzhny AE, Arvidsson U, Wu W, Wessendorf MW. Mu-opioid and delta-opioid receptors are expressed in brainstem antinociceptive circuits - studies using immunocytochemistry and retrograde tract-tracing. Journal of Neuroscience. 1996;16:6490–6503. [PubMed]
  • Kalyuzhny AE, Wessendorf MW. Relationship of mu- and delta-opioid receptors to GABAergic neurons in the central nervous system, including antinociceptive brainstem circuits. Journal of Comparative Neurology. 1998;392:528–547. [PubMed]
  • Kuypers HGJM, Maisky VA. Retrograde axonal transport of horseradish peroxidase from spinal cord to brainstem cell groups in cats. Neuroscience Letters. 1975;1:9–14. [PubMed]
  • Mantyh PW, Peschanski M. Spinal projections from the periaqueductal gray and dorsal raphe in the rat, cat and monkey. Neuroscience. 1982;7:2769–2776. [PubMed]
  • Moises HC, Rusin KI, Macdonald RL. Mu- and kappa-opioid receptors selectively reduce the same transient components of high-threshold calcium current in rat dorsal root ganglion sensory neurons. Journal of Neuroscience. 1994;14:5903–5916. [PubMed]
  • Morgan MM. Differences in antinociception evoked from dorsal and ventral regions of the caudal periaqueductal gray matter. In: Depaulis A, Bandler R, editors. The Midbrain Periaqueductal Gray Matter. New York: Plenum Press; 1991. pp. 139–150.
  • Osborne PB, Vaughan CW, Wilson HI, Christie MJ. Opioid inhibition of rat periaqueductal grey neurones with identified projections to rostral ventromedial medulla in vitro. The Journal of Physiology. 1996;490:383–389. [PMC free article] [PubMed]
  • Reichling DB, Basbaum AI. Contribution of brainstem GABAergic circuitry to descending antinociceptive controls: II. Electron microscopic immunocytochemical evidence of GABAergic control over the projection from the periaqueductal gray to the nucleus raphe magnus in the rat. Journal of Comparative Neurology. 1990;302:378–393. [PubMed]
  • Reichling DB, Kwiat GC, Basbaum AI. Anatomy, physiology and pharmacology of the periaqueductal gray contribution to antinociceptive controls. Progress in Brain Research. 1988;77:31–46. [PubMed]
  • Schroeder JE, Fischbach PS, Zheng D, McCleskey EW. Activation of μ opioid receptors inhibits transient high- and low-threshold Ca2+ currents, but spares a sustained current. Neuron. 1991;6:13–20. [PubMed]
  • Vaughan CW, Christie MJ. Presynaptic inhibitory action of opioids on synaptic transmission in the rat periaqueductal grey in vitro. The Journal of Physiology. 1997;498:463–472. [PMC free article] [PubMed]
  • Vaughan CW, Ingram SL, Connor MA, Christie MJ. How opioids inhibit GABA-mediated neurotransmission. Nature. 1997;390:611–614. [PubMed]
  • Wang Q-P, Guan J-L, Nakai Y. Immunoelectron microscopy of enkephalinergic innervation of GABAergic neurons in the periaqueductal gray. Brain Research. 1994;665:39–46. [PubMed]
  • William JT, Egan TM, North RA. Enkephalin opens potassium channels on mammalian central neurones. Nature. 1982;299:74–77. [PubMed]
  • Williams FG, Beitz AJ. Ultrastructural morphometric analysis of enkephalin-immunoreactive terminals in the ventrocaudal periaqueductal gray: Analysis of their relationship to periaqueductal gray-raphe magnus projection neurons. Neuroscience. 1990a;38:381–394. [PubMed]
  • Williams FG, Beitz AJ. Ultrastructural morphometric analysis of GABA-immunoreactive terminals in the ventrocaudal periaqueductal grey: analysis of the relationship of GABA terminals and the GABAA receptor to periaqueductal grey-raphe magnus projection neurons. Journal of Neurocytology. 1990b;19:686–696. [PubMed]
  • Yaksh TL, Al-Rodhan NRF, Jensen TS. Sites of action of opiates in production of analgesia. Progress in Brain Research. 1988;77:371–384. [PubMed]
  • Yaksh TL, Yeung JC, Rudy TA. Systematic examination in the rat of brain sites sensitive to the direct application of morphine: observation of differential effects within the periaquductal gray. Brain Research. 1976;114:83–103. [PubMed]

Articles from The Journal of Physiology are provided here courtesy of The Physiological Society
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...