Based on animal studies, one can identify several categories of adaptations produced by chronic drug administration that are candidates for triggering long-lasting changes in brain. The first is changes in gene expression, leading to altered activity of neurons expressing these genes and, ultimately, to alterations in the activity of neuronal circuits. This subject was the topic of an excellent and very recent review, which focused on two transcription factors strongly implicated in addiction, CREB and AFosB (36). These factors mediate both homeostatic and sensitizing adaptations following repeated drug administration. However, their levels return to normal after relatively short withdrawal periods (less than 1 wk for CREB, a month or two for AFosB) (36). Although there may be other drug-regulated transcription factors or regulators that are even longer-lived, it seems most appropriate to view these factors as triggers for stable changes that occur through different mechanisms.
One such mechanism is a change in the shape or the number of excitatory synapses. This mechanism contributes to long-lasting changes in synaptic strength as a result of LTP and other forms of learning (37). Recently, changes in dendritic spines have been found after chronic drug administration. Repeated treatment with either amphetamine or cocaine increased dendritic branching, spine density, and the number of branched spines in Golgi-stained medium spiny neurons in the NAc and pyramidal neurons in the prefrontal cortex, effects which persisted at least 1 mo (38,39). Very similar alterations were observed in rats allowed to self-administer cocaine for 1 mo (40). Interestingly, nicotine produced effects similar to those found with cocaine and amphetamine [perhaps more robust; (41)], whereas chronic morphine produced effects opposite to those observed after cocaine or amphetamine, i.e., decreases in spine density and dendritic branching (42). The most important point is that these changes in dendritic morphology are identical to changes implicated in other forms of experience-dependent plasticity and are among the most long-lasting reported in response to chronic drug administration; therefore, they are good candidates for mediating its persistence. Drugs of abuse produce other types of morphological change as well. Onn and Grace (43) found that withdrawal from repeated amphetamine produced long-lasting (at least 28 d) increases in gap-junction communication in the NAc and prefrontal cortex and that this was associated with increased neuronal synchronization in these brain regions.
Neurotrophic factors may promote synaptic remodeling during learning (44,45). Thus, an intriguing possibility is that drug-induced increases in neurotrophic factor expression are responsible for alterations in dendritic morphology after chronic drug administration (46). For example, three intermittent injections of amphetamine produce a long-lasting (1 mo) increase in basic fibroblast growth factor (bFGF) immunoreactivity in astrocytes of the rat VTA and substantia nigra (SN) that is blocked by coadministration of the glutamate receptor antagonist kynurenic acid (47). A very exciting finding is that induction of bFGF appears necessary for amphetamine sensitization, as intra-VTA administration of a neutralizing antibody to bFGF prior to daily amphetamine injections prevents its development (48). A 2-wk escalating-dose amphetamine regimen, which is more similar to that shown to alter dendritic morphology in NAc and prefrontal cortex (above), also elevates bFGF in these regions (significant effect in NAc, trend in the prefrontal cortex) (49).
The hypothesis of this chapter is that activity-dependent forms of neural plasticity such as LTP or long-term depression (LTD) are the first step in the cascade leading to structural changes that underlie persistent drug-induced modifications in synaptic structure. Drugs of abuse are proposed to alter activity in circuits related to motivation and reward, leading to abnormal induction of LTP or LTD. They may also directly modify the ability of "normal" neuronal activity to elicit appropriate forms of LTP and LTD. Examples of both will be discussed below. This hypothesis is consistent with evidence for many parallels between mechanisms underlying sensitization and activity-dependent plasticty. For example, as neurotrophic factors are one of the many signaling pathways implicated in activity-depen dent plasticity (e.g., ref. 45), their proposed involvement in sensitization-related structural changes (see above) is readily incorporated into this hypothesis. The same applies to protein kinase and phosphatase cascades and to transcriptional regulation, both of which are critical for LTP and other types of learning (50-52) as well as for long-term responses to drugs of abuse (36).
A model has been proposed to explain the sequential changes that may lead from LTP and LTD to alteration in the biochemical composition of the postsynaptic membrane and, ultimately, to changes in the structure of dendritic spines (53). Within the first 30 min after induction of LTP, it is proposed that AMPA receptor signaling is enhanced by Ca2+-dependent phosphorylation of AMPA receptors, which increases their single-channel conductance, and by insertion of additional AMPA receptors into the postsynaptic membrane. The latter is probably related to Ca2+-dependent enhancement of actin-depen-dent dynamics in the spines, perhaps involving increased spine motility and the formation of synapses with discontinuities within their postsynaptic densities (perforated synapses). A later stage of this process (60 min after LTP and beyond) could result in duplication of spine synapses or the formation of new synapses. An important foundation for this model is the "silent synapse" hypothesis, which postulates that some synapses are silent at normal recording potentials because they contain only NMDA receptors; LTP is proposed to result from insertion of AMPA receptors into the postsynaptic membrane (54). Although this hypothesis was originally developed based on electrophysiological findings, it is supported by recent studies showing that the surface expression of AMPA receptors is a tightly regulated process, with AMPA receptors inserted into synapses during LTP and internalized during LTD (55,56). A more detailed model has recently been proposed based on subunit-specific regulation of AMPA receptor trafficking in hippocampal neurons (57,57a). GluR2/3-containing AMPA receptors appear to undergo continuous recycling, maintaining a constant level of synaptic AMPA receptors under normal conditions. Induction of LTP enables synaptic delivery of GluR1/2 heteromers both to silent synapses and to synapses that already contain AMPA receptors. GluRl appears to be the "dominant" subunit for this process, with protein-protein interaction domains on its C-terminus preventing GluRl/2 heteromers from being delivered to synapses under normal conditions, but enabling their delivery in response to signaling pathways activated during LTP induction. An intriguing hypothesis is that LTP may also insert into the postsynaptic membrane putative "slot proteins" that serve as binding sites for AMPA receptors. After the insertion of GluRl/2 heteromers during LTP, the resulting enhancement of synaptic strength can be maintained on a longer-term basis by exchange of GluRl/2 het-eromers with intracellular GluR2/3 heteromers; the latter are then maintained by constitutive recycling mechanisms (57).
Can drugs of abuse tap into these molecular mechanisms for altering synaptic strength? We have recently obtained evidence that D1 receptors (which are stimulated during administration of cocaine or amphetamine) can influence AMPA receptor subunit phosphorylation and surface expression. These studies were performed in primary cultures prepared from the NAc of postnatal rats. First, using phosphorylation specific antibodies and Western blotting, we found that D1 receptor stimulation increases GluRl phosphorylation at the protein kinase A (PKA) site, Ser-845 (58,59). Prior studies have shown that Dl DA receptors stimulate GluRl phosphorylation at the PKA site in striatal neurons (60,61) and that phosphorylation at the PKA site is associated with enhancement of AMPA receptor currents (60,62-64), but ours is the first to demonstrate this effect in the NAc. To study AMPA receptor surface expression in NAc, we labeled surface AMPA receptors by incubating live cells with an antibody recognizing the extracellular portion of GluRl, fixing cultures, and then incubating with fluorescent secondary antibody. Dl receptor stimulation produced a rapid (5 min) increase in punctate surface GluRl labeling, whereas glutamate decreased surface GluRl labeling (65). Downregulation of GluRl by glutamate has also been demonstrated in hippocampal cultures (55). We are in the process of determining if Dl receptor-mediated phosphorylation of GluRl is linked to Dl receptor regulation of its surface expression.
In summary, our results demonstrate that D1 receptors modulate both the phosphorylation and surface expression of AMPA receptors, the two major mechanisms for modulating the strength of excitatory synapses, and that the latter mechanism is also regulated by glutamate levels. As AMPA receptors are responsible for the majority of excitatory transmission in the NAc, these results suggest that the level of excitatory drive to NAc neurons may be dynamically regulated in response to rapid changes in the activity of both glutamate and DA afferents. This suggests a direct mechanism by which cocaine and amphetamine, through promoting D1 receptor stimulation locally or altering the activity of circuits that provide glutamate input to the NAc, could tap into fundamental mechanisms for LTP and LTD. When D1 receptors are overstimulated during chronic administration of psychostimulants, we hypothesize that adaptive changes in these mechanisms occur that influence the generation of LTP and LTD. This could ultimately lead to persistent changes in the structure and function of excitatory synapses (see above). Supporting this general possibility, a recent study showed that repeated cocaine treatment altered the coupling of D1 receptors to PKA-dependent signaling pathways that modulate AMPA receptor currents (64). Of course, the findings discussed above are most relevant for plasticity occurring at synapses onto neurons containing both D1 and AMPA receptors (e.g., NAc neurons), whereas other mechanisms must account for drug modulation of plasticity in DA neurons (which possess AMPA receptors but lack D1 receptors). The remainder of this chapter will further examine the hypothesis that drugs of abuse produce long-lasting changes in brain function by influencing activity-dependent forms of plasticity such as LTP and LTD and thereby producing changes in synaptic strength in neuronal circuits related to motivation and reward.
Was this article helpful?