|Home | About | Journals | Submit | Contact Us | Français|
Drugs of abuse produce widespread effects on the structure and function of neurons throughout the brain’s reward circuitry, and these changes are believed to underlie the long-lasting behavioral phenotypes that characterize addiction. Although the intracellular mechanisms regulating the structural plasticity of neurons are not fully understood, accumulating evidence suggests an essential role for neurotrophic factor signaling in the neuronal remodeling which occurs after chronic drug administration. Brain-derived neurotrophic factor (BDNF), a growth factor enriched in brain and highly regulated by several drugs of abuse, regulates the phosphatidylinositol 3′-kinase (PI3K), mitogen-activated protein kinase (MAPK), phospholipase Cγ (PLCγ) and nuclear factor kappa B (NFκB) signaling pathways, which influence a range of cellular functions including neuronal survival, growth, differentiation, and structure. This review discusses recent advances in our understanding of how BDNF and its signaling pathways regulate structural and behavioral plasticity in the context of drug addiction.
An essential feature of drug addiction is that an individual continues to use drug despite the threat of severely adverse physical or psychosocial consequences. Although it is not known with certainty what drives these behavioral patterns, it has been hypothesized that long-term changes that occur within the brain’s reward circuitry are important (Figure 1). In particular, adaptations in dopaminergic neurons of the ventral tegmental area (VTA) and in their target neurons in the nucleus accumbens (NAc) are thought to alter an individual’s responses to drug and natural rewards, leading to drug tolerance, reward dysfunction, escalation of drug intake, and eventually compulsive use (Everitt et al., 2001; Kalivas and O’Brien, 2008; Koob and Le Moal, 2005; Nestler, 2001; Robinson and Kolb, 2004).
There has been a major effort in recent years to determine the cellular and molecular changes that occur during the transition from initial drug use to compulsive intake. Among many types of drug-induced adaptations, it has been proposed that changes in brain-derived neurotrophic factor (BDNF), or related neurotrophins, and their signaling pathways alter the function of neurons within the VTA-NAc circuit and other reward regions to modulate the motivation to take drugs (Bolanos and Nestler, 2004; Pierce and Bari, 2001). A corollary of this hypothesis is that such growth factor-induced cellular and molecular adaptations are reflected in morphological changes of reward-related neurons. For example, chronic stimulant administration increases branching of dendrites and the number of dendritic spines and dynamically increases levels of BDNF in several brain reward regions, whereas chronic opiate administration decreases dendritic branching and spines as well as BDNF levels in some of the same regions (for review see (Robinson and Kolb, 2004; Thomas et al., 2008). Moreover, chronic morphine decreases the size of VTA dopamine neurons, an effect reversed by BDNF (Russo et al., 2007; Sklair-Tavron et al., 1996). However, direct, causal evidence that these structural changes drive addiction remains lacking.
The proposal that BDNF may be related to structural plasticity of the VTA-NAc circuit in addiction models is consistent with a large literature which has implicated this growth factor in regulation of dendritic spines. For instance, studies using conditional deletions of BDNF or the TrkB receptor show that they are required for proliferation and maturation of dendritic spines in developing neurons as well as the maintenance and proliferation of spines on neurons throughout the adult brain (Chakravarthy et al., 2006; Danzer et al., 2008; Horch et al., 1999; Tanaka et al., 2008a; von Bohlen Und Halbach et al., 2007).
Although the exact molecular mechanisms by which BDNF mediates structural plasticity of the brain’s reward circuitry remain unknown, recent studies suggest that specific pathways downstream of BDNF are modulated by drugs of abuse, and that these neurotrophic factor-dependent signaling changes correlate with morphological and behavioral end-points in animal models of drug addiction. In this review, we discuss new advances in our understanding of how opiates and stimulants regulate neurotrophic factors signaling and the cellular and behavioral consequences of these effects. We also propose areas for future investigation to address the paradoxically opposite effects of stimulants and opiates on neuronal morphology and certain behavioral phenotypes consistent with addiction.
Uncovering the signaling pathways that mediate neuronal development and survival has been a long-time goal of neuroscience research. However, neurotrophic factor signaling in the adult central nervous system (CNS) has over the past decade become an important area of interest, as neurotrophic signaling has been shown to modulate neural plasticity and behavior throughout an organism’s life (for review see (Chao, 2003)). The first neurotrophic factor identified, nerve growth factor (NGF), was isolated in 1954 (Cohen et al., 1954); cloning of the gene itself did not occur until 1983 (Scott et al., 1983). This discovery was followed closely by the purification and identification of additional NGF-like growth factors that defined a neurotrophin family: BDNF (Barde et al., 1982; Leibrock et al., 1989), neurotrophin-3 (NT3) (Hohn et al., 1990; Maisonpierre et al., 1990), and neurotrophin-4/5 (NT4/5) (Berkemeier et al., 1991). Neurotrophin family members are paralogs and share significant homology (Hallbook et al., 2006); all are polypeptides that homodimerize and are found in both immature and mature forms in the CNS. While it has long been thought that the cleaved ~13 kDa mature form was the active signaling molecule, recent studies indicate that the pro- (immature) forms of the neurotrophins, which retain their N-terminus, are detectable in the brain (Fahnestock et al., 2001) and mediate signaling cascades distinct from the mature peptides. The actions of NGF in the adult CNS are largely localized to cholinergic cells in the basal forebrain, while the distribution of the other neurotrophins is much more widespread.
Further specificity of the neurotrophin signal is produced through the differential expression of neurotrophin receptors, which can be separated into two categories, the tropomyosin-related kinase (Trk) and p75 neurotrophin (p75NTR) receptors. The p75NTR was first identified as a receptor for NGF (Johnson et al., 1986), but actually binds both the immature and mature forms of all four neurotrophins (Lee et al., 2001; Rodriguez-Tebar et al., 1990; Rodriguez-Tebar et al., 1992). As opposed to p75NTR, the Trk family of receptors exhibits specificity for its ligands. The TrkA receptor preferentially binds NGF (Kaplan et al., 1991; Klein et al., 1991), the TrkB receptor binds BDNF (Klein et al., 1991) and NT4/5 (Berkemeier et al., 1991), and the TrkC receptor binds NT3 (Lamballe et al., 1991). While the mature neurotrophins have an increased affinity for Trk receptors compared to the propeptides, both the immature and mature forms can bind p75NTR with high affinity. Additionally, p75NTR has been shown to form complexes with Trk receptors, and these receptor complexes exhibit an increased affinity for the respective Trk ligands compared to homodimeric Trk.
Trk receptors are single transmembrane spanning proteins composed of an extracellular ligand binding domain and an intracellular region containing a tyrosine kinase domain. Similar to other receptor tyrosine kinases, Trk receptors homodimerize in response to ligand binding, which allows for trans-phosphorylation within the activation loop to increase catalytic activity of the receptor kinase. Trans-phosphorylation at tyrosine residues in the juxtamembrane domain and in the C-terminus generates attachment sites for SH2 (Src homology 2)-type “linker” proteins, such as Src homology domain-containing protein (Shc), and phospholipase Cγ (PLCγ), respectively. Shc binding initiates downstream signaling cascades leading ultimately to the activation of mitogen-activated protein kinase (MAPK) and phosphatidylinositol 3′-kinase (PI3K) pathways. Stimulation of the MAPK pathway includes activation of extracellular signal-regulated kinase (ERK), while binding of insulin receptor substrate (IRS) leads to recruitment and activation of PI3K and to the activation of downstream kinases such as thymoma viral proto-oncogene (Akt), also known as protein kinase B (PKB). Phosphorylation and activation of PLCγ leads to the formation of inositol(1,4,5)triphosphate (IP3) and diacylglycerol (DAG) and to the stimulation of protein kinase C (PKC) and cellular Ca2+ pathways. These three main signaling pathways—PI3K, PLCγ, and MAPK/ERK—induced by Trk receptor activation are illustrated in Figure 2. Interestingly, there is evidence of differential activation of these three cascades depending on the neurotrophin, receptor type, and signal strength and duration involved (see (Segal, 2003). The differential activation of these downstream pathways seems particularly relevant to drug-induced changes in neuronal morphology and behavior, as will be detailed in later sections of this review.
Compared to the extensive knowledge of the consequences of Trk receptor activation, much less is known about the role of p75NTR signaling in neurotrophin function. Activation of the Trk effectors generally leads to pro-survival and differentiation signals, whereas the activation of p75NTR initiates both pro-survival and pro-death signaling cascades. Survival signaling through the p75NTR requires downstream Nuclear Factor kappa B (NFκB) which is thought to be activated indirectly through TNF (tumor necrosis factor) receptor-associated factor 4/6 (TRAF4/6) or receptor interacting protein 2 (RIP2) (for review see (Chao, 2003)). Although neurotrophin signaling allows for a complex variety of signals that depend on the expression pattern of neurotrophins and receptors and the processing of the neurotrophin peptides, this review focuses on drug-induced changes in neurotrophin signaling pathways downstream of BDNF.
Changes in levels of BDNF protein and mRNA have been examined in multiple brain regions following administration of many classes of addictive substances. Stimulants produce a widespread, but transient, induction of BDNF protein in the NAc, prefrontal cortex (PFC), VTA, and the central (CeA) and basolateral (BLA) nuclei of the amygdala (Graham et al., 2007; Grimm et al., 2003; Le Foll et al., 2005). Both contingent and non-contingent (i.e., animals yoked to self-administering animals) cocaine administration causes elevated levels of BDNF protein in the NAc (Graham et al., 2007; Liu et al., 2006; Zhang et al., 2002). Likewise, long-term withdrawal of up to 90 days after cocaine self-administration is correlated with increased BDNF protein in the NAc, VTA, and amygdala (Grimm et al., 2003; Pu et al., 2006), and there is early evidence that epigenetic regulation at the bdnf gene may be involved in mediating this persistant induction (Kumar et al., 2005).
Though fewer studies have been conducted to examine BDNF mRNA and protein levels after exposure to opiates, it appears that BDNF levels are regulated by opiates in certain reward-related brain regions. Acute morphine administration increases BDNF mRNA levels in the NAc, medial PFC (mPFC), VTA, and orbitofrontal cortex. In the VTA, chronic morphine, given via subcutaneous (s.c.) implants, is reported to be ineffective in altering BDNF mRNA expression (Numan et al., 1998). This, however, is in contrast to changes in BDNF protein observed after chronic morphine treatment. Using escalating doses of intraperitoneal (i.p.) morphine, it has been shown that the number of BDNF immunoreactive cells in the VTA is decreased (Chu et al., 2007), suggesting downregulation of BDNF function. Although no reports have examined BDNF expression in the hippocampus or caudate-putamen (CPu) after administration of stimulants or opiates, such studies are warranted since robust morphological changes have been observed in pyramidal neurons of the hippocampal CA3 region and medium spiny neurons (MSNs) of the CPu under these conditions ((Robinson and Kolb, 2004); see Table 1).
Several proteins in neurotrophin signal cascades have been shown to be regulated within the mesolimbic dopamine system by opiates and stimulants; these include drug effects on IRS–PI3K–Akt, PLCγ, Ras–ERK, and NFκB signaling (Figure 3). Stimulants dramatically increase ERK phosphorylation in numerous brain regions, including the NAc, VTA, and PFC, following acute or chronic drug administration (Jenab et al., 2005; Shi and McGinty, 2006, 2007; Sun et al., 2007; Valjent et al., 2004; Valjent et al., 2005). These findings are consistent with stimulant-induced increases in neuronal branching and spine number, given Ras–ERK’s established role in neurite outgrowth. The effects of opiates on ERK signaling are less clear. Recently, it has been reported that ERK phosphorylation is decreased in the NAc (Muller and Unterwald, 2004), PFC (Ferrer-Alcon et al., 2004), and VTA (unpublished observations) after chronic morphine, an effect that is consistent with decreased neurite branching seen in these regions in morphine-dependent animals. However, earlier work from our group and others reported increased ERK activity, including increased ERK phosphorylation and catalytic activity, in the VTA after chronic morphine (Berhow et al., 1996b; Liu et al., 2007; Ortiz et al., 1995). Further studies are needed to determine the explanation for these discrepant findings. Moreover, it is important to use multiple approaches to measure protein activity so that biochemical events can be correlated with morphological and behavioral endpoints. For example, inhibition of ERK in VTA dopamine neurons does not affect cell size (Russo et al., 2007), such that future studies are required to address the functional relevance of drug-induced changes in ERK activity in this and other brain areas as they relate to addictive phenotypes.
Several recent reports have shown that IRS–PI3K–Akt signaling is influenced by drugs of abuse (Brami-Cherrier et al., 2002; McGinty et al., 2008; Muller and Unterwald, 2004; Russo et al., 2007; Shi and McGinty, 2007; Wei et al., 2007; Williams et al., 2007). Chronic opiate administration decreases Akt phosphorylation in both the NAc and VTA (Muller and Unterwald, 2004; Russo et al., 2007). These biochemical alterations correspond to decreased neuronal branching and dendritic spine density or, in the case of VTA dopamine neurons, decreased cell body size (Diana et al., 2006; Robinson et al., 2002; Robinson and Kolb, 1999b; Russo et al., 2007; Spiga et al., 2005; Spiga et al., 2003)
The effects of stimulants on IRS–PI3K–Akt signaling in these regions are less clear. For example, chronic cocaine increases PI3K activity in the NAc shell and decreases its activity in the NAc core (Zhang et al., 2006). These data are in line with a previous report showing that chronic cocaine selectively increased BDNF mRNA levels in the NAc shell and decreased TrKB receptor mRNA in the NAc core (Filip et al., 2006). Thus, shell and core differences in PI3K activity could be explained by differential upstream regulation of BDNF and TrKB by cocaine. Interestingly, when a more general dissection of striatum is used (including NAc and CPu), it has been shown that amphetamine decreases Akt activity in synaptosome preparations (Wei et al., 2007; Williams et al., 2007), and we have observed similar effects of chronic cocaine in the NAc without distinguishing between core and shell (Pulipparacharuvil et al., 2008). Additionally, these studies are complicated by the time-course used to study Akt signaling changes, as recent work by McGinty and colleagues suggests that chronic amphetamine causes a transient and nuclear-specific change in Akt phosphorylation in striatum (McGinty et al., 2008). At early timepoints after amphetamine administration there is a nucleus-specific increase in Akt phosphorylation, however, after two hours Akt phosphorylation is decreased, suggesting a compensatory mechanism to turn off this activity. Understanding the dynamic relationship between stimulants and Akt signaling will be important to determine whether this signaling pathway is driving stimulant-induced structural plasticity in the NAc, as is the case for opiates in the VTA (see Section 6).
Alterations in the PLCγ and NFκB signaling pathways in drug abuse have not been as well studied as ERK and Akt; however, recent work shows that both pathways are regulated by drugs of abuse. Chronic administration of morphine increases total levels of PLCγ protein as well as levels of its activated tyrosine-phosphorylated form (Wolf et al., 2007; Wolf et al., 1999). Moreover, viral-mediated PLCγ overexpression in the VTA was found to increase ERK activity in this brain region (Wolf et al., 2007), thereby mimicking a similar increase in ERK activity seen after chronic morphine in earlier studies (Berhow et al., 1996b). PLCγ overexpression in the VTA also regulates opiate reward and related emotional behaviors, with distinct effects seen in rostral vs. caudal VTA (Bolanos et al., 2003). Likewise, Graham and colleagues (Graham et al., 2007) observed increased phosphorylation of PLCγ in the NAc following acute, chronic yoked, and chronic self-administered cocaine, an effect that was dependent on BDNF.
An earlier study from our group showed that the NFκB subunits p105, p65, and IκB are increased in the NAc in response to chronic cocaine administration (Ang et al., 2001). This is consistent with findings from Cadet and colleagues (Asanuma and Cadet, 1998), who demonstrated that methamphetamine induces NFκB binding activity in striatal regions. Given that some of the drug-regulated NFκB proteins activate NFκB signaling, whereas others inhibit it, it was unclear from these original studies whether the observed protein changes reflect an overall increase or decrease in NFκB signaling. We have more recently resolved this question by showing that chronic cocaine administration upregulates NFκB transcriptional activity in the NAc, based on findings in NFκB-LacZ transgenic reporter mice (Russo, Soc. Neurosci. Abstr. 611.5, 2007). More recent evidence has directly implicated the induction of NFκB signaling in the NAc in the structural and behavioral effects of cocaine (see Section 6). These early findings are intriguing and warrant further exploration including an examination of the effect of opiates on NFκB signaling in brain reward regions.
The brain’s reward circuitry has evolved to direct one’s resources to obtain natural reward, but this system can be corrupted or hijacked by drugs of abuse. Within this circuit, structural plasticity is generally characterized by altered dendrite branching or arborization and by changes in the density or morphometry of dendritic spines. Although the direct behavioral relevance of experience-dependent morphological changes is still under investigation, it is believed that synaptic function is determined not only by the number, but also the size and shape of each individual spine head. As spines form, they send out thin immature structures which take on either stubby, multisynaptic, filopodial, or branched shapes (for review see (Bourne and Harris, 2007; Tada and Sheng, 2006). In the adult brain, under basal conditions, it is estimated that at least 10% of spines have these immature shapes suggesting that plasticity is a continuous process throughout life (Fiala et al., 2002; Harris, 1999; Harris et al., 1992; Peters and Kaiserman-Abramof, 1970). These structures are transient and can form within hours of stimulation and persist as long as a few days in vivo (Holtmaat et al., 2005; Majewska et al., 2006; Zuo et al., 2005).
It is believed that stabilization of a transient, immature spine into a more permanent, functional spine occurs through an activity-dependent mechanism (for review see (Tada and Sheng, 2006). Stimulation protocols which induce long-term depression (LTD) are associated with shrinkage or retraction of spines on hippocampal and cortical pyramidal neurons (Nagerl et al., 2004; Okamoto et al., 2004; Zhou et al., 2004), whereas induction of long-term potentiation (LTP) is associated with the formation of new spines and enlargement of existing spines (Matsuzaki et al., 2004; Nagerl et al., 2004; Okamoto et al., 2004). At a molecular level, it is believed that LTP and LTD initiate changes in signaling pathways, and in the synthesis and localization of proteins, which eventually alter the polymerization of actin to affect spine maturation and stability and ultimately to produce a functional spine (LTP) or retraction of an existing spine (LTD) (for review see (Bourne and Harris, 2007; Tada and Sheng, 2006). Upon stabilization, spines become mushroom shaped, have larger postsynaptic densities (Harris et al., 1992), and have been shown to persist for months (Holtmaat et al., 2005; Zuo et al., 2005). These changes reflect a highly stable cellular event that may be a plausible explanation for at least some of the long-term behavioral changes associated with drug addiction.
Most classes of addictive substances, when administered chronically, alter structural plasticity throughout the brain’s reward circuitry. Most of these studies are correlative and associate structural changes in specific brain regions with a behavioral phenotype indicative of addiction. Over the past decade, Robinson and colleagues have led the way in understanding how drugs of abuse regulate structural plasticity (for review see (Robinson and Kolb, 2004). Since these original observations, other drug abuse researchers have added to this growing literature to uncover drug class-specific effects on neuronal morphology. As depicted in Table 1 and Figure 3, opiates and stimulants differentially affect structural plasticity. Opiates have been shown to decrease the number and complexity of dendritic spines on NAc MSNs and mPFC and hippocampus pyramidal neurons, and to decrease the overall soma size of VTA dopaminergic neurons, with no effect seen on non-dopaminergic neurons in this brain region (Nestler, 1992; Robinson and Kolb, 2004; Russo et al., 2007; Sklair-Tavron et al., 1996). To date, there is a single exception to these findings, where it was reported that morphine increases spine number on orbitofrontal cortical neurons (Robinson et al., 2002). In contrast to opiates, stimulants such as amphetamine and cocaine have been shown to consistently increase dendritic spines and complexity in NAc MSNs, VTA dopaminergic neurons, and PFC pyramidal neurons, with no reports of decreased structural plasticity (Lee et al., 2006; Norrholm et al., 2003; Robinson et al., 2001; Robinson and Kolb, 1997, 1999a; Sarti et al., 2007).
Although the molecular mechanisms downstream of neurotrophic factor signaling which underlie these changes are poorly understood, many of these structural changes are accompanied by alterations in levels or activity of proteins well known to regulate the neuronal cytoskeleton. These include, but are not limited to, drug-induced changes in microtubule associated protein 2 (MAP2), neurofilament proteins, activity-regulated cytoskeletal-associated protein (Arc), LIM-kinase (LIMK), myocyte enhancer factor 2 (MEF2), cyclin-dependent kinase s5 (Cdk5), postsynaptic density 95 (PSD95), and cofilin, as well as changes in actin cycling, in the NAc or other brain reward regions (Beitner-Johnson et al., 1992; Bibb et al., 2001; Chase et al., 2007; Marie-Claire et al., 2004; Pulipparacharuvil et al., 2008; Toda et al., 2006; Yao et al., 2004; Ziolkowska et al., 2005). Since many of the biochemical changes induced by stimulants and morphine are similar, it will be important to identify distinct opiate- and stimulant-regulated gene targets related to dendritic function, as they may provide insight into the generally opposite effects of opiate and stimulants on neurotrophic factor-dependent structural plasticity.
The opposite morphological changes induced in brain reward regions by opiates and stimulants are paradoxical since the two drugs cause very similar behavioral phenotypes. For example, specific treatment regimens of opiates and stimulants, both of which result in locomotor sensitization and similar patterns of escalation of drug self-administration, cause opposite changes in dendritic spine density in the NAc (Robinson and Kolb, 2004). Thus, if these morphological changes are important mediators of addiction, either they must have bidirectional properties, whereby a change from baseline in both directions produces the same behavioral phenotype, or they mediate distinct behavioral or other phenotypes that are not captured with the experimental tools used. Additionally, these findings must be considered in the context of the drug administration paradigm in question. In our studies, for instance, animals receive high doses of subcutaneous morphine, continuously released from pellet implants, a paradigm more consistent with opiate tolerance and dependence. In contrast, most stimulant paradigms utilize once to several times daily injections of the drug, allowing blood levels to peak and return to baseline before the next administration, paradigms more consistent with drug sensitization. Patterns of opiate and stimulant use by humans can vary widely from person to person. Therefore, future studies will need to address the behavioral relevance of drug-induced morphological changes in brain reward regions in the context of dose and drug administration paradigms that mirror exposures seen in humans.
Changes in growth factor signaling are hypothesized to be a major factor influencing the structural and behavioral plasticity associated with drug addiction. Human studies are limited. Drug-induced changes in serum BDNF have been observed in humans addicted to cocaine, amphetamine, alcohol, or opiates (Angelucci et al., 2007; Janak et al., 2006; Kim et al., 2005), yet the source of this BDNF, and the relevance of these changes to the onset and maintenance of addiction have remained unclear. It would be interesting in future studies to examine BDNF and its signaling pathways in human postmortem brain tissue.
Over the past decade, work in rodents has established the influence of BDNF on various phases of the addiction process. Early studies showed that local infusion of BDNF into the VTA or NAc augments locomotor and rewarding responses to cocaine, while global loss of BDNF exerts the opposite effects (Hall et al., 2003; Horger et al., 1999; Lu et al., 2004). More recent work has shown that cocaine self-administration increases BDNF signaling in the NAc (Graham et al., 2007). In addition, an intra-NAc infusion of BDNF potentiates cocaine self-administration and cocaine-seeking and relapse, while infusion of antibodies against BDNF, or local knockout of the bdnf gene in the NAc (achieved via viral expression of Cre recominase in floxed BDNF mice), blocks these behaviors. Based on these studies, Graham and colleagues (2007) concluded that BDNF release in the NAc during the initiation of cocaine self-administration is a necessary component of the addiction process.
These data support the view that BDNF is a candidate molecule to mediate the structural changes in NAc neurons produced by chronic exposure to cocaine or other stimulants. According to this hypothesis, stimulant-induced increases in BDNF signaling in the NAc would induce an increase in dendritic arborization of NAc neurons, which would underlie sensitized behavioral responses to the stimulants as well as strong drug-related memories crucial for relapse and addiction. Consistent with this hypothesis are findings from cultured hippocampal neurons, where it has been shown that BDNF secretion induces protein synthesis-dependent enlargement of individual dendritic spines (Tanaka et al., 2008b). The weakness of this hypothesis is that there has been no direct experimental evidence that enhancement of dendritic spines of NAc neurons per se is necessary or sufficient for sensitized drug responses. In fact, there are data that suggest a more complex relationship between the two phenomena: inhibition of Cdk5 in the NAc blocks the ability of cocaine to increase dendritic spines on NAc neurons, despite the fact that such inhibition potentiates locomotor and rewarding responses to cocaine (Norrholm et al., 2003; Taylor et al., 2007). Clearly, further work is needed to study the relationship between this structural and behavioral plasticity.
Another important caveat to this hypothesis is that changes in BDNF signaling may produce profoundly different effects on neuronal morphology and behavior depending on the brain region examined. Recent reports have drawn clear distinctions between BDNF function in the hippocampus versus VTA (Berton et al., 2006; Eisch et al., 2003; Krishnan et al., 2007; Shirayama et al., 2002): BDNF infusions in the hippocampus are antidepressant-like, whereas infusions of BDNF in the VTA or NAc produce prodepressant-like effects. Similar patterns are emerging in the addiction field. Notably, increased BDNF in the NAc enhances cocaine-induced behaviors (Graham et al., 2007; Horger et al., 1999), whereas in the PFC BDNF suppresses these same behaviors (Berglind et al., 2007). Not surprisingly, the induction of BDNF by cocaine is also differentially regulated in these two brain regions, a pattern which further substantiates the behavioral differences (Fumagalli et al., 2007).
Preliminary evidence has implicated NFκB signaling in the regulation of cocaine-induced structural and behavioral plasticity. Although the direct mechanism by which these changes occur is unknown, previous work has shown that the p75NTR, which is upstream of NFκB, is localized at the synapse and that p75NTR activation by BDNF is necessary for LTD. Although BDNF-TrkB interactions have been extensively studied in drug abuse, these data suggest an alternative pathway through NFκB that warrants further investigation. In line with this hypothesis, we have recently observed that viral-mediated overexpression of a dominant negative antagonist of the NFκB pathway in the NAc prevents the ability of chronic cocaine to increase the density of dendritic spines on NAc MSNs. Such inhibition of NFκB signaling also blunts sensitization to the rewarding effects of cocaine (Russo, Soc. Neurosci. Abstr. 611.5, 2007). These data, unlike the situation for Cdk5 cited above, support a link between increased dendritic arborization and behavioral sensitization to cocaine, further emphasizing the complexity of these phenomena and the need for further study.
Although limited work has addressed the relevance of neurotrophic factor signaling in opiate-induced behaviors, work from our laboratory has uncovered a role for BDNF and the downstream IRS2-PI3K-Akt pathway in the regulation of VTA dopaminergic cell size and subsequent reward tolerance (Russo et al., 2007; Sklair-Tavron et al., 1996). Specifically, chronic opiate administration in rodents produces a state of reward tolerance and physical dependence during relatively early periods of withdrawal that is thought to contribute to an escalation of drug-taking behavior. Early experiments found that intra-VTA infusion of BDNF prevents the morphine-induced decrease in VTA neuron size (Sklair-Tavron et al., 1996). More recently, we have shown that the timeline of reward tolerance, as measured by conditioned place preference, parallels the timeline of reduced dopaminergic cell size and that these phenomena are mediated via BDNF signaling cascades (Russo et al., 2007). As mentioned earlier, the biochemical signaling pathways in the VTA that are downstream of BDNF and the TrKB receptor are differentially regulated by chronic morphine: morphine activates PLCγ(Wolf et al., 2007; Wolf et al., 1999), decreases activity of the IRS–PI3K–Akt pathway (Russo et al., 2007; Wolf et al., 1999), and produces variable effects on ERK (see above). In light of recent evidence that Akt regulates the size of many cell types in the central nervous system (Backman et al., 2001; Kwon et al., 2006; Kwon et al., 2001; Scheidenhelm et al., 2005), we utilized viral gene transfer techniques to directly show that morphine produces reward tolerance through inhibition of the IRS2–PI3K–Akt pathway and reduced size of VTA dopamine neurons. These effects were not observed by altering ERK or PLCγ signaling, again pointing to the importance of IRS–PI3K–Akt signaling for this phenomenon. Future studies will address the relevance of BDNF and IRS–PI3K–Akt pathways in the escalation of opiate self-administration, a more clinically relevant paradigm to measure addiction. A greater understanding of the upstream changes in neurotrophic factors or their receptors and downstream targets of Akt will address the specific mechanisms of opiate reward tolerance in addiction models. Moreover, it will be important to understand the role of BDNF signaling in the regulation of VTA function within a neural circuit context. In this regard, it is interesting to note that Pu et al. (2006) showed that following withdrawal from repeated cocaine exposure, excitatory synapses onto dopamine neurons in the VTA are more responsive to potentiation by weak presynaptic stimuli, an effect requiring endogenous BDNF-TrkB signaling.
While the above discussion focuses on BDNF and its signaling cascades, there is evidence that several other neurotrophic factors and their downstream signaling pathways also influence behavioral or biochemical responses to drugs of abuse. NT3, like BDNF, has been shown to promote sensitized responses to cocaine at the level of the VTA (Pierce and Bari, 2001; Pierce et al., 1999). Chronic administration of morphine or cocaine up-regulates glial cell line-derived neurotrophic factor (GDNF) signaling in the VTA-NAc circuit, which in turn feeds back and suppresses the behavioral effects of these drugs of abuse (Messer et al., 2000). Amphetamine induces basic fibroblast growth factor (bFGF) in the VTA-NAc circuit and bFGF knockout mice have a blunted response to locomotor sensitization induced by repeated amphetamine injections (Flores et al., 2000; Flores and Stewart, 2000). The cytokine, ciliary neurotrophic factor (CNTF), administered directly into the VTA, enhances the ability of cocaine to induce biochemical adaptations in this brain region; cocaine increases intracellular signaling cascades through Janus kinase (JAK) and signal transducers and activators of transcription (STATs), an effect that was potentiated by an acute infusion of CNTF (Berhow et al., 1996a). There is also evidence that chronic morphine alters levels of insulin-like growth factor 1 (IGF1) in the VTA and other brain regions (Beitner-Johnson et al., 1992). These isolated findings indicate that a diverse array of neurotrophic mechanisms control VTA-NAc function to regulate plasticity to drugs of abuse in complex ways and highlight the need for much future research in this area.
Over the past decade, we have expanded our understanding of how drugs of abuse regulate neurotrophic signaling pathways and the morphology of diverse neuronal populations throughout the brain’s reward circuitry. Recent advances in viral gene transfer allow for manipulations of specific downstream neurotrophic signaling proteins within a given brain region of interest of fully developed adult animals to study the relationships among drug abuse, neuronal morphology, and behavioral plasticity. With novel bicistronic viral vectors, it is possible to express a protein that manipulates neurotrophic signaling pathways as well as a fluorescent protein to visualize neuronal morphology (Clark et al., 2002). Thus, with improved immunohistochemical techniques to label specific neuronal populations, it is possible to assess drug-induced morphological changes and associated biochemical adaptations in neurotrophic signaling in a cell type-specific manner, and therefore provides crucial information for drug-induced regulation of heterogeneous brain reward regions. Using multidisciplinary approaches with behavioral, physiological, biochemical, and morphological endpoints, it will be increasingly possible to define the mechanisms of addiction with far greater precision, including the precise role of neurotrophic factor signaling in experience-dependent plasticity and the addiction process. This knowledge may lead to the development of novel medical interventions to normalize the maladaptive plasticity induced by drugs of abuse in brain reward regions and to thereby reverse the addiction process in humans.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.