Psychosomatic Medicine 61:630-650 (1999)
© 1999 American Psychosomatic Society
SPECIAL ISSUE: PSYCHOPHARMACOLOGY AND PSYCHOSOMATIC RESEARCH |
Drugs of Abuse and Brain Gene Expression
German Torres, PhD and
Judith M. Horowitz, PhD
From the Behavioral Neuroscience Program, Department of Psychology, State University of New York at Buffalo (G.T.), and Department of Social Sciences, Medaille College (J.M.H.), Buffalo, NY.
Address reprint requests to: German Torres, PhD, Behavioral Neuroscience Program, Department of Psychology, State University of New York at Buffalo, Buffalo, NY 14260. Email: gtorres{at}acsu.buffalo.edu
 |
ABSTRACT
|
|---|
Addictive drugs like cocaine, ethanol, and morphine activate signal transduction pathways that regulate brain gene expression. Such regulation is modulated by the presence of certain transcription factor proteins present in a given neuron. This article summarizes the effects of several addictive drugs on transcriptional processes contributing to the development of a drug-dependent state. The characterization of drug-induced changes in gene expression shows promise for improving our understanding of drug-addiction phenomena and cellular modes of cocaine, ethanol, and morphine action.
Key Words: addiction • dopamine • genes • mutagenesis • rodents • striatum
Abbreviations: ACTH = adrenocorticotropin hormone; cAMP = cyclicadenosine monophosphate; CART = cocaine- and amphetamine-regulatedtranscript; CBP = CREB-binding protein; CRE = cAMP-regulatedenhancer; CREB = cAMP response-element binding protein; CRF =corticotropin-releasing factor; DA = dopamine; FRAs =Fos-related antigens; G-proteins = guaninenucleotide–binding proteins; GABA = 
-aminobutyric acid; IEG = immediate-early gene; JAK2-STAT = Janus kinase–signaltransducer and activator of transcription; LC = locus ceruleus; NE = norepinephrine; NGFI = nerve growth factor I; NMDA= N-methyl-D-aspartate; PKA = proteinkinase A; PVh = paraventricular nucleus of thehypothalamus; 5-HT = 5-hydroxytryptamine; SN = substantianigra; THP = tetrahydropapaveroline; VMAT2 = vesicularmonoamine transporter 2; VTA = ventral tegmental area.
|
INTRODUCTION
|
|---|
The various factors that influence drug abuse in humans are traditionally partitioned into those reflecting the actions of genes and those based on social environments. Although deceptively simplistic, this partitioning has helped frame a comprehensive understanding of the complex factors underlying the use of, abuse of, and dependency on several drugs that target a common neural pathway that mediates their rewarding psychoproperties. The realization that cocaine, alcohol (ethanol), and morphine modify signal transduction programs that regulate cellular gene expression has reinforced the notion that genes, via the additive effects of DNA-DNA or protein-protein interactions, contribute to the development of addiction. The multiple determinants and directions of these genetic effects bring two main considerations. First, drugs of abuse indirectly induce the expression of a number of genes, which, in the context of protein synthesis, activate several networks of biochemical pathways in brain neurons. Second, the effects of drugs of abuse on cellular gene expression are multifactorial in space, time, and level, making it seem virtually impossible to trace a path from a given addictive behavior to a particular gene. With respect to the effects of social environments on drug abuse, such effects may impact only genetically influenced characteristics of drug-prone phenotypes.
Drug abuse can be viewed as a brain disease (1) characterized by neurobiological disturbances and behavioral pathologies, such as compulsive drug use and states of constant drug craving. It is hypothesized that such disturbances reflect neuroadaptive changes in signal transduction function and cellular gene expression produced by chronic drug use (2, 3). Indeed, significant evidence demonstrates that most drugs of abuse indirectly stimulate transcription of specific genes by increasing intracellular cAMP, which inevitably results in activation of multifunctional protein kinases and phosphorylation of several cellular proteins (4, 5). In this review, we explore this issue by summarizing what is known about the effects of cocaine, ethanol, and morphine on cellular gene expression. In addition, the effects of coadministration of several drugs on transcriptionally operating proteins are reviewed, because drug combination patterns are the most common clinical characteristics of drug use (6).
|
INNER WORKINGS OF TRANSCRIPTION FACTORS ON GENE EXPRESSION
|
|---|
During the course of normal and pathophysiological brain conditions, genes are transcribed in strikingly complex patterns. In response to drug-derived signals, single genes can be expressed in specific brain regions, each of which may contribute either directly or indirectly to the development and maintenance of a drug-dependent state. Not surprisingly, expression of DNA fragments is finely regulated by signal transduction factors, transcriptionally operating proteins that add specificity and stability to DNA-protein interactions. For instance, transcription factor proteins can bind to DNA through homologous groups of sequence motifs, such as homodimers and heterodimers, to turn genes on (or off), thereby regulating which transcripts are processed into functional mRNA and which are degraded (Figure 1 ). Hundreds of regulatory transcription factor proteins function by binding DNA sequences within their target genes. The best characterized in the mammalian brain are the IEGs, whose encoded proteins (eg, Fos and Jun) are rapidly, transiently, and stereotypically induced by arrays of extracellular signals, including hormones, neurotransmitters, and drugs of abuse (7–9). Such transcription factor proteins exert their regulatory function on gene expression by binding to promoters or modulatory DNA structures containing a number of cis-acting regulatory elements that function as binding sites (ie, AP-1 sites with core sequence TGACTCA) for several species of transcription factor proteins (10, 11). Promoters, therefore, constitute prime target structures through which initiation of cellular transcription by RNA polymerase can occur. It should be noted that cellular transcription can be initiated from two or more promoter sites within a single gene, thereby adding flexibility in the control of expression of that gene. Alternative promoter usage, for instance, can increase the probability of a gene being expressed in a given neuron independent of the pool of available transcription factor proteins (12). After binding, the bound transcription factor protein regulates transcription by directly activating RNA polymerase itself or by facilitating the binding of other transcription factor protein species.
View larger version (23K):
[in this window]
[in a new window]
|
Fig. 1. Schematic diagram depicting activation of signal transduction pathways in response to extracellular signals. Exposure to a neurotransmitter leads to stimulation of cell-membrane receptors (R). This stimulation leads to the brief generation of cAMP, allowing the catalytic subunits of PKA to translocate to the cell nucleus, where they can phosphorylate CREB. By means of a process not fully understood, phosphorylated CREB initiates transcription of IEGs. The protein products of these genes interact with one another and bind specifically to DNA sequences (eg, -TGACGTCA-; AP-1 site), modulating gene expression. These (late) genes encode proteins (eg, dynorphin), which then target nearby neural signaling pathways. The diagram is greatly simplified. Alterations in this intracellular signaling cascade by cocaine, ethanol, or morphine may give rise to significant neural plasticity and behavioral change. AC = adenylyl cyclase; ATP = adenosine triphosphate; GI = guanine protein, inhibitory; GS = guanine protein, stimulatory; TATA = ubiquitous sequence bound by a complex of proteins known as TFIIB. Phosphorylated CREB interacts with CBP, which itself interacts with TFIIB and possibly with the TATA-box–binding protein, TBA.
|
|
Many transcription factor proteins, like Fos proteins, are also modified extensively by phosphorylation events, which can affect the interaction of both the transcription factor protein and the transcriptional machinery (13). The best characterized example of this type of modification is the phosphorylation of CREB by kinase A holoenzymes. Such a phosphorylation event effectively facilitates the transcription (at TATA sites) and induction of CREB-regulated genes (14, 15). Current evidence indicates that most neurons, including those that serve as common recipients for the rewarding effects of psychostimulant drugs, express low basal levels of AP-1 transcription factor proteins. However, after a strong synaptic stimulus generated by drug use, there is a dramatic increase in the quantity of AP-1 transcription factor proteins and DNA-binding activity. Although these intracellular actions seem to be well documented, relating these actions to the physiology and behavior of an organism on drugs is a harder task. For instance, the identity of most target late-response genes modified by transcription factor proteins in response to drug exposure is still unclear. Therefore, in its most contemporary form, induction of transcription factor proteins by psychostimulant drugs should be viewed as a group of concerted events that occur against a complex background of intra- and intercellular signal transduction pathways. Only after such target genes are identified can we begin to understand fully the brain mechanisms underlying the addictive actions of drugs of abuse and the genetic factors that contribute to a drug-prone state.
Summary
c-fos is an IEG induced by multiple second messenger pathways. Its protein product (Fos) regulates the transcription of late-response genes. This induction is triggered by a variety of external signals (eg, neurotransmitters) that cause intracellular changes, ultimately resulting in the phosphorylation of AP-1 and CREB proteins (see Appendix 1). This sequence of events traces the transfer of external information from the cell membrane to the nucleus.
|
BRAIN CIRCUITS AND DRUGS OF ABUSE
|
|---|
Converging evidence from a variety of disciplines has identified a substantial population of cells bearing the phenotypes for DA, opioid peptides, and GABA as predominant circuits mediating the psychostimulant properties of drugs (16). These three populations of cells are essentially separate and exhibit a high degree of topographical organization within the brain parenchyma. Superimposed on this compartmentalization is a similarly high degree of synaptic codependence and extrajunctional petal and fugal influences that substantially broaden the integrative capabilities of these cell populations to respond to drug-derived signals. The prominence of these neurons in providing the stimulatory tone of most drugs of abuse strongly suggests that addiction or compulsive use of a drug reflects structural, biochemical, and/or genomic alterations within dopaminergic, opioidergic, and/or GABAergic brain systems. Indeed, as described below, most drugs of abuse alter intracellular messenger pathways into short- and long-term changes in gene expression by signal-regulated transcription factors in the aforementioned brain systems.
Psychostimulant drugs, like cocaine and amphetamine, affect a neural circuit that includes the mesocorticolimbic dopaminergic system, which has been implicated in drug-reward pathologies (Figure 2 ). DA-synthesizing perikarya or cell bodies are discretely localized to the VTA, whose axonal projections and terminal boutons ramify to the nucleus accumbens, olfactory tubercle, frontal cortex, and amygdala. There is also a nigrostriatal dopaminergic system in the mammalian brain consisting of DA cell bodies housed in the SN that project preferentially to dorsal aspects of the caudate putamen (16). DA molecules have a broad-based function; relevant to this review is that they are intimately involved in the processing of information related to rewards (17). Rewards are typically appetitive events that generate emotional experiences by inducing (subjective) feelings of euphoria and hedonia (2). Most drugs of abuse exert their addictive properties through reward-related activities of DA at individual synapses in the nucleus accumbens and frontal cortex (18). Stimulation of D1 receptor subtypes by indirect DA receptor agonists (eg, amphetamine or cocaine) increases the expression of NGFI-A, c-fos mRNAs, 46- and 35-kd FRAs, and AP-1 DNA binding activity in discrete mesocorticolimbic and nigrostriatal dopaminoreceptive structures. Coincident with these changes in gene transcription, the stimulated brain structures seem to mount a parallel response geared toward attaining regulatory adaptations in the regional complement of DA receptors, intracellular signaling pathways, or RNA editing before translation and protein synthesis. These neuroadaptations to (chronic) drug use might be involved in the etiopathogenesis of addictive behavior.
View larger version (19K):
[in this window]
[in a new window]
|
Fig. 2. Schematic representation of the chemical structures of cocaine, ethanol, and morphine. These drugs gain rapid access to rat brain tissue, where they target several neural pathways heavily represented by dopaminergic, opioidergic, and GABAergic systems. There is considerable overlap between these functional neurotransmitter systems, which provides several potential sites at which interactions relevant to the development of drug addiction may occur. AA = amygdala; Acb = nucleus accumbens; Arc = arcuate nucleus; CdP = caudate putamen; Cer = cerebellum; FC = frontal cortex; GP = globus pallidus; H = hypothalamus; Hip = hippocampus; LC = locus ceruleus; NCx = neocortex; PAG = periaquaductal gray area; RN = raphe nucleus; SN = substantia nigra; TH = thalamus; Tu = olfactory tubercle; VTA = ventral tegmental area. Neuroanatomical nomenclature is derived, in part, from the rat atlas of Kruger et al. (213).
|
|
Opioid peptide-synthesizing neurons, when compared with those of DA, exhibit a more circuitous trajectory within the brain parenchyma (Figure 2). These neurons and their axonal fibers are present near telencephalic (eg, nucleus accumbens, ventral pallidum, and amygdala), diencephalic (eg, arcuate nucleus, lateral hypothalamus, and thalamus), and mesencephalic (eg, VTA, SN, and LC) structures, where they generate antinociception, neurotransmitter/neuroendocrine flux, and drug-reward experiences (16, 19, 20). These divergent effects are mediated by the interactions of ß-endorphins, Leu-Met-enkephalins, and dynorphins, with G-protein–coupled receptor subtypes, namely µ, 
, and 
(21, 22). It is noteworthy that recent studies with rodents have revealed additional opioid receptors and ligands (eg, orphanin, nociceptin, endomorphin-1, and nocistatin) occurring in brain tissue (23–25). However, their functional significance in terms of opioid dependence or opioid withdrawal is not known. What is well established is that µ, , and receptor subtypes are superimposed on postsynaptic mesocorticolimbic and nigrostriatal neurons, where they actively regulate dopaminergic tone (26). Conversely, enkephalin- and dynorphin-synthesizing neurons in rostral forebrain express D1 and D2 receptor mRNAs, where they regulate the levels of neurotransmitter available for release in synaptic endings (27–29). These data indicate that opioid peptides and their receptors regulate processes thought to be mediated by brain DA. Moreover, because the opioidergic system can also modulate the effects of drugs of abuse that act via DA neurons, it illustrates the reciprocal interplay between two classes of neurocircuits in generating appropriately integrated responses to euphoria and reward experiences. With use of this scheme, drugs of abuse have been shown to induce the expression of AP-1 proteins (such as Fos and Jun) in striatal neurons that are postsynaptic to DA terminals. Of significance is that these neurons are rich in enkephalin and dynorphin peptides and constitute the major striatal efferent projections to the SN (ie, striatonigral) and globus pallidus (ie, striatopallidal) (30, 31). Alterations in patterns of IEG expression in opioid-containing neurons are proceeded by, and proportional to, changes in AP-1 DNA-binding activity and mRNA levels that encode the two precursor neuropeptides (32, 33). The active component of opium, morphine, is also known to induce wave patterns of c-fos expression in the caudate putamen and nucleus accumbens of the rat brain (34, 35), a finding that is consistent with the hypothesis that psychostimulant and opiate narcotic drugs target similar cellular IEGs that control expression of late-response genes. The manner in which such changes in neuronal gene expression translate to changes in behavior-determining, reward-related pathologies are completely obscure. Clearly, one of the challenges in drug addiction research is to associate the phenomenology of cellular signal transcription with behavioral function.
Synaptic inhibition in brain results largely from the binding of the small amino acid GABA to its receptors. Almost every major division of the brain and spinal cord contains GABA synapses (Figure 2). GABA molecules act mainly through GABAA and GABAB receptor subtypes, leading to the opening of chloride (Cl-)-permeable ion channels, which invariably results in large increases in membrane conductance and strong blunting of excitatory currents (36). This increase in Cl- flux is potentiated by barbiturates, benzodiazepines, and (volatile) anesthetics, such as ethanol (16). The degree of ethanol potentiation of GABAA receptors depends on the presence of certain subunits arranged around a central ion-conductance pore (37). For instance, it now seems that two subunits in complementary positions on GABA receptors are crucial for neurons to respond to ethanol (38, 39). Approximately 20 different GABA receptor subunits are expressed in different neuronal populations (40). Because such subunits can mix and match indiscriminately to form active GABA-gated Cl- channels, it is conceivable that many brain GABAA receptors are preferentially sensitive or resistant to the actions of ethanol. These findings suggest that ethanol exerts its effects by interacting specifically and directly with brain proteins. Recent studies in neuroscience have shown that ethanol (via GABAergic systems) affects neurotransmitter and hormone signal transduction pathways, leading to profound changes in neuronal function by altering basic patterns of gene expression (for reviews, see Refs. 41 and 42). In light of these observations, it seems likely that the burst firing of synaptic GABAA channels, after their activation by ethanol and other drugs of abuse (eg, opiate narcotics), plays a significant role in mediating the acute and chronic effects of ethanol as well as the ethanol withdrawal syndrome (42). Furthermore, these observations indicate that the effects of drugs of abuse are multifaceted, with diverse actions on several brain systems (eg, see Ref. 43). Indeed, an extra layer of complexity is added by the fact that cocaine, ethanol, and morphine also interact with the NMDA subtype of glutamate receptors and serotonin (5-HT) neurons. In rat brain, NMDA receptor–gated ion channel activity has been described in the striatum, a region that is densely populated with DA, opioid, and GABA receptors (44). Likewise, 5-HT input to the striatum from the midbrain dorsal raphe nucleus is massive; here, this indoleamine tonically and selectively enhances the transcription of prodynorphin mRNA (45).
Summary
The capacity for all drugs of abuse to jointly act on several brain systems that synaptically converge on a common set of mesocorticolimbic neurons illustrates the complexity of the problem of drug addiction, which has thus far outdistanced our current knowledge of brain circuits.
|
COCAINE
|
|---|
Psychostimulant drugs like cocaine and amphetamines produce a wide range of behaviors, including euphoria, sympathetic arousal, stereotypy, and motor activity (2, 6, 46). With repeated cocaine use, craving, anxiety, paranoia, and anhedonia may also develop (47, 48). Such behavioral manifestations are reflective of neuromolecular changes in brain circuits critical for induction and expression of the psychopathologies mentioned above. Manifestations of some of these behaviors seem to rely on the inner workings of the DA reuptake transporter, a transmembrane protein that controls synaptic concentrations of DA (49–51). Cocaine blocks the DA reuptake transporter, thereby increasing the availability of this catecholamine within the brain (52–54). Increased DA levels in striatal synapses subsequently lead to a cascade of intracellular events, including cAMP-dependent protein kinase activity, phosphorylation of CREB, and induction of AP-1 IEGs. For instance, acute administration of cocaine in rats elicits rapid and transient increases in the expression of c-fos, NGFI-A mRNAs, and FRAs in striatal and cerebellar neurons (55–60). Within the striatum, induction of IEGs by cocaine requires functional D1 and NMDA receptors that are either colocalized on cocaine-sensitive spiny neurons or modulated by local DA and glutamate circuit interactions (9, 56, 61, 62). Within the cerebellum, D1, D2, GABAB, and NMDA receptors mediate the effects of this psychostimulant on various mRNA molecules (63). Additional neurotransmitter systems also regulate cocaine-induced transcription factor protein expression in striosomal compartments of the rodent striatum. For example, the selective denervation of 5-HT axonal projections by p-chloroamphetamine blunts the expression of c-fos and NGFI-A mRNAs induced by cocaine in dopamine-receptive neurons (59, 64). A new and more refined approach to axonal denervation techniques is that of inactivating specific genes in mice (see below). This experimental approach also shows that 5-HT receptors (eg, 5-HT1B receptor subtypes) contribute significantly to the induction of c-fos by cocaine because 5-HT1B knockout mice exhibit reduced expression of the IEG in the striatum (65). In this telencephalic brain region, 5-HT1B receptors are localized predominantly to GABAergic axonal terminals that synapse with the VTA, SN, and pallidal neurons (66, 67). Involvement of DA, glutamate, and 5-HT in the neurobiology of cocaine use strongly suggests that this drug targets a complex mosaic of striatal neurochemical systems that are related by their anatomical connections. Relating these systems to drug-behavior function has proven difficult because of the heterogeneous patterns of feedback mechanisms that define the striatum and its axonal projections. For instance, the medium-sized spiny neuron of the striatum uses GABA as its neurotransmitter and provides inhibitory input to the globus pallidus, which in turn provides GABAergic inhibitory input to the SN or subthalamic nucleus (68). Furthermore, striatonigral and striatopallidal neurons are also known to synthesize additional chemical messengers that are expressed in a subset of closely related striatal projection nerve cells. Thus, axons arborizing to the external pallidal segment contain enkephalin and D2 receptors, whereas axons projecting to the SN contain dynorphin and D1 receptors (68, 69). However, recent work using mRNA amplified from single spiny neurons shows a high degree of overlap between D1 and D2 receptor subtypes (70). This circuitry is, of course, oversimplified, because we are omitting powerful excitatory (glutamatergic) input from the cerebral cortex that impinges on medium-sized spiny neurons to control their activity patterns (71). Nevertheless, the nature of this anatomical pathway is used as a starting point to characterize the cellular effects of cocaine. Cocaine induces specific programs of IEG expression in the striatum, preferentially in nerve cells that contain D1 receptors, that are thought to be selectively localized to neurons projecting to the SN (72). Increases in c-fos mRNA levels are found predominantly in dorsomedial and dorsolateral quadrants of the striatum as opposed to ventral striatal segments, where expression of the IEG is generally less pronounced and less uniform (26, 55, 73). Induction of this transcriptional gene shows kinetics with maximal rates of transcription occurring 60 minutes after cocaine exposure and declining steadily thereafter, typically reaching a nadir 2 hours after injection (55, 56, 74). Increased AP-1–binding activity in the nucleus accumbens roughly parallels the quantity of AP-1 transcription factors induced by acute cocaine treatment. Similarly, the time course for AP-1–binding activity is relatively short-lived, returning to basal levels within a few hours after drug exposure (75). It should be noted that not all IEGs are induced by cocaine in the striatum, because c-jun mRNA is not increased in concert with c-fos and junB (74). Thus, the cellular effects of cocaine are not only region specific (eg, the striatum) but also gene-type specific (eg, c-fos, junB, and NGFI). Finally, acute cocaine administration to rats produces minimal effects on striatal dynorphin mRNA levels (73). However, if a receptor agonist (eg, spiradoline) is given to rats before a single cocaine injection, spiradoline significantly suppresses IEGs in dopamine-receptive neurons of the caudate putamen (26). Together, these findings indicate that altering the amounts of mRNA molecules by cocaine in striatonigral neurons is a process mediated by functional D1, NMDA, and 5-HT1B receptors and inhibited by the activation of () opioid receptors.
Chronic cocaine administration to rodents results in long-lasting behavioral alterations. Such alterations are accompanied by changes in signal transduction mechanisms that regulate gene activity. For example, chronic cocaine treatment to rats leads to the desensitization of c-fos and NGFI-A mRNAs in the forebrain, caudate putamen, and nucleus accumbens (64, 75–78). Therefore, in marked contrast to the actions of acute cocaine administration, chronic treatment with this psychostimulant progressively and considerably diminishes the ability of telencephalic IEGs to further respond to transsynaptic activity. However, AP-1–binding activity, which reflects the dimerization of Fos and Jun proteins before their binding with promoter regions of genes, remains elevated for at least 18 hours after the last cocaine injection (75). Persistence of AP-1–binding activity depends on the expression of at least four FRAs (with electrophoretic mobilities in the range of 35–37 kd) that are selectively induced by chronic, but not acute, cocaine administration (76, 77). These "chronic" FRAs, with long half-lives, are preferentially expressed and localized to projection neurons bearing the phenotype for dynorphin (79). Additional long-lived FRAs (with electrophoretic mobilities in the 40- to 45-kd range) have also been found in the striatum of rats chronically treated with cocaine (79). Interestingly, some of the FRAs in the caudate putamen require functional D1 receptors for their expression because blockade of such receptors with SCH-23390 suppresses their persistent expression and inducibility (79). The phenotypic identity of FRAs seems to be related to 
FosB, a splice variant of FosB (80–82) which, along with c-fos, fra-1, and fra-2, comprise the fos family of transcription factor proteins (83). Indeed, there is considerable evidence that chronic FRAs are products of the fosB gene and that these transcriptional products are directly linked to some of the behavioral effects generated by cocaine (84). Together, these findings suggest that chronic cocaine exposure results in the induction of both transient and persistent FRAs that accumulate in a region-specific manner in brain neurons to influence, perhaps, some of the short- and long-lasting behavioral consequences of cocaine use. Among FRAs, fosB gene products are therefore the transcription factor proteins that predominate in the striatum after chronic cocaine treatment. Furthermore, FRAs are differentially regulated by DA receptors, suggesting, then, the existence of positive and negative loops within the dopaminergic system, whose function is to maintain or terminate gene activity.
Chronic cocaine treatment also leads to changes in neuropeptide gene expression, including neuropeptide Y and dynorphin (85). For example, repeated injections of cocaine to rats increases expression of mRNA encoding the opioid peptide dynorphin (26, 86, 87). Therefore, there is an inverse relationship between dynorphin expression and induction of IEGs in striatal neurons, particularly as it relates to the time course of cocaine exposure. These findings are of significant interest because they parallel those reported in human cocaine addicts in terms of a persistent elevation of dynorphin mRNA levels in the caudate (47). Because dynorphin acts on -opioid receptors to counteract the effects of D1 receptor activation, upregulation of dynorphin may therefore serve as a negative feedback signal to regulate excessive DA release resulting from frequent cocaine consumption. This is supported by the increase in dynorphin levels, because cocaine-induced c-fos is reduced over days of drug exposure, and by the fact that spiradoline blocks induction of the IEG by cocaine (26). Such a compensatory mechanism illustrates the neuronal plasticity that results from high-dose cocaine use over long periods of time. Because compensatory mechanisms of unexpected diversity and magnitude are generated in the striatal dopaminergic system as a function of drug use, they may contribute to dysphoria phenomena and the motivational effects of cocaine withdrawal.
It is now known that the molecular actions of cocaine are not restricted to multiple transcriptional factor proteins (eg, Fos), neuropeptides (eg, dynorphin), or other gene products. In fact, the effects of cocaine are more global than previously expected. Using in situ hybridization, cocaine has been shown to elicit a robust, transient increase in activity-regulated, cytoskeletal-associated (arc) mRNA levels in the rat striatum (88). The protein product of arc is localized to perikarya and dendritic processes, where it may be involved in cytoskeletal dynamics, including neurofilament expression and axonal transport (89). Because it has been proposed that psychostimulants affect major cytoskeletal proteins in rat brain (90), it is conceivable that alterations in the neuronal morphologic structure may contribute to some of the behavioral properties of cocaine use. In a different study, isolation and characterization of corresponding cDNA clones resulted in the identification of a mRNA whose relative levels are induced four- to five-fold by cocaine or amphetamine treatment (91). This mRNA, termed CART, is present as two alternatively spliced variants in rat brain, whereas in the human only a single variant is observed (92). CART is a peptide molecule discretely expressed in the shell region of the nucleus accumbens, where it coexists with GABA-containing neurons (93, 94). Its mRNA is also found in the retina, hypothalamus, spinal cord, and peripheral neurons (95). Although the functional significance of CART in terms of psychostimulant actions is unknown, it seems to respond exclusively to both cocaine and amphetamine signals because acute or chronic morphine treatment fails to induce transcriptional effects within the brain structures examined (91). The fact that CART acts as a cotransmitter with GABA in striatal projection neurons suggests that the RNA doublet modulates GABA release, particularly during the course of cocaine use. Clearly, a variety of molecules are implicated in the mechanisms of psychostimulant actions, particularly in neural constituents associated with the reinforcing effects of cocaine. More recent studies in rats have examined the effects of withdrawal from repeated cocaine treatment. For instance, withdrawal from chronic cocaine self-administration results in the induction of a novel mRNA transcript named NAC-1 because it is expressed exclusively in accumbal neurons (96). In addition, there is evidence of downregulation of the DA transporter gene and functional reorganization of striatal FRAs after cocaine withdrawal (48, 79, 97).
Other brain structures and systems also seem to respond to cocaine-derived signals, as recently shown by the observation that chronic exposure to this psychostimulant activates the JAK2-STAT pathway in VTA DA neurons (98). This signaling pathway mediates the innate immune response in animals (including humans) by inducing the expression of inflammatory genes after specific cytokines activate cognate membrane-bound receptors (for details, see Ref. 99). Along the same lines, the PVh, a nucleus that participates directly in the control of anterior and posterior pituitary hormone secretions, mounts a time-dependent response by increasing the transcription of the CRF gene after acute, but not chronic, cocaine treatment (100, 101). These latter findings are of particular interest because hypothalamic CRF and adrenal glucocorticoids actively participate in the arousal-enhancing properties of psychostimulants as well as behavioral sensitization. Furthermore, glucocorticoids, which are also secreted in response to stress, play a critical role in drug-reward pathologies by facilitating the reinforcing efficacy of cocaine, amphetamine, and morphine (102, 103). Thus far, it seems to be a general rule that many genes that are often coregulated are expressed differentially, rather than uniformly, by cocaine. The precise manner in which each gene fragment might be regulated and integrated with brain substrates associated with drug use is not yet known, although clearly each one must ultimately influence the mesocorticolimbic dopaminergic system. What is the significance of so many genes being transcribed after cocaine use? And, more importantly, how can we begin to understand the significance of these genes in anything approaching a drug-dependent state or a relapse state after cessation of cocaine consumption?
Recently developed techniques to delete specific genes in mice have significantly enhanced our ability to unravel the multilayer networks that control gene expression in the context of drug-addiction processes. Most targeting experiments performed to date have been aimed at generation of null-mutant phenotypes, commonly referred to as knockout models. These techniques have shown that genes belonging to the same family are often functionally redundant and in most cases have multiple functions (104, 105). The relevance of gene deletion, however, has been questioned because of the number of compensatory mechanisms emerging from the removal of a gene fragment and because the new phenotypes are considerably different when compared with wild-type counterparts (for details, see Refs. 106 and 107). Taking these caveats into account, gene-targeting studies show that mice lacking the DA transporter are indifferent to the behavior-activating properties of cocaine and amphetamine (50). Therefore, targeted mutagenesis clearly establishes the importance of the mutated DA transporter in mediating the effects of cocaine on some aspects of addictive behavior. However, mice that lack the DA transporter still self-administer cocaine and exhibit cocaine-conditioned place preferences (108, 109). These latter results indicate a substantial redundancy in brain systems through which cocaine signals the acquisition and maintenance of cocaine-taking behaviors. Indeed, there is evidence of the involvement of a specific 5-HT receptor subtype (5-HT1B) in processes that participate in the reinforcing properties of cocaine use (110).
Deletion of the D1 receptor subtype results in the failure of cocaine to induce wave patterns of c-fos and NGFI-A mRNA expression in striatal neurons (111). These findings stress the importance of functional D1 receptors for the induction of IEG species by cocaine. Another example is the inactivation of the fosB gene, which revealed an unexpected role of this gene on motor behavior. Mutant mice exhibit exaggerated locomotor activities as well as strong conditioned place preferences to low doses of this psychostimulant. Moreover, fosB mutant phenotypes fail to show an increase in striatal AP-1–binding activity after repeated cocaine treatment, and chronic FRAs are completely absent under both basal and cocaine-induced conditions (84). The use of knockout mice has unveiled unexpected phenotypes and unsuspected functions of deleted genes. In this respect, disruption of the VMAT2 gene results in behavioral supersensitivity to the locomotor-activating properties of cocaine and amphetamine but diminishes both conditioned place preference and sensitization to repeated psychostimulant administration (112, 113). Therefore, it seems that the protein product of VMAT2, which is responsible for sequestering DA, 5-HT, NE, and histamine molecules from the cytoplasm into synaptic vesicles, influences behavior through apparent changes in neurotransmitter storage and release parameters. Another example of the unsuspected functions of genes in null mice is that of the retinoic acid and retinoid X receptors. These nuclear receptors (eg, RARß and RXR
) play important roles during embryonic development, and, appropriately, mRNAs for these receptors are amply distributed in fetal mouse brain and spinal cord (114). Interestingly, single and compound inactivation in the genes for RARß and RXR results in locomotor defects related to reduction of D1 and D2 receptor transcripts in the caudate putamen and nucleus accumbens. Because the integrity of the dopaminergic pathway is seriously compromised by mutagenesis of retinoic acid and retinoid X receptors, null mice do not exhibit the locomotor-activating effects of cocaine (115). Therefore, the retinoic acid signaling pathway seems to play a significant role in brain function, particularly as it relates to DA-based motor behaviors. By means of homologous recombination in embryonic stem cells, additional studies have revealed the importance of DA receptors for some of the positive, reinforcing properties produced by drugs of abuse. The many biological effects of DA molecules are mediated by receptors transcribed from five distinctive genes (116). Of the three D2-like subtypes, the D3 and D4 receptors (which inhibit adenylyl cyclase activity) seem to be the primary mediators of the reinforcing effects of cocaine (117). The topographic distribution of both of these receptor subtypes is well represented in limbic structures, including the nucleus accumbens, Calleja isles, amygdala, and hippocampus (118); these brain regions are critically associated with affective, psychotic, motivation, and reward behaviors. Phenotypes with disrupted D3 and D4 receptors show significant changes in behavioral sensitivity to drugs of abuse. For example, D3 receptor mutant mice exhibit enhanced behavioral sensitivity to cocaine and amphetamine (119). Similarly, mutant mice lacking the D4 receptor are reported to be supersensitive in terms of locomotor behavior to cocaine, ethanol, and methamphetamine (120). These results indicate the importance of D3 and D4 receptors in psychostimulant action, but the precise cellular mechanisms that may contribute to the expression of behavioral supersensitivity to different types of addictive drugs have not been identified. Nonetheless, these targeted mutagenesis experiments have allowed revision of many distinctive roles of genes and have indicated that several types of drugs affect similar programs of gene expression. This raises the possibility that studies detailing the molecular mechanisms of addiction for one type of drug may be directly applicable to studies of other drugs. Therefore, the major contribution of these studies in the future will be to distill all of the above genes into a much smaller number of biological processes of known biochemical and behavioral function.
Summary
The prevailing theory is that cocaine produces reinforcing effects by elevating extracellular DA levels in drug-sensitive brain regions. This effect alters numerous intracellular signaling processes, including gene expression (see Table 1 ), resulting in the development of drug dependence. In addition to DA, other effector molecules (particularly 5-HT) also seem to play a significant role in the brain mechanisms mediating cocaine addiction.
|
ETHANOL
|
|---|
Ethanol dependence has well-defined diagnostic criteria and a well-described cellular pathophysiology (121, 122). For instance, chronic and long-term ethanol use often results in encephalopathies, such as cerebellar degeneration and pontine myelinolysis, that seriously compromise the health of affected individuals. Therefore, a great deal of effort has been devoted to understanding the molecular mechanisms contributing to these neurological disorders as well as the cellular steps that give rise to tolerance and dependence that this simple two-carbon molecule confers on specific neurons. As with most drugs of abuse, acute ethanol treatment increases the activity of certain dopaminergic and serotonergic brain pathways (Figure 2). For example, exposure to ethanol enhances levels of DA and 5-HT in the caudate putamen and nucleus accumbens of rat brain (123–125). Acute increases in monoaminergic content involve the stimulation of appropriate receptor subtypes (eg, D1, D2, 5-HT1A, 5-HT1B, and 5-HT3) that modify resting membrane potential dynamics regulating excitatory or inhibitory synapses (42, 126, 127). Ethanol also interferes with the function of ionotropic receptor complexes by potentiating GABAA ion (eg, Cl-) currents or reducing NMDA-activated ion (eg, Ca2+) fluxes in different regions and even in different neurons of the rodent brain (128). Because random opening and closing of individual ion channels are sensitive to the actions of ethanol, it is hypothesized that the subunit composition of each receptor defines the sensitivity tone to ethanol. Indeed, molecular studies show that certain GABA- (eg, ß) and NMDA receptor subunit complexes (eg, NR2A–NR2D) can significantly affect the response of each receptor to ethanol (for review, see Ref. 127). Current work on null mice should yield more information about the level of involvement of these ubiquitous subunits in the actions of ethanol as well as their role in tolerance and withdrawal phenomena.
Alterations in the amounts of various RNA species within brain (eg, the endocrine hypothalamus) have also been detected in rodents after low to moderately intoxicating doses of ethanol. For instance, administration of ethanol to rats rapidly upregulates intronic CRF heteronuclear RNA levels in the PVh as well as the secretion of pituitary ACTH and adrenal corticosterone (for review, see Ref. 129). Along the same lines, mRNA species encoding the CRF type 1 receptor are upregulated by the presence of ethanol within PVh neurons (130). These observations indicate that ethanol (or nonspecific stress responses triggered by ethanol) alters hypothalamic constituents that influence a variety of physiological functions, including the immune system, cardiovascular tone, reproductive function, and gastrointestinal responses.
Chronic ethanol treatment has considerable effects on GABA and glutamate receptor subunits, G-proteins, protein kinase C, nitric oxide synthase activities, phosphorylation of CREB, prodynorphin mRNA levels, and proopiomelanocortin gene expression (131–137). Of significant interest in terms of gene transcription is that acute ethanol treatment triggers an increase in CREB content in granule cell layers of the rat cerebellum, with peak levels of phosphorylated CREB detected 30 minutes after drug exposure. In contrast, 5 weeks of ethanol consumption results in significantly lower levels of phosphorylated CREB in striatal or cerebellar neurons (137). These findings indicate that the actions of ethanol vary considerably as a function of drug use and emphasize the global effects of this two-carbon molecule on intracellular signaling pathways. Ethanol, when administered over a period of days or months, also interacts with opioid neurocircuits, such as those that regulate the secretion of ß-endorphins and prodynorphin-derived peptides (for review, see Ref. 138). Such an interaction invariably results in activation of the endogenous opioid system, leading to positive reinforcing and anxiolytic effects that trigger development of ethanol dependence (139). The fact that opioid peptide secretion is influenced by ethanol consumption adds another level of complexity to the neuronal networks (eg, GABA, glutamate, 5-HT, and DA) contributing to ethanol-seeking behaviors. Therefore, the precise mechanisms involved in the interaction of ethanol with opioid constituents have not been fully characterized. Regardless of such complexity, blockade of opioid receptors (eg, µ and ) by naltrexone (Revia or Trexan) significantly reduces ethanol intake in both humans and rodents. The anticraving properties of naltrexone seem to influence the mesocorticolimbic DA reward circuitry; this pathway is closely regulated by opioid receptors and opioid neuropeptide secretion (140).
The functional consequences of ethanol treatment on IEGs have also been studied in rat brain. In all of these studies, ethanol by itself did not alter total brain c-fos, c-jun, or NGFI-ß mRNA expression (141–147). This phenomenon is well documented in different cell populations of the caudate putamen, nucleus accumbens, hippocampus, and hypothalamus, where low to moderate doses of ethanol reduce the amounts of transcriptionally operating factor proteins (143). It should be noted, however, that other investigators report induction of IEGs by acute ethanol treatment in the stria terminalis, PVh, amygdala, hippocampus, LC, and parabrachial nucleus (148–150). The discrepancy in results remains unexplained but may reflect differences in rat strains, ethanol doses, individual RNA molecules, or location of the IEGs within brain parenchyma. Nevertheless, evidence that ethanol induces transcription signals that stimulate cellular gene activity is, at best, weak. The significance of this work on ethanol and IEGs extends far beyond the description of the two-carbon molecule preventing the transcription of striatal or hypothalamic genes. It also impinges on a much greater issue, polydrug use and abuse, which, despite its importance and relevance to drug addiction pathology, has been seriously neglected. As noted above, ethanol by itself does not induce wave patterns of IEG expression in brain. Furthermore, if ethanol is coadministered with cocaine, the expected expression of the c-fos gene in striatal neurons is significantly blunted (143, 146). Therefore, in the context of polydrug use, ethanol alters the effects of cocaine on gene transcription (Figure 3 ). Because ethanol affects the transcriptional machinery of selective neurons, an important question is how this drug modifies signal transduction pathways that regulate transcription factor protein function. Although the mechanisms are not well characterized, it is thought that ethanol suppresses cocaine-induced Fos-like proteins through inhibition of NMDA-gated ion channels (143). This hypothesis is based on two lines of evidence. First, MK-801 and CPP, powerful NMDA receptor antagonists, block induction of IEGs by cocaine (58, 61). Thus, when specific NMDA receptor antagonists alter AP-1 transcription factor proteins, one can infer that ethanol, which antagonizes the effects of glutamate, is also involved in that process. Second, ethanol at concentrations of 10 to 40 mM (levels associated with moderate ethanol intake in humans) potently and selectively blocks NMDA receptor function (124, 151). Consistent with this, ethanol at concentrations of approximately 40 mM reduces the expected expression of striatal Fos-like proteins by cocaine (143). The specific nature of ethanol-cocaine interactions at the gene level has not been well defined, but the fact that human self-administration often occurs in the context of polydrug use warrants further detailed and systematic studies in this area.
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 3. Schematic representation depicting the effects of ethanol ( 40 mM), cocaine (20 mg/kg), and ethanol-cocaine treatments on Fos-like protein in striatonigral neurons. Ethanol by itself has no effect on the cellular expression of this transcription factor protein. In contrast, cocaine treatment markedly induces the expression of Fos-like protein, particularly in the dorsal caudate putamen. However, concomitant administration of ethanol and cocaine reduces the expected expression of the c-fos gene. Sterile NaCl was used as vehicle.
|
|
Another important issue related to polydrug use and abuse is that of the metabolic interaction resulting from the combined ingestion of several addictive drugs. In the context of ethanol and cocaine consumption, cocaethylene, an active cocaine metabolite, has been implicated in promoting and sustaining substance dependence by acting on specific brain systems critical for drug-reinforcing phenomena (152–155). Cocaethylene produces neurochemical effects qualitatively similar to those of cocaine, including augmentation of DA molecules across the synapse in both rat and nonhuman primate striata (156, 157). Cocaethylene also induces IEG expression (Figure 4 ) in discrete neurons of the rat caudate putamen and nucleus accumbens (143, 158). This nuclear event is purportedly mediated by D1- and NMDA receptor subtypes because blockade of these membrane-bound receptors prevents induction of the c-fos gene (158). The effects of cocaethylene on gene transcription also extend to other species of the IEG family, namely NGFI-ß (Figure 5 ). This gene is constitutively induced in parvocellular (and magnocellular) neurons of the PVh in response to ethanol, cocaine, and cocaethylene treatment (144, 159). The fact that cocaethylene alters the amounts of NGFI-ß mRNA molecules within the PVh, a cellular nucleus regulating homeostasis in the mammalian body, indicates that the ethanol-cocaine metabolite may be an additional factor augmenting stressful internal cues associated with compulsive drug craving. In this respect, cocaethylene also triggers the secretion of stress hormones, including ACTH and corticosterone, in laboratory animals (159). These hormones, particularly corticosterone, seem to facilitate the acquisition of drug self-administration by increasing the reinforcing efficacy of cocaine and ethanol on brain reward circuits (103). Therefore, cocaethylene may be indirectly involved in the pathophysiology of stress, depression, and panic attacks that characterize the major psychiatric consequences of chronic polydrug abuse and withdrawal in humans. It should be noted that the capacity of ethanol and cocaine to yield the formation of a neuroactive metabolite does not seem to be a promiscuous attribute of all drugs that are concomitantly used and abused. We are unaware of clear demonstrations of additional metabolites affecting signal-regulated transcription factor proteins that control cellular gene expression. It is important to point out, however, that to date few studies have investigated the effects of polydrug use on gene transcription; therefore, such effects might still be observed in other drug combination patterns. For instance, long-term ethanol consumption induces the formation of certain morphine precursors (eg, THP and norcoclaurine) in striatal neurons of rat brain (160). Under a number of experimental preference conditions, THP treatment in rats and nonhuman primates enhances ethanol intake (161, 162). These findings provide supporting evidence for the previously discussed hypothesis that derivatives of drug combination patterns may yield active molecules with significant biological value and function.
View larger version (67K):
[in this window]
[in a new window]
|
Fig. 4. Photomicrographs illustrating the striking effects of cocaethylene on brain IEGs. Cocaethylene is derived from the combined ingestion of cocaine and ethanol. This neuroactive metabolite induces the expression of Fos-like protein in dopaminoreceptive neurons of the caudate putamen (cdP). Top, Schematic illustration drawn through a comparable level of the cdP, where induction of Fos-like immunoreactivity is persistently observed after acute cocaethylene treatment. Middle, Representative photomicrograph through the cdP of a rat injected with sterile NaCl. Note the relative absence of Fos-like protein on medium-sized spiny neurons. Bottom, Representative photomicrograph through the cdP of a rat injected with cocaethylene (20 mg/kg). Note the distribution pattern and relative strength of the transcription factor protein. Asterisk indicates unreactive myelinated bundles of corticofugal fibers coursing through the cdP.
|
|
View larger version (61K):
[in this window]
[in a new window]
|
Fig. 5. Photomicrograph illustrating the induction of NGFI-ß mRNA molecules in the PVh of the hypothalamus by cocaethylene (cocaeth). Top, Schematic coronal view depicting the topographical organization of the PVh with an emphasis on parvocellular neurosecretory neurons. Middle, Photomicrograph through a common level of the PVh of a rat injected with NaCl. Hybridized neurons are rare in parvocellular aspects of this hypothalamic nucleus. Bottom, Representative photomicrograph of the PVh of a rat injected with cocaethylene. A reliable increase in the number of positively hybridized neurons in the parvocellular (and magnocellular) aspects of the PVh is apparent 30 minutes after cocaethylene treatment. * = third ventricle; pv = periventricular. Reprinted with permission from Ref. 159.
|
|
Research efforts over the last 10 years, particularly in quantitative and molecular genetics, have recognized that ethanol abuse has a strong genetic contribution. This in itself is not surprising given that genetic variation in susceptibility is a feature of many of the most common diseases catalogued in humans. Thus, a number of family studies show that genetic susceptibility to ethanol dependency seems to be associated with relatively common alleles of one or several loci (163, 164). An important focus currently is to locate and characterize reliable trait markers that determine vulnerability to the actions of ethanol. In this context, a proposed marker, the A1 allele of the D2 receptor gene, shows a significantly higher prevalence in ethanol users compared with nonusers (164, 165). Although there is good reason for optimism at this point, it is clear that we know far too little about the polymorphism and distribution of this or other alleles that determine susceptibility to ethanol and still less about how these alleles work and what they do to bring about such a disorder. Despite these caveats, the contribution of polygenic determinants of ethanol liability in humans is further supported by comparative studies of behavioral responses to ethanol in inbred rodents that show differences in brain transmitter function (166–168). For example, differences in 5-HT neurotransmission are linked to ethanol effects on brain tissue and to ethanol abuse (see Ref. 41 and references therein). This indoleamine seems to play a critical role in the modulation of ethanol consumption and in the development of tolerance to the two-carbon molecule (169). The general consensus is that there is an inverse relationship between brain 5-HT transmission and ethanol consumption in mammals. For instance, low 5-HT neurotransmission is correlated with excessive ethanol intake in a subgroup of human alcoholics (170) and in ethanol-preferring rats (171). Changes in 5-HT function may also contribute to the ethanol withdrawal syndrome through enhanced 5-HT activity at the 5-HT2 receptor subtype (172). In animal models of ethanol withdrawal, there are significant changes in transcriptional mechanisms when this drug is withheld after chronic use. These animals show transient increases in the expression of c-fos, c-jun, and NGFI-A mRNAs in the cerebral cortex, olfactory bulb, granular layers of the cerebellum, and hippocampus (173, 174). The hyperexcitability in membrane-dependent events leading to the induction of IEGs seems to reside, at least in part, in increased activity of 5-HT and glutamate receptors and decreased inhibitory function of GABAA receptors, all of which are upregulated or downregulated, respectively, during chronic ethanol intake (127, 175, 176). Significant evidence also supports a role of the 5-HT1B receptor subtype in ethanol consumption, as demonstrated in null-mutant mice lacking the indoleamine receptor gene mentioned above. This targeted mutagenesis experiment shows that knockout mice drink more ethanol than their wild-type counterparts but that they exhibit reduced sensitivity to ethanol ataxia in the grid test paradigm (177). These findings also support the involvement of serotonergic systems in the regulation of ethanol intake and facilitate both the understanding of function of the 5-HT1B gene and the view that certain behaviors related to ethanol are controlled by several genes. This view adds significantly to the hypothesis that ethanol use and abuse are composites of large biochemical complexes and cascades, where each step is refined and directly dependent on many genes. For example, mutant phenotypes lacking a functional gene encoding for the enzyme Fyn tyrosine kinase, which has been linked to spatial learning and long-term potentiation, are hypersensitive to the hypnotic effects of ethanol (178). Fyn tyrosine kinase is involved in intracellular signaling and is strongly expressed in brain, where it phosphorylates GABAA and NMDA receptors (179). Phosphorylation by Fyn tyrosine kinase may determine the ethanol sensitivity of GABAA and NMDA receptors for their appropriate native neurotransmitters. Furthermore, disruption of the mouse D1 receptor by mutagenesis results in attenuation of ethanol-seeking behaviors (180). Thus, a highly integrated and redundant network of gene responses to ethanol intake are revealed by the mouse models described here.
Summary
There is evidence of the existence of ethanol "receptors" on cell membranes. Behavioral and pharmacogenetic assays using (primarily) rodent lines support a neurotransmitter (DA, 5-HT, GABA, and glutamate) theory of ethanol action on the brain. Mutagenesis studies also support this theory by demonstrating the importance of specific ligands and receptors necessary for the hypnotic actions of ethanol.
|
MORPHINE
|
|---|
Morphine and its synthetic analog, heroin, are thought to produce reinforcement phenomena via stimulation of µ, , and opioid receptors in reward-relevant brain systems (181, 182). Receptor-binding studies show that morphine has the highest affinity for µ, followed by , and the lowest affinity for (182, 183). These G-protein–coupled receptors, with seven membrane domains, have stimulatory effects (eg, increasing the rate of neuronal firing) as well as inhibitory effects (eg, closing of voltage-sensitive Ca2+ channels or inhibition of cAMP synthesis) on neurotransmission (184). The overall impact of morphine on brain tissue is therefore determined by which neurons express µ, , or receptors. As described earlier, opioid receptors are heavily invested in brain systems that regulate stress perception, pain control, reward behavior, and neurohormone secretion (185). Antagonist compounds (eg, naltrexone) at these receptors alleviate addiction to morphine and ethanol dependence (140). Morphine exposure facilitates drug self-administration behavior, enhances the reinforcing value of electrical brain stimulation and induces conditioned place preference in laboratory animals (186–188). These positive motivational actions are indicative of the rewarding properties of morphine that could drive liability to produce narcotic addiction in humans (20). In vivo microdialysis shows that morphine increases extracellular levels of DA in the nucleus accumbens shell of the rat brain (189, 190) and elicits stereotyped behaviors that are reminiscent of those seen after cocaine or amphetamine treatment (see Ref. 188 and references therein). These studies provide strong evidence that DA mediates the reinforcing and behavioral effects of morphine. It is noteworthy that morphine and related congeners may also act independently of DA neurons to maintain opiate-reward effects via activation of opioid receptors expressed on accumbal neurons (20). Morphine also indirectly generates nerve impulses in VTA DA neurons through hyperpolarization of local GABA interneurons containing µ-opioid receptors (191). Morphine therefore has the ability to act on opioid receptors in both the VTA and nucleus accumbens to influence some of the cellular features underlying opiate addiction. Nonetheless, DA-opioid interactions in the nucleus accumbens are still necessary synaptic events for the expression of certain negative states associated with morphine dependence, including somatic signs of opiate withdrawal. For instance, stimulation of D2 receptors by DA agonists in morphine-dependent rats significantly attenuates a variety of withdrawal symptoms precipitated by the opiate antagonist naloxone (192). Conversely, injections of D2 receptor antagonist into the nucleus accumbens shell produce signs of opioid withdrawal (192). Thus, endogenous dopaminergic systems in accumbal neurons are directly implicated in the regulation of symptoms associated with morphine withdrawal behavior. The fact that drugs of abuse with very different initial actions (eg, cocaine and morphine) converge on a common substrate (eg, the nucleus accumbens) suggests that this brain region is one locus generating reward and withdrawal motivation syndromes (2, 193).
Exposure to morphine results in rapid and profound behavioral changes (eg, tolerance and sensitization) that modify the function of specific targeted neurons. For instance, chronic morphine treatment in rats leads to a compensatory upregulation of G-protein subunit levels, adenylyl cyclase activity, and cAMP-dependent protein kinase activity in nucleus accumbens and LC neurons (for review, see Ref. 20). The LC, a small pontine nucleus, contains a substantial population of cell bodies that synthesize NE for synaptic release into various brain circuits associated with attentional and autonomic functional states (194). The LC has served as a useful model system for describing basic cellular and molecular events that underlie morphine action. Briefly, on contact with µ-opioid receptors, morphine increases the conductance of a K+ channel, which is operated by networks of G-protein signaling systems. This opens the way for inhibition of adenylyl cyclase activity, which in turn reduces the level of cAMP. Inhibition of cAMP by morphine shuts down PKA activity and decreases the phosphorylation rate of CREB (20). This regional and selective effect of morphine on cAMP protein phosphorylation, in addition to altering gene transcription, may also indirectly generate specific behaviors commonly associated with addiction (eg, tolerance and withdrawal). Indeed, if morphine treatment persists and is then abruptly terminated, LC neurons show excitation at action potential frequencies thought to be signs of withdrawal pathologies (20). However, it is not clear whether the expression of morphine withdrawal in the LC is an intrinsic event or whether it is driven by extrinsic afferent signals (eg, glutamatergic) that impinge on NE-coding neurons (for review, see Ref. 195). Nonetheless, there are clear precedents of neural adaptations in the rat LC and VTA after chronic morphine use, including oversynthesis of tyrosine hydroxylase (the rate-limiting enzyme in the production of DA), alterations in the composition of glutamate receptor subunits, increases in G-protein–coupled receptor kinase 2, decreases in neurofilament proteins (eg, NF200, NF160, and NF68), and increases in glial fibrillary acidic proteins (20, 196). The findings concerning the effects of morphine on the cytological structure of VTA neurons are of significant interest because morphometric changes in DA neurons could restrict the flow of electrical currents spreading through neuronal processes, thereby providing a structural base for behavioral impairments (eg, motivational dependence) associated with frequent opiate use. This hypothesis has merit because chronic morphine treatment specifically reduces the cell body size of VTA DA neurons by an average of 25%. Remarkably, reduction in neuronal size can be reversed either by naltrexone or intra-VTA infusion of brain-derived neurotrophic factor molecules (197). The fact that neurotrophins prevent the "shrinking" effects of morphine on rat DA neurons indicates that trophins could be used as therapeutic agents of target control for neuronal survival and function in the opiate-addicted brain.
Chronic morphine treatment increases Leu- and Met-enkephalin levels in the VTA and Leu-enkephalin and preprodynorphin in the nucleus accumbens of rats (198, 199). It is noteworthy that increases in opioid mRNA levels due to morphine are dependent, at least in part, on genetic factors that control cAMP-dependent phosphorylation in DA neurons (198). In this respect, prominent differences exist in levels of tyrosine hydroxylase and related morphine-regulated phosphoproteins in the VTA and SN of several inbred rat strains (200) that could give rise not only to differences in opiate preference but also to differences in opioid hormone secretion. Therefore, overall changes in the basal steady state of opioid peptide gene expression after morphine treatment are not fixed but manifest plasticity as a function of background genes. Morphine acts at µ-opiate receptors to indirectly release DA in striatal synapses. Such a membrane event triggers transcriptional activation of the prototypic IEGs, c-fos and junB, in caudate and accumbal cell bodies (34). Initial approaches to defining the receptor mechanisms involved in this stimulus-transcription coupling event reveal that both D1 and NMDA receptors mediate the induction of such genes by morphine (34). Therefore, different types of addictive drugs affect similar neuroreceptors and neuroanatomical pathways, often causing profound changes in neuronal function by altering basic patterns of gene expression. In the LC, however, there is a divergence of intracellular effects. Acute or chronic morphine administration suppresses the induction of c-fos mRNA, whereas after withdrawal, an abrupt and significant induction of the IEG is noted in the LC, amygdala, neocortex, hypothalamus, and brainstem regions (201, 202).
Transcriptional activation of cAMP-responsive IEGs by morphine is also regulated by the presence of nuclear CREB (14, 203). CREB is phosphorylated by the catalytic subunit of PKA at a single specific amino acid, Ser-133, thereby promoting the association of CREB with CBP, leading to efficient transcriptional activation of CREB-regulated genes (204). These late response genes (eg, preproenkephalin and preprodynorphin) encode proteins necessary for short- or long-term adaptation of synaptic events induced by DA release. It is speculated that alterations in signal transduction (like the one for nuclear CREB) may underlie the molecular basis for tolerance, dependence, and withdrawal symptomatology to morphine (205). Several lines of research support this hypothesis. First, acute morphine treatment decreases rates of CREB phosphorylation in the rat LC, an effect not seen after chronic morphine exposure. However, when opiate withdrawal is precipitated by naltrexone, PKA-dependent phosphorylation of CREB is abruptly increased (206). Second, chronic morphine treatment decreases CREB phosphorylation in accumbal neurons, suggesting that two morphine-sensitive brain regions display contrasting features of adaptation, regulation, and expression of phosphorylated CREB (207). The implication is that alterations in CREB biology may impair the temporal or spatial induction of genes involved in the control of attentional (eg, LC) or reward (eg, nucleus accumbens) behaviors. The importance of phosphorylated CREB for the induction of c-fos, NGFI-A, and neuropeptide mRNA molecules after morphine treatment is also evidenced in amphetamine and ethanol studies showing impaired phosphorylation of CREB and CRE-binding activity in striatonigral neurons (29, 57, 137, 208, 209). Antisense technology, the introduction of modified oligonucleotides into neurons that alters the processing, transport, or translation of targeted mRNA species (210), has also been used to further illustrate the impairment of CREB in transcriptional activation of striatal IEGs and prodynorphin gene expression by drugs of abuse (29, 209). Thus, phosphorylation of CREB at a serine residue by PKA can be significantly modified by cocaine, ethanol, or morphine treatment, thereby producing potential effectors of long-term neuroadaptations to psychostimulants, depressants, or narcotics.
The importance of µ and D2 receptors and the cAMP signaling pathway for the actions of morphine is also evidenced in gene-targeted line mutations. For example, disruption of the µ-opioid receptor in mice by homologous recombination results in animals that do not exhibit morphine-induced analgesia, reward, or withdrawal symptoms after a naloxone challenge (211). Although and receptors exhibit normal expression in the mutant brain, stimulation of these cell-surface receptors by morphine produces no salient behavioral responses. These findings suggest, therefore, that µ-opioid receptors are the specific molecular targets of opiates in mammals and that such receptors represent obligatory components of the pharmacological effects of morphine. Significant new evidence using current knockout technology shows that disruption of the and isoforms of CREB results in morphine-dependent mice having less severe signs of withdrawal symptoms after a naloxone challenge (203). CREB activity, therefore, is implicated in naloxone-precipitated withdrawal manifestations of morphine dependence in mice. It should be noted that CREB mutagenesis does not affect the predicted upregulation of IEGs in LC neurons, nor does it affect the increased level of cAMP that typically occurs after morphine abstinence. Thus, a key component of the cAMP signal transduction pathway that regulates gene expression seems to be the common mediator in the mechanisms associated with physical dependence and subsequent physical withdrawal from morphine use (205). Another example is the inactivation of the D2 receptor by mutagenesis, which reveals the importance of this receptor subtype for the rewarding properties of morphine. Mice lacking brain (eg, nucleus accumbens) D2 receptors do not show the normal positive responses to morphine, as observed in place-preference tests (212). Surprisingly, physical dependence on morphine is not suppressed in mutant mice, which exhibit a full profile of withdrawal symptoms, including ptosis, diarrhea, lacrimation, and weight loss. Thus, it seems that the D2 receptor is required, at least for the motivational component of morphine addiction. However, these findings differ considerably from those obtained in wild-type animals, in which the involvement of D2 receptors for the expression of somatic symptoms of morphine withdrawal is stipulated (192). These discrepancies may be due to either the introduced null mutation or background genes linked to the targeted locus (for review, see Ref. 106). For instance, deletion of the D2 receptor from the mouse genome might have perturbed specific developmental programs of DA gene expression in selective neurons of the nucleus accumbens. Included in this reasoning is the expression of alternative alleles in nascent neurons to compensate for the disruption of the D2 receptor gene. In this context, allelic compensatory processes can have profound interactive effects with the mutated DNA fragment. It is also conceivable that the phenotypic consequences of the mutated D2 receptor gene may simply reflect polymorphisms in the genetic background of the selected mouse line (for review, see Ref. 107). Thus, target mutagenesis approaches are not devoid of constraints. Perhaps new approaches with more refined and sophisticated techniques will restrict inactivation of the target gene to specific neuron type or to defined developmental stages of the animal. This approach, aptly called "conditional" gene disruption, will be more informative and will no doubt lead to further advances in the neurobiology of opiate abuse.
Summary
Opioid receptors are targeted by exogenous opiates, the prototype of which is morphine. Rapid activation of the µ-opioid receptor by morphine results in a euphoric phenotype, thus conferring the reinforcing effects of the drug. This activation is accompanied by extracellular DA release, which alters several events related to the cAMP signal transduction pathway. Of particular significance is that CREB seems to be modified by morphine, thereby affecting addictive behavioral phenomena, such as withdrawal symptomatology.
|
CONCLUSIONS
|
|---|
TOP
ABSTRACT
INTRODUCTION
