Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

The mechanisms of tolerance and dependence to morphine have been studied extensively, in part due to the negative effects of these phenomena on long-term opiate analgesia. Despite the opioid receptor down-regulation in response to chronic morphine exposure in vitro and in vivo (Morris and Herz, 1989; Ronnekleiv et al., 1996), there is a strong up-regulation of the cAMP system (Nestler et al., 1993). In addition to this mechanism, it also has been proposed that nitric oxide and the protein kinase C (PKC) pathways might also play a role. Thus, Mayer et al. (1995) recently showed that the development of tolerance to the analgesic effects of morphine was markedly reduced by agents blocking PKC translocation. Similarly, n(G)-nitro-L-arginine methylester, a nitric oxide synthase inhibitor, reduced the development of morphine tolerance and dependence (Majeed et al., 1994). Recently, Chen and Huang (1991) showed that the selective µ-opiate receptor agonist [D-Ala²-N-Me-Phe⁴,Gly-ol⁵]-enkephalin increased the amplitude of N-methyl-D-aspartate (NMDA) currents in cultured spinal cord cells, via activation of a PKC pathway that decreased Mg⁺⁺ blockade.

It is now thought that the NMDA glutamate receptor subtype plays a key role in opioid tolerance and dependence. Thus, the NMDA receptor antagonist MK801 reduces some aspects of these phenomena (Herman et al., 1995; Marek et al., 1991; Trujillo and Akil, 1991, 1995). In addition, recent studies (Fundytus et al., 1995; Fundytus and Coderre, 1997a, b) have pointed to the role of metabotropic glutamate receptors (mGluRs) by showing that chronic i.c.v. administration of -methly-4-carboxyphenylglycine, a specific antagonist of group 1 and 2 mGluRs, markedly attenuated abstinence syndromes. This effect is particularly interesting, because it has been now clearly established that mGluR agonists like trans-1-aminocyclopentane-1,3-decarboxylic acid (trans-ACPD) and L(+)-2-amino-4-phosphonobutyric acid (L-AP4) strongly inhibit NMDA receptor function (Pin and Duvoisin, 1995), including that in neurons of the nucleus accumbens (NAcc) (Martin et al., 1997a). This brain region is thought to play a major role in opiate dependence and has an abundance of opiate peptides and receptors (Ding et al., 1996; Gracy et al., 1997; Svingos et al., 1996). Therefore, we used intracellular recording in a slice preparation of NAcc to examine the effect of chronic morphine on the interaction between NMDA and mGluR receptors as a possible cellular underpinning of morphine tolerance and dependence. We now report that chronic morphine treatment selectively alters the regulation of glutamate release via presynaptic mGluRs. Thus, the data show that the presynaptic inhibitory effect of trans-ACPD on the NMDA current is significantly enhanced, whereas the postsynaptic inhibition mediated by L-AP4-sensitive receptors is comparable with that observed in control rats.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Animal and Slice Preparations. We used male Sprague-Dawley rats (100 and 170 g) to prepare accumbens slices from fresh brain tissue, as described previously (Yuan et al., 1992; Nie et al., 1993). We rapidly removed the brain and transferred it to a cold (4°C) and oxygenated artificial cerebrospinal fluid (ACSF) of the following composition (in mM): NaCl, 130; KCl, 3.5; NaH₂PO₄, 1.25; MgSO₄·7H₂O, 1.5; CaCl₂, 2; NaHCO₃, 24; glucose, 10. When the slices were obtained from morphine-treated but not control or placebo-treated rats, 1 µM morphine was added to the ACSF to prevent withdrawal. We glued a tissue block containing NAcc to a Teflon chuck and cut it transversely with a Vibroslice cutter (Campden Instrument, England) and immediately incubated the slices (400 µM thick) in the recording chamber. During initial incubation in an "interface" configuration, the tops of the slices were exposed to a mixture of oxygen (95%) and CO₂ (5%). After 30 min we submerged and superfused the slices with warm (34°C) carbogenated ACSF, at a rate of 3 to 4 ml/min.

Recording. We pulled sharp microelectrodes from borosilicate capillary glass (outer diameter, 1.2 mm; inner diameter, 0.8 mm) on a Brown-Flaming puller (Sutter Instruments, Novato, CA) and filled them with 3 M KCl. Tip resistances were 60 to 100 M. We used an Axoclamp 2B amplifier (Axon Instruments, Foster City, CA) to record neurons in voltage-clamp mode. Throughout all experiments we continuously monitored electrode settling time and capacitance neutralization (on a separate oscilloscope) in discontinuous single-electrode voltage-clamp studies. We recorded neurons within the core NAcc just ventrally to the anterior commisure. Current and voltage levels were monitored and stored on a polygraph, digitized by a TL-1 interface (Axon Instruments) and stored on a 486 personal computer using Clampex 6.0 software (Axon Instruments). We then analyzed the digitized records with Axograph software (Axon Instruments). To generate current-voltage (I-V) curves, we used standard voltage-clamp methods with a 400-ms step duration and measured the current before the step and at a "steady state" 380 ms after the onset of the current pulse. The first voltage step was 20 mV below the holding potential (80 mV) with an increment of +10 mV for the five subsequent steps.

Synaptic Stimulation. We studied the NMDA component of EPSCs (NMDA-EPSCs) in voltage-clamp mode, using the I-V protocol (400-ms step duration) described above to measure EPSC amplitudes evoked at different membrane potentials. The NMDA-EPSCs were elicited by local ("focal" or "proximal") stimulation triggered 100 ms after the onset of, and therefore superimposed upon, the voltage step. We averaged two traces at the same voltage-step size with their superimposed NMDA-EPSCs. To minimize the influence of the stimulation artifact on the NMDA current, we neutralized it by injecting into the amplifier blanking circuit a 2-ms duration pulse 1 ms before and after the stimulation.

Drug Administration. To pharmacologically isolate the NMDA-EPSC component, we superfused the slices with antagonists specific for non-NMDA (kainate/-amino-3-hydroxy-5-methly-4-isoxazolepropionic acid) receptors [10 µM 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX)] and -aminobutyric acid-A receptors (15 µM bicuculline), for at least one-half hour before recording. Our standard drug-testing protocol was as follows: after recording a stable membrane potential and NMDA receptor-evoked events for at least 15 min, we superfused the slices with the ACSF-antagonist solution described above but containing either trans-ACPD (5-10 µM; 15-25 min), quisqualate (1 µM), or L-AP4 (20 µM). Each cell served as its own control. The superfusion system allowed switching of drug-ACSF solutions without disrupting the rapid flow of the ACSF through the recording chamber (Moore et al., 1994).

Chronic Morphine Treatment. Morphine pellets (75 mg of base) and placebo pellets were provided by the National Institute on Drug Abuse. Two pellets (either morphine or placebo) were implanted s.c. in the neck in each rat under light halothane anesthesia. All electrophysiological testing was done 3 to 6 days after pellet implantation. This treatment was based on the model designed by Gold et al. (1994), who showed that it led to tolerance 12 h postimplantation, that plasma morphine levels remained constant for several days, and that withdrawal symptoms could be observed up to 13 days after implantation.

Statistical Analyses. We expressed all grouped data as mean ± S.E.M. Statistical significance between control, drug, and washout within each group of cells was analyzed with one-factor analysis of variance for repeated measures, with a post hoc analysis by Newman-Keul's or Fisher's PLSD comparison tests. We analyzed differences (expressed as percentage of control) within and between groups of cells from untreated, morphine-treated, and sham-operated rats by one-way analysis of variance between subjects. We considered p values of less than .05 statistically significant.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Effect of trans-ACPD on NMDA-EP. As we previously reported, the group 1 and 2 metabotropic agonist trans-ACPD had no effect on membrane potential or input conductance, but markedly depressed NMDA-EPSCs in NAcc neurons (Martin et al., 1997b). Figure 1 shows the inhibition of NMDA-EPSCs, evoked by local stimulation in the presence of 10 µM CNQX and 15 µM bicuculline, recorded at holding potentials around 60 mV in slices obtained from untreated and morphine-treated rats. As we previously showed, the amplitude and duration of this pharmacologically isolated EPSC component increased as the cell was depolarized, and the EPSCs were nearly completely blocked by 60 µM D-2-amino-5-phosphonovaleic-acid (Martin et al., 1997a,b). Superfusion of 5 µM trans-ACPD weakly decreased the NMDA-EPSC amplitude in untreated rats (Fig, 1A, left panel). However, the same trans-ACPD concentration applied to slices from a rat chronically treated with morphine markedly decreased NMDA-EPSC amplitudes.

View larger version (32K): [in this window] [in a new window]   Fig. 1.   Chronic morphine augments the depressant effect of trans-ACPD on NMDA-EPSC areas. A, Pharmacologically isolated NMDA-EPSCs from different NAcc neurons recorded at hyperpolarized and depolarized potentials in the presence of CNQX (10 µM) and bicuculline (15 µM). These recordings were obtained from slices of untreated (left panel) and morphine-treated rats (right panel). In cells of untreated rats, 5 µM trans-ACPD slightly decreased NMDA-EPSC amplitudes, whereas the depression was more robust after morphine treatment. These effects were followed by an nearly complete recovery on washout (right panel). B, In another neuron, NMDA-EPSC traces recorded before and during 10 µM trans-ACPD superfusion in NAcc neurons from untreated (left panel) and morphine-treated rats (right panel). Note the difference between the depressant effect of trans-ACPD in these two neurons. In this and subsequent figures, vertical arrows indicate time of local pathway stimulation. C, Grouped data; mean effect (±S.E.M.) in percentage of control of 5 µM trans-ACPD on NMDA-EPSC time-integrals in NAcc neurons from untreated (n = 7), morphine-treated (n = 7), and sham-operated (n = 6) rats evoked over the same range of membrane potentials from 90 to 55 mV. D, The mean effect (±S.E.M.) of 10 µM trans-ACPD on NMDA-EPSC areas from a pool of 7, 6, and 7 neurons from untreated, morphine-treated, and sham-operated rats, respectively.

View larger version (26K): [in this window] [in a new window]   Fig. 2.   Dose-response relationships for trans-ACPD reduction of NMDA-EPSC time-integrals, averaged over all tested membrane potentials and plotted as percentage of inhibition of EPSC time-integral versus trans-ACPD concentration (1-10 µM), in a group of untreated (black column), morphine-treated (white column), and sham-operated rats (gray column). Numbers in columns indicate number of neurons.

Effect of Quisqualate and L-AP4 on NMDA-EPSCs. We previously reported that the group 1 metabotropic agonist quisqualate had no effect on NMDA-EPSCs in NAcc neurons (Martin et al., 1997a). In the present study we verified this lack of effect, but also now in NAcc neurons of morphine-treated rats (Fig. 3). We also re-examined the effect of L-AP4 (20 µM) on NMDA-EPSCs. We previously found that this group 3 mGluR agonist markedly depressed, probably at the postsynaptic level, NMDA-EPSCs and currents evoked by exogenous NMDA in NAcc neurons (Martin et al., 1997a). Figure 3A shows NMDA-EPSCs recorded at three membrane potentials from a cell (resting membrane potential = 84 mV) recorded in a slice from the morphine group. As with slices from untreated animals, L-AP4 markedly attenuated NMDA-EPSC time-integrals. This effect, statistically significant compared with controls (F (2,46) = 78.85; p < .001, n = 4), was followed by recovery on washout. Figure 3C shows the mean decrease of NMDA-EPSC time-integrals by L-AP4 over all potentials in four cells each from untreated and morphine-treated rats. Although the inhibitory effect of L-AP4 on NAcc neurons from morphine-treated rats was slightly greater compared with those of the untreated group, the difference was not statistically significant (F(1,41) = 16.69; p = .203). A possible explanation for this absence of significant effect between the two groups is that the inhibition elicited by 20 µM L-AP4 represents a ceiling effect that would prevent the chronic morphine treatment from further decreasing NMDA-EPSCs amplitude. However, this possibility is not likely, because the inhibition mediated by 40 µM L-AP4 in untreated rats was larger than that with 20 µM (data not shown).

View larger version (26K): [in this window] [in a new window]   Fig. 3.   Chronic morphine does not alter the lack of effect of quisqualate (Quis) or the L-AP4-mediated inhibition of NMDA-EPSCs in NAcc. A, NMDA-EPSCs in a NAcc neuron from a chronically treated rat, held at 61 and 99 mV before (control) and during (Quis 1 µM) superfusion of the group 1 mGluR agonist, quisqualate. Note that the two traces in B are virtually indiscriminable. B, Grouped data: the mean effect (±S.E.M.) of 1 µM quisqualate on NMDA-EPSC time-integrals averaged over all potentials as percentage of control from untreated and morphine-treated rats. C, NMDA-EPSCs in a NAcc neuron from a morphine-treated rat, recorded at 56 to 94 mV before (control), during (L-AP4), and after (washout) 20 µM L-AP4 superfusion. D, Group data: the mean effect (±S.E.M.) of 20 µM L-AP4 on NMDA-EPSC time-integrals averaged over all potentials, as percentage of control, from untreated and morphine-treated rats. Numbers in columns indicate numbers of neurons.

Paired-Pulse Facilitation. Our previous studies (Martin et al., 1997a) showed that trans-ACPD probably acts presynaptically to NAcc neurons to decrease glutamate release. Therefore, our present findings suggest that chronic morphine treatment increases the presynaptic inhibitory control of mGluRs on glutamate release in NAcc neurons. To test this hypothesis further, we measured the amplitude of NMDA-EPSCs elicited by a paired-pulse paradigm, often used to discriminate pre- from postsynaptic site of drug action. Figure 4 shows that in untreated rats, an interstimulus interval of 50 ms markedly facilitated the second of two NMDA-EPSC amplitudes; in controls, the secondary NMDA-EPSC amplitudes, measured at 60 mV, were facilitated significantly to 153 ± 5.1% of control (p < .001). This effect disappeared at an interval of 200 ms, because the NMDA-EPSC amplitude was only 95 ± 4.2% of control. Interestingly, a paired-pulse inhibition occurred in controls at an interval of 400 ms, because the amplitude of the synaptic response was attenuated by 21 ± 11%. After chronic morphine treatment, the paired-pulse facilitation observed at a 50-ms interstimulus interval was significantly enhanced: the mean amplitude of the secondary NMDA-EPSCs was 174 ± 6% of control (F(1,19) = 6.88; p < .016). A similar upward shift of the secondary response in the morphine group was also observed at 200 ms, because the response was significantly boosted to 123 ± 6% of control (F(1,19) = 14,94; p < .0012). However, the paired-pulse inhibition at 400 ms was not significantly altered between the untreated and morphine groups (F(1,16) = 1.87; p = .19).

View larger version (25K): [in this window] [in a new window]   Fig. 4.   Paired-pulse studies show enhanced paired-pulse facilitation after chronic morphine. A, NMDA-EPSCs recorded at 60 mV from NAcc neurons of an untreated (Control; upper traces) and morphine-treated (MTR) rat (lower traces) at various interstimulus intervals. Note that the shorter interstimulus interval of 50 ms elicits a strong potentiation of the amplitude of the second NMDA-EPSC, whereas a longer interval of 200 ms causes paired-pulse facilitation only in the morphine-treated cells. An interval of 400 ms causes a small paired-pulse inhibition. B, Group data: the mean secondary NMDA-EPSC amplitudes (expressed as percentage of the primary EPSC amplitude) recorded at 50-, 200-, and 400-ms interstimulus intervals in untreated and morphine-treated rats, showing a significant increase in paired-pulse facilitation after chronic morphine. * p < .016; ** p < .0012. Numbers in parentheses indicate number of cells.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

In previous studies carried out on NAcc slices of naive untreated rats, we reported that NMDA receptor-mediated neurotransmission was under the control of both µ-opioid and metabotropic glutamate receptors (Martin et al., 1997a, b). More specifically, we found evidence that glutamate release was strongly attenuated by agonists selective for µ-opiate receptors and group 2 mGluR receptors located presynaptically. In contrast, the receptors expressed at the postsynaptic level have a different pattern of effects: a µ-opiate agonist enhanced the current evoked by exogenous NMDA, whereas L-AP4, a specific agonist of the group 3 mGluR subtype, decreased it. In the present study, we found that chronic treatment with morphine, known to act on µ- opiate receptors, up-regulated the presumed presynaptic metabotropic depression of NMDA-EPSCs in NAcc neurons but had no influence on the postsynaptic depression of these EPSCs by L-AP4.

Chronic morphine also decreased paired-pulse facilitation of these EPSCs. In the past, paired-pulse facilitation of synaptic potentials has been used as an indirect test of the pre- or postsynaptic site of drug action. Several reports have suggested that there is an inverse relationship between transmitter release and changes in paired-pulse facilitation (Bonci and Williams, 1997; Mennerick and Zorumski, 1995; Salin et al., 1996). This relationship would suggest that chronic morphine further decreases glutamate release in NAcc, because paired-pulse facilitation of NMDA-EPSCs increased after chronic morphine treatment. Therefore, this finding is consistent with our data suggesting up-regulation of pre- but not postsynaptic metabotropic glutamate receptors that depress NMDA-EPSCs. However, it also should be noted that a recent study has found that the magnitude of paired-pulse facilitation of hippocampal CA1 non-NMDA-EPSCs can also be regulated by postsynaptic mechanisms (Wang and Kelly, 1996, 1997).

Mechanisms Underlying Effects of Chronic Morphine. The question arises as to the mechanisms responsible for the chronic morphine effect. At least two mechanisms could be responsible for our findings. First, it is possible that voltage-dependent Ca⁺⁺ channels (VDCC) are involved, because they are known to control the release of neurotransmitters. In fact it has been reported that group 2 mGluR agonists decrease Ca⁺⁺ currents in isolated cortical neurons (Choi and Lovinger, 1996), cerebellar granule cells (Chavis et al., 1995), and striatal neurons (Stefani et al., 1994). Interestingly, Glaum and Miller (1995) reported that this effect observed in isolated cells could also be exerted at the level of presynaptic terminals. The metabotropic glutamate receptor agonist (2S,3S,4S)--(carboxycyclopropyl)glycine decreased Ca⁺⁺ currents of neurons of the tractus solitarius, and this effect was inhibited by cyclic GMP-dependent protein kinase inhibitors. Interestingly, it has been shown that voltage-dependent Ca⁺⁺ channels are regulated by adenylyl cyclase, the enzyme known to be linked to the group 2 mGluRs (Pin and Duvoisin, 1995).

Relevance to Behavioral Effects of Chronic Opiates. The chronic treatment model used here was reported to lead to dependence and tolerance by 12 h postimplantation and to induce withdrawal symptoms up to 13 days after implantation of the pellets (Gold et al. 1994). Our combined data with mGluR agonists and paired-pulse facilitation suggest that this chronic morphine treatment may reduce glutamate release in the NAcc core. Interestingly, this action should decrease NAcc neuronal excitability. It is noteworthy that a similar decreased excitability should occur in the ventral tegmental area, where -aminobutyric acid release has been found to increase with chronic morphine (Bonci and Williams, 1997). Furthermore, as noted in the introduction, behavioral experiments have suggested that changes in both mGluRs and NMDA receptors might be responsible for some aspects of opiate tolerance and dependence (Trujillo and Akil, 1991; Fundytus and Coderre, 1997; Fundytus et al., 1997). In addition, the NAcc is thought to play a major role in opiate-seeking behavior or the rewarding properties of opiates. To date, the cellular mechanisms responsible for these phenomena remain unknown. The present studies were undertaken to determine whether the interactions of mGluRs with glutamate release and NMDA receptor function might provide a cellular substrate for some aspects of chronic morphine exposure such as tolerance and dependence. Our previous studies provided data suggesting that activation of NMDA receptors is under the complex, balanced control of both pre- and postsynaptic µ-opioid receptors, and we hypothesized that a disruption of this balance occurred with chronic opiate treatment, perhaps underlying the phenomenon of dependence. Because the presynaptic but not the postsynaptic mGluR regulation of NMDA neurotransmission is enhanced with chronic morphine, our present findings further suggest that chronic morphine treatment also may presynaptically unbalance metabotropic regulation of glutamatergic function in the accumbens.