The most direct evidence for the hypothesis presented above comes from studies on the mechanisms underlying the induction of behavioral sensitization. As reviewed above, induction of sensitiza-tion is dependent on glutamate transmission in the VTA. Of course, the induction of LTP and LTD share this requirement for glutamate transmission, which was the basis for early hypotheses for involvement of these phenomena in sensitization (18,66).
Many lines of evidence suggest that sensitization involves an increase in glutamate transmission in the VTA (4). This could occur as a result of an increase in glutamate release, an increase in glutamate receptor number, or an increase in glutamate receptor sensitivity. Recent evidence supports a version of the third hypothesis in which sensitization is accompanied by LTP-like changes that increase in the efficiency of glutamate transmission in the VTA.
LTP is expressed as a potentiation of AMPA receptor transmission (67,68). If the early phase of sensitization is associated with LTP at synapses onto midbrain DA neurons, the DA neurons should exhibit increased responsiveness to AMPA. To test this, single-unit recording studies examined the responsiveness of VTA DA neurons to glutamate agonists either 3 or 14 d after discontinuing repeated administration of cocaine or amphetamine (69,70). When glutamate or AMPA was applied directly to DA cell body regions by microiontophoresis, DA neurons recorded from amphetamine- or cocaine-treated rats showed enhanced excitatory responses compared to neurons recorded from saline controls. No difference was observed in responsiveness to NMDA. The enhanced responsiveness to glutamate and AMPA dissipated after 10-14 d of withdrawal, consistent with a role in induction mechanisms.
Based on these electrophysiological results, we predicted that increased AMPA receptor responsiveness should be detectable in microdialysis experiments as an increase in the ability of AMPA to drive meso-accumbens DA cells and, thus, elicit DA release in the NAc. We tested this using rats treated with the same amphetamine regimen used for electrophysiological studies (or repeated saline) and dual-probe microdialysis. We found that intra-VTA administration of a low dose of AMPA produced significantly greater DA efflux in the ipsilateral NAc of amphetamine-treated rats (71). This augmented response to AMPA was transient, because it was present 3 d, but not 10-14 d, after the last injection. It was specific for AMPA, because intra-VTA NMDA administration produced a trend toward increased NAc DA levels that did not differ between groups. Thus, our microdialysis data are in complete agreement with prior electrophysiological studies using the same amphetamine regimen. Both suggest an LTP-like enhancement of AMPA transmission onto VTA DA neurons during the early phase of drug withdrawal.
Of course, an alternative explanation for enhanced responsiveness to AMPA is an increase in AMPA receptor expression. Indeed, there are reports that GluRl levels in VTA, quantified using Western blots, are increased 16-24 h after discontinuation of repeated cocaine, morphine, ethanol, or stress paradigms (72-74) but not after 3 w withdrawal from cocaine (74). Increased GluRl was not observed in the substantia nigra after repeated treatment with cocaine or morphine (73; the substantia nigra was not examined in stress experiments), but after repeated ethanol, there was a greater increase in GluRl in the substantia nigra than in the VTA (72). In contrast, our own quantitative immunoautoradiographic studies (75) found no change in levels of GluRl immunoreactivity in VTA, substantia nigra, or a transitional area after 16-24 h of withdrawal from repeated amphetamine or cocaine treatment, or after 3 or 14 d of withdrawal from amphetamine; this study used the same amphetamine regimen that resulted in enhanced electrophysiological (70) and neurochemical (71) responsiveness to intra-VTA AMPA at the 3-d withdrawal time and one of the same cocaine regimens used previously (73). Possible reasons for these discrepant results at the protein level have been discussed (75). In contrast, all studies agree that mRNA levels for GluR1 (and GluR2-4) in the VTA are not altered during withdrawal from repeated amphetamine or cocaine (75-77). Taking all data into account, we believe that an overall increase in GluR1 expression in the midbrain is unlikely to account for increased responsiveness of VTA AMPA receptors at short withdrawal times. We propose that more complex mechanisms, related to LTP, are responsible for AMPA receptor plasticity in the VTA. Although mechanisms might involve increased AMPA receptor surface expression, models for this process do not necessary involve increases in total levels of AMPA receptors at a cellular or regional level (Section 3). As a first step toward evaluating this hypothesis, several laboratories have begun to characterize synaptic plasticity in midbrain DA neurons.
Overton et al. (78) recorded from a parasagittal slice preparation containing the substantia nigra and the subthalamic nucleus (STN); the latter sends glutamatergic projections to substantia nigra DA neurons. Tetanic stimulation of the STN region produced long-lasting changes in the amplitude of excitatory postsynaptic currents (EPSPs) in approximately half of the DA neurons evaluated, whereas many of the others showed an immediate but short-lived potentiation of EPSP amplitude (short-term potentiation). No LTP was observed in experiments conducted in the presence of NMDA receptor antagonists. Bonci and Malenka (79) examined the types of plasticity exhibited by both DA and non-DA neurons of the VTA. Many of the non-DA cells contain GABA and project to the forebrain (e.g., ref. 80). Synapses onto DA neurons exhibited paired-pulse depression as well as depression in response to a train of 10 stimuli, whereas non-DA cells displayed facilitation under both conditions. Using a protocol that reliably elicits LTP in hippocampal CA1 pyramidal cells, it was found that DA neurons (but not non-DA neurons) exhibited LTP and that its induction required NMDA but not metabotropic glutamate receptor activation.
Next, two studies showed that DA neurons also exhibit LTD (81,82). Both found that the induction of LTD did not require NMDA receptor or metabotropic glutamate receptor activation. However, it was prevented by the Ca2+ chelator BAPTA (81,82) or by voltage-clamping cells at depolarized potentials that prevent the activation of voltage-dependent Ca2+ channels (82). Conversely, it was induced by driving Ca2+ into the DA neuron with repetitive depolarization; importantly, the LTD induced by this mechanism occluded synaptically driven LTD (81). Together, these results suggest that LTD induction in midbrain DA neurons requires a postsynaptic rise in intracellular Ca2+. The L-type Ca2+ channel antagonist nifedipine did not block LTD, suggesting a role for high-threshold Ca2+ channels
(82). Perhaps the most exciting result from both articles is that D2 receptor activation appears to oppose LTD. Thomas et al. (82) applied DA for 15-20 min and blocked LTD. This was mimicked by quinpirole but not SKF 38393, suggesting mediation by D2 receptors. Jones et al. (81) found that incubating slices for 15 min with amphetamine blocked LTD and that this was prevented by the D2 receptor antagonist eticlopride. Interestingly, amphetamine had no effect on LTD at hippocampal synapses (81) and DA had no effect on LTD in the NAc (82), suggesting specificity for certain circuits. The ability of DA agonists to inhibit LTD in the midbrain is consistent with a role for high-threshold Ca2+ channels, because DA inhibits N- and P/Q-type Ca2+ currents in midbrain DA neurons
(83). The importance of postsynaptic Ca2+ for plasticity in DA neurons is consistent with in vivo studies indicating a requirement for VTA Ca2+ signaling during induction of sensitization, although these in vivo studies implicated L-type Ca2+ channels rather than the high-threshold Ca2+ channels implicated in LTD (84-86).
Jones et al. (81) proposed that "LTD normally acts to protect VTA dopamine neurons from excessive glutamatergic excitation, but that in the presence of amphetamine this brake is removed, permitting unrestricted excitation of dopamine neurons." Loss of this "braking mechanism" may promote "pathological" strengthening of excitatory synapses on DA neurons via LTP. This situation may be exacerbated by stimulant-induced increases in glutamate levels in the VTA, although exactly how drugs modulate VTA glutamate levels is a matter of some debate. We have shown that amphetamine produces a delayed but persistent increase in extracellular glutamate levels in the VTA (87-89) through a mechanism involving glutamate transporters and reactive oxygen species (90). We have argued that this delayed increase in glutamate levels is relevant to sensitization because treatments that prevent amphetamine from eliciting sensitization (pretreatment with MK-801, SCH 23390, or lesions of prefrontal cortex) also prevent amphetamine from eliciting the delayed increase in VTA glutamate levels (89). On the other hand, Kalivas and Duffy (91,92) have reported a short-duration increase in VTA glutamate levels in response to cocaine challenge in naive and chronic cocaine-treated rats. They propose that cocaine elevates DA levels in the VTA, leading to activation of D1 receptors on glutamate nerve terminals that promote the release of glutamate. Our data, and other results, are not consistent with this model (88). However, the most important point is that drugs of abuse can elevate VTA glutamate levels and this is likely to be important in triggering subsequent forms of neuroplasticity.
Another independent mechanism has been identified by which DA-releasing psychostimulants could enhance the excitability of DA neurons. Paladini et al. (93) found that amphetamine selectively inhibits slow mGluR-mediated inhibitory postsynaptic potentials (IPSPs) in DA neurons recorded from midbrain slices without affecting ionotropic glutamate receptor-mediated EPSCs. Unlike amphetamine's effect on LTD (81,82), this effect does not involve D2 receptors. Rather, the DA released by amphetamine activates postsynaptic a1 adrenergic receptors on DA neurons, leading to desensitization of inositol 1,4,5-triphosphate (InsP3)-mediated Ca2+ release from internal stores and, thus, to inhibition of mGluR-mediated IPSPs (which depend on InsP3-induced calcium release). The mechanism of the desensitization is not clear. However, this postsynaptic a1 receptor-mediated effect is functionally relevant because it leads to prolonged responses to repetitive stimulation. The authors speculate that amphetamine may thus increase bursting, which may be relevant to attentional or motivational aspects of behavior. Putting this together with the results of Jones et al. (81) and Thomas et al. (82), it appears that DA-releasing psychostimulants may promote excitation of DA neurons through two distinct mechanisms—inhibition of LTD and inhibition of mGluR-mediated IPSPs. Nicotine may also promote LTP in the VTA, albeit by a different mechanism (94).
The critical question is whether drug regimens producing sensitization in a whole animal can influence synaptic plasticity in the VTA. The first study to address this question was published recently (95). The authors prepared midbrain slices from naive mice and mice injected the day before with a single injection of saline or 15 mg/kg cocaine. The relative contribution of AMPA receptors and NMDA receptors to excitatory postsynaptic currents was compared to obtain a measure of potentiated
AMPA receptor transmission. In slices prepared from cocaine-treated mice, the authors found a significantly larger contribution of AMPA receptors (AMPA receptor/NMDA receptor ratio). This could reflect a change in the probability of transmitter release, increased AMPA receptor function, or decreased NMDA receptor function. Arguing against the first possibility, cocaine-treated mice showed no change in their responses to paired pulses. Consistent with an increase in AMPA receptor function, the cocaine-treated mice showed a significant increase in both the amplitude and frequency of miniature AMPA receptor-mediated EPSCs and also showed larger currents in response to exogenously applied AMPA. In contrast, there was no difference between NMDA-induced currents in cocaine- and saline-injected mice. The increase in AMPA/NMDA receptor contributions to postsynaptic currents was transient (present 5 but not 10 d after cocaine injection), like other adaptations at the level of the VTA, and was not seen in hippocampus or on GABA neurons of the VTA. Importantly, the authors confirmed that the single injection of cocaine used in their study produced behavioral sensitization in the mice. Coadministration of MK-801 with cocaine blocked sensitization in the mice, as expected, and also blocked the cocaine-induced change in the AMPA receptor/NMDA receptor ratio. Is the cocaine-induced potentiation equivalent to LTP? This appears to be so, because the cocaine-induced potentiation prevented the subsequent in vitro induction of LTP. Finally, it should be noted that the cocaine-induced potentiation was not accompanied by changes in GluR1 or GluR2 immunoreactivity as determined by Western blotting (95), in agreement with the idea that drugs of abuse modulate AMPA transmission in the VTA through mechanisms more subtle than altered expression of AMPA receptor subunits (75).
In conclusion, it is encouraging that the importance of augmented AMPA receptor transmission in VTA for sensitization has now been demonstrated using widely different experimental approaches, including in vitro electrophysiology (95), in vivo electrophysiology (69,70), and microdialysis (71), and that its dependence on NMDA receptor transmission is consistent with results of behavioral studies showing that intra-VTA administration of NMDA receptor antagonists prevents sensitization (28,30,31).
Was this article helpful?
Being addicted to drugs is a complicated matter condition that's been specified as a disorder that evidences in the obsessional thinking about and utilization of drugs. It's a matter that might continue to get worse and become disastrous and deadly if left untreated.