Neuropharmacology 47 (2004) 33–46 www.elsevier.com/locate/neuropharm

Structural plasticity associatedwith exposure to drugsof abuse Terry E. Robinson a,, Bryan Kolb b a Department of Psychology (Biopsychology) and Neuroscience Program, The University of Michigan, 525 E. University (East Hall), Ann Arbor, MI 48109, USA b Department of Psychology and Neuroscience, Canadian Centre for Behavioural Neuroscience, University of Lethbridge, Lethbridge, Alta., Canada T1K 3M4

Received19 April 2004; receivedin revisedform 24 May 2004; accepted30 June 2004

Abstract

Persistent changes in behavior andpsychological function that occur as a function of experience, such those associatedwith learning andmemory, are thought to be dueto the reorganization of synaptic connections (structural plasticity) in relevant brain circuits. Some of the most compelling examples of experience-dependent changes in behavior and psychological function, changes that can last a lifetime, are those that accrue with the development of . However, until recently, there has been almost no research on whether potentially addictive drugs produce forms of structural plasticity similar to those associated with other forms of experience-dependent plasticity. In this paper we summarize evidence that, indeed, exposure to , cocaine, or morphine produces persistent changes in the structure of dendrites and dendritic spines on cells in brain regions involvedin incentive motivation andreward(such as the ), andjudgmentandthe inhibitory control of beha- vior (such as the prefrontal cortex). It is suggestedthat structural plasticity associatedwith exposure to drugsof abuse reflects a reorganization of patterns of synaptic connectivity in these neural systems, a reorganization that alters their operation, thus con- tributing to some of the persistent sequela associated with drug use—including . # 2004 Elsevier Ltd. All rights reserved.

Keywords: Amphetamine; Cocaine; Morphine; Nicotine; Psychostimulants; Opiates; Dendrites; Dendritic spines; Sensitization; Golgi; Synaptic plasticity

1. Introduction where experience has been shown to alter the physical structure of andsynapses (i.e., producestruc- Persistent changes in behavior andpsychological tural plasticity). For example, changes in behavior that function that occur as a consequence of experience are result from learning (Chang andGreenough, 1982; thought to be mediated by the reorganization or Moser et al., 1994; Stewart andRusakov, 1995; Leuner strengthening of synaptic connections in specific neural et al., 2003), living in an isolatedversus complex circuits. This idea has been a fundamental assumption environment (Greenough et al., 1990; van Praag et al., underlying research on the neurobiology of learning 2000; Kolb et al., 2003a) or recovery of function after andmemory, as well as other forms of experience- brain damage (Kolb andGibb, 1991; Jones et al., 1996; dependent plasticity, since the time of Ramon y Cajal Biernaskie andCorbett, 2001 ) are all associatedwith (1928), andHebb (1949) formalizedthis postulate in structural alterations in relevant neural circuits. Thus, his seminal book, ‘‘The Organization of Behavior’’. a major aim of modern research on the neurobiology Although experimental evidence directly relating of behavioral plasticity, including learning and mem- structural plasticity in the brain to changes in specific ory, is elucidating the molecular mechanisms involved behaviors is very limited, there are numerous examples in the structural reorganization of neuronal circuits (Lamprecht andLeDoux, 2004 ). Corresponding author. Tel.: 1-734-763-4361; fax: +1-734-763- A focus of much research on structural plasticity has 7480. been on the morphology of dendrites and dendritic E-mail address: [email protected] (T.E. Robinson). spines. The vast majority of synaptic inputs onto

0028-3908/$ - see front matter # 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.neuropharm.2004.06.025 34 T.E. Robinson, B. Kolb / Neuropharmacology 47 Supplement No. 1 (2004) 33–46 neurons are on dendrites or dendritic spines, and the function. Furthermore, there is growing evidence that amount of synaptic input cells receive varies with the drugs of abuse usurp many of the same cellular and amount of dendritic surface available (Harris and molecular mechanisms involvedin other forms of Kater, 1994). Furthermore, it is estimatedthat over synaptic plasticity (Berke andHyman, 2000; Hyman 90% of excitatory synapses are on dendritic spines, and andMalenka, 2001; Nestler, 2001 ). It is surprising, synaptogenesis associatedwith experiences like learning therefore, that until recently there has been almost no or environmental complexity is reflectedby changes in research on whether any of the long-lasting behavioral the number of dendritic spines (Greenough et al., 1990; consequences of repeatedexposure to drugsof abuse in Kolb et al., 1998; Woolley, 1999; Rampon et al., 2000). adulthood are accompanied by the kinds of structural Even changes in the shape of existing spines may mod- plasticity normally associatedwith other forms of ify synaptic efficacy by altering the chemical micro- experience-dependent plasticity. (Note that structural environment or the electrotonic properties of the changes in dendrites have been described when drugs synapse (Horner, 1993; Rusakov et al., 1996; Shepherd, are given early in development, but this literature will 1996; Nimchinsky et al., 2002; Tsay andYuste, 2004 ), not be reviewedhere; see StanwoodandLevitt, 2004 or by altering fast synaptic neurotransmission (Kasai for a recent review). et al., 2003). Indeed, dendrites and dendritic spines are We have begun to address this issue in a series of thought to be a locus of experience-dependent structur- experiments in which we askedwhether repeated al plasticity (Harris andKater, 1994; Nimchinsky et al., exposure to cocaine, amphetamine, morphine or nic- 2002; Kasai et al., 2003; Lamprecht andLeDoux, otine, in adult rats, whether administered by an 2004), andtherefore, they providean obvious focus of experimenter (EA) or self-administered (SA), have study in trying to understand how experiences can alter long-lasting effects on the structure of dendrites and brain organization to produce life-long changes in dendritic spines in brain regions thought to mediate behavior andpsychological function. drug-induced changes in incentive motivation and Nearly all research on structural plasticity in the reward(such as the nucleus accumbens; Acb) andin brain has involvedhow learning ( Greenough and cognitive function (such as the prefrontal cortex; PFC). Bailey, 1988; Andersen and Trommald, 1995; Kolb and The purpose of this paper is to summarize what we Whishaw, 1998; Lamprecht andLeDoux, 2004 ), long- have learnedthus far. term potentiation (Andersen and Soleng, 1998; Yuste andBonhoeffer, 2001 ), stress (McEwen, 2000; Vyas et al., 2002), environmental manipulations (Greenough 2. The method et al., 1990; van Praag et al., 2000), recovery of function (Kolb andWhishaw, 1998 ), changes in the A common approach to examine the impact hormonal milieu (Woolley, 1999; Leranth et al., 2003), of experience on synaptic organization is to use pathological states (Fiala et al., 2002), etc., change Golgi-stainedmaterial to quantify the structure of den- synapses or dendritic structure. But one of the most drites and the density of dendritic spines (Greenough, compelling examples of experience-dependent plas- 1984; Greenough et al., 1990; Kolb et al., 1998) and ticity, whereby experience at one point in life changes that is the methodusedin most of the studies summar- behavior andpsychological function for a lifetime, is ized here. In all cases, the structure of dendrites or the addiction. The propensity of addicts to relapse, even density of spines on neurons was quantified using one months to many years after the discontinuation of drug of three measures (Greenough andChang, 1985; Kolb use, andlong after withdrawalsymptoms have sub- andWhishaw, 1998 ). (1) Total dendritic length was sided, provides stark evidence that drug use has long- estimatedby counting the number of ring intersections lasting consequences for behavior andpsychological using an overlay of concentric rings (Sholl, 1981). (2) function. Similarly, very long-lasting changes in The total number of dendritic branches (indicated by behavior produced by repeated exposure to drugs of bifurcations) was countedat each orderaway from the abuse have been described in controlled animal studies, cell body (Coleman andRiesen, 1968 ). (3) Spine den- as exemplified, for example, by phenomena like sity was estimated along a specific segment of dendrite behavioral sensitization (Robinson andBecker, 1986 ). by tracing the dendritic segment, calculating its exact Repeatedintermittent exposure to a variety of drugs of length, andcounting the number of spines along that abuse can produce a hypersensitivity (sensitization) to length (to yieldspines/10 lm). In some experiments, their psychomotor activating andincentive motiva- the frequency of branchedspines (i.e., spines with mul- tional effects that can persist for months to years after tiple heads) was also quantified (Comery et al., 1996). the discontinuation of drug treatment (Paulson et al., There are, of course, limitations in interpreting chan- 1991; Robinson andBerridge, 2003 ). ges in dendritic structure estimated from Golgi material The behavioral evidence leaves no doubt that drugs that need to be kept in mind. This approach does not of abuse can produce very persistent changes in brain provide a direct measure of synapses, but only an T.E. Robinson, B. Kolb / Neuropharmacology 47 Supplement No. 1 (2004) 33–46 35 indirect measure based on alterations in the post- 3. Repeated exposure to drugs of abuse produces synaptic surface of cells. However, a strong relation- long-lasting changes in the structure of dendrites ship between measures of dendritic structure and and dendritic spines synapses has been confirmedin many studies using electron microscopy to directly quantify synaptic den- The most extensive data available are from studies sity. Typically changes in the dendritic surface of cor- with the psychomotor stimulant drugs, amphetamine tical neurons assessedusing Golgi-stainedmaterial are andcocaine. For these drugs,the effects of both exper- accompaniedby changes in the number of synapses per imenter-administered (EA) and self-administered (SA) assessedwith EM ( Greenough andBailey, drug have been studied, primarily in the nucleus accumbens (Acb) andmedialprefrontal cortex (mPFC) 1988; Greenough et al., 1990; Kolb andWhishaw, (Table 1). EA andSA cocaine andamphetamine have 1998; Woolley, 1999). Of course, without ultra- very similar effects on the density of spines in both structural studies one cannot be positive that changes the Acb andmPFC. Both amphetamine ( Robinson in dendritic surface and spines are accompanied by an andKolb, 1997, 1999a; Ferrario et al., 2003; Heijtz increase in synaptic contacts, but normally in adult rats et al., 2003; Kolb et al., 2003b; Li et al., 2003; nearly all spines in the cortex andstriatum have a syn- Crombag et al., 2004) andcocaine ( Robinson and aptic contact (Gray, 1959; Peters andFeldman,1976; Kolb, 1999a; Robinson et al., 2001; Li andRobinson, Wilson et al., 1983). 2003; Norrholm et al., 2003), and both modes of drug It is also important to emphasize that the kinds of administration, increase spine density on medium spiny structural alterations in neurons seen at this level of neurons in both the Acb shell (AcbS) andthe Acb core analysis provide no information about exactly how the (AcbC; amphetamine SA data not available for AcbC, operation of cells or circuits is altered. Indeed, very dif- see Table 1). Where data are available (Table 2) these ferent changes in structure, such as an increase versus a increases in spine density in the Acb are accompanied decrease in spine density on a given dendritic segment by increases in dendritic branching (Robinson and of a cell, couldhave exactly the same outcome in terms Kolb, 1997, 1999a; Robinson et al., 2001; Kolb et al., of how cell signaling andthe operation of the circuit is 2003b). Similarly, both amphetamine (Robinson and altered—depending on how different synaptic inputs Kolb, 1997, 1999a; Ferrario et al., 2003; Heijtz et al., are rearrangedaroundthe alteredpostsynaptic surface. 2003; Crombag et al., 2004) andcocaine ( Robinson By the same token, apparently similar alterations in andKolb, 1999a; Robinson et al., 2001; Ferrario et al., dendrites produced by different treatments could have 2003) and both modes of administration, increase spine very different outcomes for the operation of cells and density (and dendritic branching where studied) on the circuits, if synaptic inputs are rearrangeddifferently. apical dendrites of pyramidal cells in the mPFC. Simi- Thus, it is impossible to tell how dendritic reorganiza- lar effects are seen on the basilar dendrites of mPFC tion characterizedat this level of structural analysis pyramidal cells, but the effect of amphetamine here is translates into alterations in properties of cells much weaker than that of cocaine (Robinson and (although see Kasai et al., 2003). Some information Kolb, 1999a). The effects of EA nicotine are very simi- may be gleanedfrom ultrastructural studies, which lar to the psychostimulants: nicotine increases spine have the potential to tell how different inputs are rear- density in both the AcbS and mPFC (Brown andKolb, ranged, but even this will not tell exactly how the syn- 2001; Gonzalez et al., 2004). Deprenyl, which is meta- aptic rearrangement changedcell signaling. This will bolizedinto amphetamine, also is reportedto increase require electrophysiological approaches. Nevertheless, dendritic branching in prefrontal pyramidal cells in as described below, the approach used here can provide Bonnett monkeys (Shankaranarayana Rao et al., 1999). a great deal of information about the structural plas- In sharp contrast, both EA andSA morphine mark- ticity associatedwith exposure to drugsof abuse— edly decrease spine density in the AcbS and the mPFC, including, what brain regions are affected, which cells andwhere studied(following EA; Table 3) this is and which portions of the dendritic surface are altered, accompanied by a decrease in dendritic branching whether different drugs have similar or different effects, (Robinson andKolb, 1999b; Robinson et al., 2002 ). the conditions that lead to structural plasticity (the Chronic morphine treatment andwithdrawal also doses, treatment conditions, mode of drug administra- alters the morphology of VTA neurons tion, etc.), whether there is any relationship between (Sklair-Tavron et al., 1996; Spiga et al., 2003). Thus, persistent structural changes andpersistent behavioral (1) exposure to four different drugs of abuse (cocaine, changes produced by exposure to drugs of abuse, and amphetamine, nicotine andmorphine) producesstruc- how drug-induced structural plasticity interacts with tural plasticity; (2) the effects of the three stimulants the structural plasticity associatedother kindsof life are very similar (but not identical, see below)—they experience. So, what have we learned? increase spines in the Acb andmPFC; (3) the effect of 36 T.E. Robinson, B. Kolb / Neuropharmacology 47 Supplement No. 1 (2004) 33–46

the narcotic, morphine, is very different—it decreases spine density in the Acb and mPFC.

3.1. Persistence

For all drugs studied thus far structural changes are evident long after the discontinuation of drug treat- ment. Amphetamine- andnicotine-inducedincreases in spine density are seen up to 3.5 months after the last drug treatment (Kolb et al., 2003b; Gonzalez et al., 2004), although most studies have been conducted approximately one month after the last treatment (Robinson andKolb, 1997, 1999a ). For cocaine, spine changes have been foundas little as 24–48 h after the last of 28 daily injections (Norrholm et al., 2003), and as long as 2 weeks (Li et al., 2004) to a month later (Robinson andKolb, 1999a; Robinson et al., 2001 ). Similarly, the structural plasticity associatedwith mor- phine treatment has been described a month after the last treatment (Robinson andKolb, 1999b; Robinson et al., 2002). Interestingly, in one experiment, rats were given a cocaine treatment regimen that produced only relatively transient behavioral sensitization; i.e., psy- chomotor sensitization dissipated by 3 months after the last treatment. In this case spine changes were also no longer evident after 3 months of withdrawal (Kolb et al., 2003b). This suggests that, depending on the treatment regimen, spine changes may be reversible, and that they wax and wane as drug-induced beha- vioral plasticity waxes andwanes.

3.2. Treatment regimens

There have been no systematic studies on the dose or number of drug treatments required to produce this form of structural plasticity, but we can glean some information from the variety of treatment regimens that have been used. In most cases fairly aggressive treatment regimens have been used[e.g., for ampheta- mine, 5 weeks of treatment with doses escalating from 1 to 8 mg/kg (Robinson andKolb, 1997 ); 20 injections of 3 mg/kg (Robinson andKolb, 1999a )—for cocaine, 20 injections of 15 mg/kg (Robinson andKolb, 1999a)]. But such aggressive dosing regimens do not seem to be necessary to produce structural plasticity. For example, a single injection of 2 mg/kg of ampheta- mine produces a very small, but significant, increase in spine density in the AcbS, although the effects of repeatedtreatments are much greater ( Kolb et al., Fig. 1. Camera lucida drawings (courtesy of Grazyna Gorny) of the 2003b). Also, eight daily injections of 15 mg/kg major cell types discussed in this paper (not drawn to scale). A high of cocaine are sufficient to see spine changes when power illustration of a distal dendrite on a medium spiny neuron animals are studied 2 weeks after the last treatment illustrating dendritic spines is also shown at the bottom right. Zilles’ (Li et al., 2004). (1985) terminology: Cg3 (cingulate cortex, area 3; medial frontal cor- tex [mPFC]); AID (agranular insular cortex; which we refer to as In general, animals allowed to self-administer drug orbital frontal cortex [oPFC]); CA1 (fieldCA1 of Ammon’s horn take much greater amounts than usedin studieswhere [hippocampus]); Dentate (dentate gyrus); Acb (nucleus accumbens). drug is EA, and therefore, this approach may lead to T.E. Robinson, B. Kolb / Neuropharmacology 47 Supplement No. 1 (2004) 33–46 37

Table 1 Effects of stimulant drugs on spine density

Acb mPFC oPFC Par1 Oc1 CS AB A B AB AB

Cocaine EA "" "" –– –– –– SA "" "" NC NC "" NC # Amphetamine EA "" "NC - - ## ## SA – """##–– –– Nicotine EA – """– – NC NC – – Abbreviations: EA, experimenter-administered; SA, self-administered; C, core; S, shell; A, apical; B, basilar; NC, no change; –, no data; ", increase; #, decrease (see text for references).

Table 2 Effects of stimulant drugs on dendritic branching

Acb mPFC Par1 Oc1 CSA BA B AB

Cocaine EA – "" "–– –– SA – "" "NC " NC NC Amphetamine EA """ NC NC NC NC NC Nicotine EA – " NC " NC NC – – Abbreviations: see Table 1.

Table 3 Effects of morphine in spine density and dendritic branching

Acb mPFC oPFC Par1 Oc1 HPC C S A B A B A B A B CA1 DG

Spines EA – ###""##NC # NC NC SA – ###""NC NC ### # Branches EA – ###––##––– – Abbreviations: see Table 1. especially robust structural plasticity. But even here the possibly changedas a consequence of drugexperience. amount of drug intake has an effect. For example, rats This issue has been addressed for medium spiny neu- given 6 h of daily access to intravenous cocaine over rons in the Acb andthe caudate-putamen (CPu). These approximately 20 days, by which time they have dra- neurons receive different inputs onto the proximal ver- matically escalatedtheir intake ( AhmedandKoob, sus distal portions of their dendritic tree (Sesack and 1998), show a significantly greater increase in spine Pickel, 1990; Smith andBolam, 1990; Groenewegen density in the AcbC than animals given 1 h of daily et al., 1991). The distal dendrites are the primary locus access (who show a stable pattern of intake; unpub- of DA inputs arising from the midbrain, as well as lishedstudies). In conclusion, although there are no excitatory (presumably glutamate) inputs from the neo- systematic studies, it seems likely that both the magni- cortex, hippocampus andamygdala. Typically, excit- tude of the structural changes and their persistence atory inputs form asymmetric synapses on the heads of depend on dose, the number of drug treatments and spines, whereas the DA inputs make symmetric con- even the pattern of intake. tacts with the neck of spines or the dendritic shaft, forming the so-called‘‘triads’’that are thought to 3.3. Locus provide the means by which DA modulates the excit- atory drive on these cells. It is interesting, therefore, The locus of structural alterations on neurons pro- that changes in both dendritic branching (Robinson vides some information about what synaptic inputs are andKolb, 1999a ) andspine density ( Li et al., 2003) 38 T.E. Robinson, B. Kolb / Neuropharmacology 47 Supplement No. 1 (2004) 33–46 produced by amphetamine or cocaine are confined to reportedthat, ‘‘Each of the branches ... made contact distal dendrites of medium spiny neurons in both the with a standard perforant path bouton and showed all Acb andCPu. Furthermore, in the Acb there is not the normal attributes of an asymmetric excitatory spine only an increase in the number of spines, but an synapse’’ (pp. 224–5), and, ‘‘after reconstruction of 27 especially large increase in the number of spines with bifurcating spines no case was foundin which the same multiple heads (Robinson andKolb, 1997, 1999a ; see two spine heads were served by the same axon’’ (see below). Thus, the morphological evidence suggests that Andersen and Soleng, 1998 for a review). We do not prior exposure to psychostimulant drugs may change know if the same holds true for branched spines in the synaptic efficacy primarily on that portion of medium Acb, but if it does, branched spines may reflect a fun- spiny neurons that integrates DA andglutamate signal- damental reorganization of synaptic inputs onto the ing. This is especially important because both DA and distal dendrites of medium spiny neurons in the Acb as glutamate have been prominently implicatedin persist- a consequence of past drug experience. ent forms of drug experience-dependent behavioral plasticity, such as behavioral sensitization (Kalivas, 1995; Wolf, 1998), as well as in drug experience- 4. Different drugs produce different effects dependent changes in synaptic signaling measured on dendritic structure in different regions electrophysiologically (Thomas et al., 2001; Li and It is already obvious from the above that different Kauer, 2004). Of course, these same transmitters have drugs have different effects on dendritic structure. The been implicatedin other forms of synaptic plasticity most striking differences are between the stimulants, all andin learning ( Berke andHyman, 2000; Hyman and of which increase spine density in the Acb and mPFC, Malenka, 2001; Lamprecht andLeDoux, 2004 ). versus morphine, which decreases spine density in these 3.4. Branched spines regions. Nevertheless, it is worth emphasizing that even closely relateddrugslike amphetamine andcocaine, In a number of experiments there was an especially although similar in many respects, do not produce intriguing change in the shape of dendritic spines on identical patterns of structural plasticity. For example, medium spiny neurons in the Acb following treatment we have consistently foundthat in adult rats EA with amphetamine or cocaine (Robinson andKolb, cocaine increases spine density on the basilar dendrites 1997, 1999a). Amphetamine or cocaine approximately of pyramidal cells in the mPFC (Robinson andKolb, doubled the proportion of spines with multiple heads 1999a; Robinson et al., 2001; Ferrario et al., 2003; Li (branchedspines). In most brain regions branched et al., 2004), but EA amphetamine has either no effect spines are quite rare (Sorra et al., 1998), but there is or a weak effect on these dendrites [in one study the increasing evidence that they are associated with synap- effect of amphetamine was non-significant (Robinson tic plasticity. Trommaldet al. (1990, 1996) have reported andKolb, 1997 ) andin another only marginally signifi- that after long-term potentiation (LTP) evokedby cant andmuch less than for cocaine ( Robinson and stimulation of the perforant path there is an increase in Kolb, 1999a)]. In contrast, EA amphetamine has been the number of branched spines on dendrites of dentate reported to increase spine density on the basilar den- granule cells, and Geinisman andMorrell (1989) founda drites of mPFC pyramidal cells when given to juvenile similar effect after kindling induced by perforant path (P22–P34) rats (Heijtz et al., 2003). stimulation. Comery et al. (1996) reportedthat there is One of the most striking regional differences comes an increase in branchedspines on mediumspiny neurons from comparing the effects of drugs on pyramidal cells in the CPu of rats raisedin a complex environment. in the mPFC versus the orbital prefrontal cortex Although little is known about branchedspines (oPFC). The mPFC andoPFC are two closely related (Harris andKater, 1994 ), recent evidence suggests they frontal regions; e.g., they both receive projections from may represent a more radical alteration in synaptic the mediodorsal nucleus of the (defining them organization than might seem at first glance. A com- as prefrontal cortex), andthey are the primary cortical mon hypothesis is that branchedspines are formedby targets of mesocortical DA inputs in rats (Kolb, 1990). ‘‘splitting’’ an existing presynaptic bouton to form two Experience with SA amphetamine or EA nicotine new synapses (Sorra et al., 1998). But in an evaluation increases spine density in the mPFC, but decreases of serial EM sections in CA1 of the hippocampus spine density in the oPFC (Crombag et al., 2004; Sorra et al. (1998) foundthat of 91 branchedspines, Gonzalez et al., 2004). Experience with SA cocaine ‘‘Different branches of the same spine never synapsed increases spine density in the mPFC, but has no effect with the same presynaptic bouton’’, andthat the bou- on spine density in the oPFC (Ferrario et al., 2003). tons on each branch head, ‘‘were not even neighboring Both EA and SA morphine decrease spine density in boutons splitting along the same axon’’ (p. 236). the mPFC, but increase spine density in the oPFC Similarly, in their LTP material Trommaldet al. (1996) (Table 1; Robinson et al., 2002). It appears, therefore, T.E. Robinson, B. Kolb / Neuropharmacology 47 Supplement No. 1 (2004) 33–46 39 that different drugs reorganize these closely related pre- administration is also an important factor. Thus, to the frontal regions in very specific andvery differentways. extent that studies of drug experience-dependent plas- Another curious drug-dependent difference is seen in ticity in animals are intended to model some of the the motor cortex. A priori we wouldnot anticipate that changes that may occur in addiction it is important to psychomotor stimulant drugs would affect the motor determine whether EA (usually i.p.) and SA (usually cortex, especially if the mode of action is via effects on i.v.) drugs have similar or different effects. dopamine neurotransmission. As expected, ampheta- In some brain regions both EA andSA drugpro- mine does not affect dendritic branching in motor duce the same changes in spine density. Thus, both EA cortex (unpublishedobservations). But Brown and andSA amphetamine andcocaine increase spine den- Kolb (2001) foundthat nicotine doesincrease branch- sity in the Acb andmPFC ( Robinson andKolb, 1999a; ing in motor cortex, although more recent studies have Robinson et al., 2001; Ferrario et al., 2003; Crombag foundit has no effect on the adjacentsomatosensory et al., 2004), andboth EA andSA morphine decrease cortex (Gonzalez et al., 2004). Of course, nicotine has spine density in the Acb and mPFC (Robinson and actions on cholinergic afferents to the cortex so this Kolb, 1999b; Robinson et al., 2002). There are, how- may be the route of action but what is puzzling is why ever, brain regions in which EA andSA morphine have the effects are specific to motor cortex. different effects. For example, in the Par 1 of the rat Finally, a lesson learnedfrom examining other brain EA morphine decreases spine density on pyramidal regions is that drugs can have different effects on spines cells, but SA morphine has no effect. In the hippocam- versus branches. For example, EA amphetamine has no pus SA morphine decreases spine density but EA effect on dendritic branching on pyramidal cells in either morphine has no effect (Robinson et al., 2002). the parietal cortex (Par1) or occipital cortex (Oc1), but It is not clear what accounts for differences between decreases spine density on these cells (while increasing EA and SA drug administration on structural plas- branching and spine density on pyramidal cells in the ticity. Some of the difference may simply be pharmaco- mPFC) (Robinson andKolb, 1997; Kolb et al., 2003b ). kinetic, or relatedto the total amount of drug It is not clear what accounts for differences in the exposure. This usually varies considerably between stu- effects of even closely relateddrugslike amphetamine dies using EA versus SA drug, which also typically andcocaine. Some differences may be relatedto phar- involve different routes of administration. Further- macokinetics or even potency (e.g., the doses of more, there is evidence the neurobiological impact of amphetamine typically usedwouldincrease extra- drugs varies as a function of rate of drug delivery cellular DA to a much greater extent than the doses of (Porrino, 1993), even when the same route of adminis- cocaine used). A more interesting possibility is that the tration is used( Samaha et al., 2004). On the other differences in structural plasticity between ampheta- hand, some of the differences between EA and SA drug mine and cocaine (and other drugs) reflect differences administration could be related to other ‘‘psychologi- in their spectrum of action. For example, cocaine cal’’ differences between these two modes of drug (unlike amphetamine) has approximately equal affinity administration. for DA, 5-HT andNE transporters. Therefore, these two drugs produce quite different patterns of change in 5.1. Role of learning monoamine transmission, andthis may be reflectedby different patterns of structural plasticity. An interesting issue addressed by comparing EA and SA drug is the extent to which drug-associated structural plasticity is induced ‘‘unconditionally’’ as a 5. Whether drugs are EA or SA influences structural consequence of drug actions on the brain, or whether plasticity in some brain regions but not others they are relatedto learning about the relationship between an action and drug delivery (i.e., operant or In most studies on drug-induced structural plasticity instrumental learning). Many of the structures in which (andbehavioral plasticity for that matter) drugsare we have seen drug-induced changes in dendritic struc- administered by an experimenter (EA) rather than SA. ture are brain regions that have been implicated This is potentially an important issue because the neu- in various aspects of learning (Cardinal et al., 2002; robiological effects of drugs may vary depending on Killcross andCoutureau, 2003; CardinalandEveritt, whether it is SA or EA (Smith andDworkin, 1990; 2004; HollandandGallagher, 2004; Kelley, 2004 ). For Wilson et al., 1994; Dworkin et al., 1995; Mark et al., example, Kelley et al. (1997) have suggestedthat the 1999; Stefanski et al., 1999). Furthermore, the rate at Acb is critical in instrumental (response-) which intravenously administered cocaine reaches the learning. If EA andSA drug have the same effect on brain has a dramatic effect on its ability to induce dendrites, presumably learning about the contingency immediate early genes in mesocorticolimbic structures between an action (a lever press or nose-poke) and (Samaha et al., 2004), andtherefore, modeof drug delivery of a drug reward is not responsible for the 40 T.E. Robinson, B. Kolb / Neuropharmacology 47 Supplement No. 1 (2004) 33–46 structural plasticity. Thus, the fact that EA andSA completely convincing argument, because learning amphetamine andcocaine have similar effects on spines about different rewards could have different effects. in the Acb andmPFC suggests this structural plasticity Nevertheless, at least we know that the structural is not due to instrumental learning. (An important changes are relatively stimulus (drug) specific. caveat is requiredhere. Although EA andSA psychos- timulant administration has similar effects on spine 5.2. Stress density in the Acb and mPFC, we do not know whe- Of course, psychostimulant drugs are also stressors, ther the synaptic reorganization reflectedby these activating the HPA axis, andsome of their effects on changes is qualitatively the same; see Section 2 above). dendritic structure could be related to these actions. Another way to address the role of instrumental Repeated intermittent stress alters dendritic mor- learning in structural plasticity is to study whether phology, for example, decreasing dendritic branching on learning an instrumental action for a non-drug reward pyramidal cells in the CA3 region of the hippocampus (food) has a similar effect as learning an action for a (McEwen, 1999) and increasing dendritic branching in drug reward. We have found that learning to work for the amygdala (Vyas et al., 2002). We have not studied a foodrewardhas no effect on spine density in the Acb the CA3 region or amygdala, but amphetamine SA andmPFC ( Robinson et al., 2001; Ferrario et al., 2003; experience increases spine density on CA1 pyramidal Crombag et al., 2004). In contrast, we have foundthat cells andhas no effect on dentategranule cells ( Crombag in the hippocampus responding for either a food et al., 2004). Interestingly, like psychostimulant drugs, rewardor for amphetamine increases spine density sodium depletion (a stressful experience) increases den- (Crombag et al., 2004). Thus, the available evidence dritic branching on medium spiny neurons in the AcbS, suggests that drug-induced spine changes in the Acb andenhances (sensitizes) the later psychomotor response andthe mPFC are not a consequence of instrumental to an amphetamine challenge (Roitman et al., 2002). learning, whereas in the hippocampus it is possible the In summary, there are a number of different actions spine changes are relatedto instrumental learning. of drugs that may contribute to structural plasticity in It is much more difficult to determine the extent to different brain regions. It will be a major challenge to which the kinds of changes in dendritic structure we untangle the extent to which drug-induced structural have described here with EA and SA drug are related changes in any given brain region, or any given cell to Pavlovian learning about the relationship between population, are related to unconditional drug effects, stimuli in the environment and drug administration learning, the actions of drugs as stressors, etc., or to (Cardinal et al., 2002). This kindof Pavlovian learning interactions amongst these factors. occurs whether drugs are EA or SA. That is, in both of these situations stimuli in the environment (especially the context) are paired with drug administration (the 6. Structural changes in some brain regions are US) andthese stimuli can acquire conditionedstimulus associated with the development of behavioral (CS) properties, whereby subsequent exposure to the plasticity (sensitization) whereas changes in other CS can either elicit a conditioned response or modulate brain regions are not the ability (set the occasion) of other stimuli to elicit responses (Anagnostaras andRobinson, 1996; One form of behavioral plasticity produced by the Anagnostaras et al., 2002; Cardinal et al., 2002). It is repeatedintermittent administration of psychostimu- possible, therefore, that some of the changes in den- lant drugs, whether they are EA or SA, is behavioral drites described here are due to Pavlovian learning (or sensitization (Robinson andBerridge, 2000; Vezina, even higher level cognitive learning), rather than to an 2004). Sensitization refers to an increase in a drug unconditional drug effect. effect that occurs as a consequence of past drug Although learning may contribute to drug-induced administration, and a number of different drug effects structural plasticity, there are reasons to believe that at have been reportedto sensitize. Sensitization to the least some of the changes in dendritic structure are due psychomotor activating effects of drugs have been to ‘‘unconditional’’ drug effects, rather than a conse- best characterized, but sensitization of drug reward quence of associative learning. Probably the most com- andincentive-sensitization have also been reported pelling reason is that different rewards produce very (Robinson andBerridge,2003 , for review), andall different effects. The psychostimulants increase spine these forms of sensitization persist for long periods of density in the Acb and mPFC, whether EA or SA, time after the discontinuation of drug treatment. It has whereas morphine decreases spine density, and food been suggested that drug sensitization is due to non- has no effect. If the changes in spine density in these associative changes in the neural substrates that regions were relatedto Pavlovian learning one might mediate unconditional drug effects, including psycho- expect to see the same changes whether the rewardwas motor andincentive motivational effects, although the cocaine, morphine or food. Nevertheless, this is not a expression of behavioral sensitization can come under T.E. Robinson, B. Kolb / Neuropharmacology 47 Supplement No. 1 (2004) 33–46 41 strong CS control (i.e., its expression is modulated by 7. Drug experience-dependent structural plasticity learning; Stewart, 1992; Stewart andBadiani, 1993; may influence the structural plasticity associated Anagnostaras andRobinson, 1996; Anagnostaras et al., with other experiences 2002). It has also been suggestedthat the neuroadapta- tions responsible for sensitization are critical in the As discussed above, drugs of abuse usurp many of transition from recreational or circumstantial drug use the cellular andmolecular mechanisms responsible for to the compulsive patterns of drug-seeking and drug- experience-dependent plasticity, and both experience taking behavior that characterize addiction (Robinson and drugs alter dendritic structure. This raises the andBerridge,1993, 2000, 2003 ). We have been inter- possibility that changes in synaptic organization pro- ested, therefore, in the extent to which drug-induced duced by experience may interact with those produced structural plasticity in specific brain regions is related by exposure to drugs of abuse. We recently tested this to the development of behavioral sensitization. hypothesis by studying the influence of treatment with One approach to this question is to determine whe- amphetamine or cocaine on an especially well-studied form of experience-dependent structural plasticity; that ther manipulations that influence behavioral sensitiza- associatedwith variation in environmental complexity tion have a similar effect on structural plasticity (i.e., (Kolb et al., 2003b). There is a wealth of evidence that the extent to which the two phenomena are dissoci- housing adult rats in a relatively complex environment able). A simple example is that as dose increases the (compared to standard cages) increases dendritic degree of behavioral sensitization increases (Kalivas branching, spine density and the number of synapses in andStewart, 1991 ), and so does the magnitude of the a variety of brain regions (Greenough et al., 1990; van increase in spine density in the Acb (Kolb et al., Praag et al., 2000; Kolb et al., 2003a). Therefore, we 2003b). More interesting is a procedure that allows one used this procedure to study the effect of past exposure to holddrughistory constant while manipulating whe- to psychostimulant drugs on the ability of housing in a ther drug treatment produces behavioral sensitization complex environment to alter dendritic structure. or not. There are doses of cocaine that induce robust Rats were given repeatedinjections of amphetamine behavioral sensitization when cocaine is given in a dis- or cocaine, using an injection regimen that produced tinct test environment (‘‘novel’’ condition) but not behavioral sensitization, andafter the last injection when cocaine is given in the home cage (‘‘home’’ con- they were housedfor 3–3.5 months either in a complex dition; Badiani et al., 1995; Browman et al., 1998). If environment or socially in standard lab cages. After drug-induced changes in dendritic structure are related this, their brains were obtained and dendritic branch- to the development of behavioral sensitization they ing andspine densityon cells in the AcbS andsomato- shouldbe seen in the ‘‘novel’’ condition but not the sensory cortex (Par1) were quantified. We found that ‘‘home’’ condition, even though both groups receive past treatment with amphetamine or cocaine interfered the same drug treatments (just in different environ- with the ability of experience in a complex environment ments). Indeed, Li et al. (2004) foundthat in the AcbC to produce structural plasticity in these brain regions repeatedcocaine treatment increasedspine density only (Kolb et al., 2003b). Although the topic requires much in the ‘‘novel’’ group (the group that developed beha- further investigation, this study suggests that in some vioral sensitization), but not in the ‘‘home’’ group (that situations, andin some brain regions, exposure to failedto sensitize). In contrast, cocaine increasedspine drugs of abuse may limit or occlude the ability of sub- density in the AcbS in both groups, that is, inde- sequent experience to promote synapse formation and/ pendent of sensitization. Furthermore, if the dose of or synaptic reorganization. cocaine (andnumber of treatments) was increased, This notion has important implications for thinking such that cocaine induced behavioral sensitization even about the long-term consequences of drug use on when given at ‘‘home’’, an increase in spine density was behavior andpsychological function. For example, now seen in the AcbC (Li et al., 2004). Thus, the many other kinds of experiences besides housing in a induction of this particular form of behavioral plas- complex environment are associatedwith structural ticity (sensitization) is associatedwith structural plas- plasticity, including learning and recovery of function ticity in the AcbC, but not the AcbS, although mere following brain damage, and the synaptic reorganiza- exposure to cocaine is sufficient to produce structural tion associatedwith these experiences is thought to plasticity in the AcbS (but not the AcbC). It appears, have desirable functional consequences. Thus, if prior therefore, that the conditions necessary for cocaine to exposure to drugs of abuse interferes with the ability of produce structural changes in these two subregions of experience to reorganize neural circuits this couldinter- the Acb are quite different, and only changes in the fere with the behavioral/cognitive advantages that AcbC are associatedwith a form of drugexperience- accrue with experience. Consistent with this notion, dependent behavioral plasticity that has been linked to Gonzalez et al. (2004) recently foundthat treatment the compulsive pursuit of drugs. with nicotine can affect motor learning. Rats were 42 T.E. Robinson, B. Kolb / Neuropharmacology 47 Supplement No. 1 (2004) 33–46 given nicotine while they learnedone skilledmotor et al., 2003; Chao andNestler, 2004 for review) as well task andthen went drug free for 2 months. They then as some behavioral effects of cocaine, andis a down- were trainedon a new skilledmotor task, but didnot stream target of delta-FosB, which is involved in receive further drug treatment. Past treatment with nic- actions of cocaine andis alteredby repeatedcocaine otine interferedwith the learning of the new motor administration (Kelz et al., 1999; Chao andNestler, task andeven with extendedtraining nicotine treated 2004). The potential role of neurotrophic factors in rats failedto learn the task. Saline-treatedanimals mediating long-term behavioral and neurobiological learnedthe new task in only a few days.Thus, prior adaptations produced by drugs of abuse is also an exposure to nicotine may have impairedthe process of important area of investigation (Bolanos andNestler, synaptic reorganization requiredfor new learning. 2004; Flores andStewart, 2000 ). For example, Flores These studies also raise an interesting way to think andtheir colleagues have shown that astrocytic basic about some of the behavioral andpsychological deficits fibroblast growth factor (bFGF) expression is seen in addicts. There is accumulating evidence that enhancedby repeatedtreatment with amphetamine, addicts present with a variety of neuropsychological andthat this is necessary for the induction of psycho- deficits indicative of frontal cortical dysfunction (Bolla motor sensitization (Flores et al., 1998; Flores and et al., 1998; Rogers andRobbins, 2001 ), andthese cog- Stewart, 2000; Flores et al., 2000). nitive deficits are often attributed to either a kind of Another approach to exploring potential mechanisms ‘‘lesion’’ effect, that is, frank neurotoxicity, or at least a comes from recent studies from Margaret Gnegy’s lab kindof ‘‘functional’’ lesion. But the interaction at Michigan using culturedPC12 cells. These cells con- between drugs and other experiences in producing tain endogenous DA (and NE) that is released by structural plasticity suggests an alternative hypothesis. amphetamine, andrepeatedintermittent treatment with It is possible that some of the neuropsychological defi- amphetamine enhances later amphetamine-stimulated cits seen in addicts are due to limits on structural and DA release from PC12 cells (Kantor et al., 2002), a synaptic plasticity imposedby drug use, rather than to phenomenon that has been associatedwith behavioral a kindof ‘‘lesion effect’’. sensitization in vitro (Robinson andBecker, 1982 ) and in vivo (Robinson et al., 1988; Vezina, 2004). Recently Park et al. (2002) have shown that repeatedintermittent 8. Mechanisms treatment with amphetamine also causes neurite out- growth in PC12 cells, andthe conditionsthat leadto Of course, there is considerable interest in the molecu- neurite outgrowth (andenhancedDA release) are very lar mechanisms responsible for structural plasticity similar to the conditions required to see robust beha- associatedwith both experience andexposure to drugs vioral sensitization andenhancedDA release in the of abuse (Hyman andMalenka, 2001; Nestler, 2001; intact animal. For example: (1) Amphetamine produces Ujike et al., 2002; Bolanos andNestler, 2004; Chao and much greater neurite outgrowth when given intermit- Nestler, 2004; Lamprecht andLeDoux, 2004 ). For tently than when given continuously or a single time. (2) example, these mechanisms are thought to involve a The degree of neurite outgrowth is greater after a period host of changes initiatedin part by calcium entry via of withdrawal than immediately after the last treatment. glutamate NMDA receptors, activation of numerous (3) The effect of amphetamine is attenuatedby blocking intracellular signaling cascades that alter gene the DA transporter. Park et al. (2003) have also expression, andeventually to changes in growth factors, exploredthe intracellular signaling pathways respon- cytoskeletal andadhesionmolecules, andmany other sible for this amphetamine-evokedenhancement in proteins needed to form new synapses. A discussion of neurite outgrowth andamphetamine-stimulatedDA these putative mechanisms is beyondthe scope of this release. They foundthat inhibition of MAP kinase or article, but suffice it to say that the molecular mechan- PKC (but not PKA) preventedthe neurite outgrowth isms involvedin development andin mediatingother andenhancedDA release producedbyrepeatedamphe- forms of experience-dependent plasticity are likely to be tamine treatment, whereas inhibition of PKA prevented sharedwith those involvedin mediating the structural the enhancement in DA release but not neurite out- plasticity associatedwith drugs of abuse. growth. Importantly, these intracellular signaling path- We are aware of very few studies that directly exam- ways have been implicatedin mediating the actions of ine the molecular basis of drug-induced structural plas- psychostimulant drugs, behavioral and neurochemical ticity. Norrholm et al. (2003) recently reportedthat sensitization, various forms of experience-dependent repeatedexposure to cocaine increases spine densityin plasticity, as well as structural plasticity (Nestler, 2001; the Acb, andthat this is preventedby inhibition of Chao andNestler, 2004 ). cyclin-dependent kinase 5 (Cdk5). This kinase is known Although the studies on drug-induced changes in den- to be involvedin regulating cytoskeletal proteins and dritic structure reviewed here do not address potential neurite outgrowth (e.g., Sasaki et al., 2002; Hallows mechanisms they do have important implications in T.E. Robinson, B. Kolb / Neuropharmacology 47 Supplement No. 1 (2004) 33–46 43 thinking about mechanisms, andhow to go about effects of drugs. Some may be produced uncondition- studying them. For example, the regional specificity of ally as a consequence of exposure to drugs, and some the effects, andthe variation in the effects of different may be relatedto learning about the relationship drugs, suggests that repeated drug treatment does not between actions or stimuli associated with drug deliv- alter dendritic structure due actions on some ubiqui- ery. Some may be relatedto the induction of forms of tous, brain-wide growth factor. To the extent that drug- behavioral plasticity, andothers not. For example, induced structural plasticity is mediated by molecules cocaine-induced changes in spine density in the AcbS that regulate neuronal andsynaptic structure ( Hyman are seen whether cocaine induces behavioral sensitiza- andMalenka, 2001; Nestler, 2001; Chao andNestler, tion or not (i.e., they occur as a function of mere drug 2004; Lamprecht andLeDoux, 2004 ) the actions of such history), but changes are seen in the AcbC only if agents must also vary regionally, andas a function of cocaine induces behavioral sensitization. mode of drug administration. Indeed, the specificity of It is not known how these structural changes alter the the effects described here may provide an excellent operation of cells andcircuits, but presumably the reor- avenue to delineate causal mechanisms. If one hypothe- ganization of these brain regions contributes to some of sizes, for example, that a given molecule is critical for the persistent sequelae associatedwith repeateddrug use, morphine to alter synaptic organization (at least as indi- including the hypersensitivity to the incentive motiva- catedby changes in spine density), then in Par1 only EA tional effects of drugs and drug-related stimuli, and cog- morphine shouldhave an effect on the molecule under nitive alterations, that are the hallmarks of addiction study, in the hippocampal formation only SA morphine (Robinson andBerridge, 2003 ). Finally, it is important shouldhave an effect, andin mPFC andoPFC one that repeatedexposure to drugsof abuse influences the might predict opposite effects (Table 3). Of course, it is ability of other life experiences to produce structural also possible that parsimony does not rule, and that the plasticity. Our initial studies suggest that repeated effects of different drugs of abuse on synaptic organiza- exposure to drugs of abuse limits the ability of other tion in different types of cells in different brain regions experiences to reorganize synapses in some brain are mediated by entirely different mechanisms. regions. Thus, some of the long-term behavioral and Whatever the case, the available data strongly suggest cognitive deficits seen in addicts could be to drug- that to delineate the molecular mechanisms for structur- induced limits on plasticity (Kolb et al., 2003b). How- al plasticity, andassociatedchanges in patterns of ever, to endon a positive note, given that drugs and synaptic connectivity, will require studying very specific other experiences interact in remodeling synapses, it is brain regions andcell types, andperhaps even specific also possible that some life experiences couldmitigate portions of a dendritic tree. the ability of drugs to reorganize synapses, and thus miti- gate some of the negative neurobehavioral consequences of drug abuse. That will be a topic of future study. 9. Conclusions

The available literature establishes that repeated Acknowledgements exposure to a number of different drugs of abuse (amphetamine, cocaine, nicotine andmorphine) alters We especially thank G. Gorney, Y. Li andR. Gibb the morphology of dendrites and dendritic spines on for excellent technical assistance with the Golgi stain- cells in brain regions associatedwith incentive motiv- ing andquantification, andother students andpost- docs who have contributed to this work (including ation, rewardandlearning, such as the Acb, CPu and H. Crombag, C. Ferrario andA. Samaha). Supported prefrontal cortex. Drug-induced structural plasticity is by grants from the National Institute on Drug Abuse evident long (many months) after the discontinuation (USA) andthe National Science andEngineering of drug treatment, suggesting that drugs of abuse pro- Research Council of Canada. duce a persistent reorganization of patterns of synaptic connectivity in these brain regions. The form of drug- induced changes in dendritic structure (e.g., whether branching or spine density increases or decreases) References varies depending on the drug, the brain region and Ahmed, S.H., Koob, G.F., 1998. Transition from moderate to even the portion of the dendritic field on a cell. This excessive drug intake: change in hedonic set point. Science 282, suggests that drugs do not produce these effects by pro- 298–300. moting a ubiquitous ‘‘growth factor’’, but drug-induced Anagnostaras, S.G., Robinson, T.E., 1996. Sensitization to the structural plasticity represents very specific patterns of psychomotor stimulant effects of amphetamine: modulation by associative learning. Behav. Neurosci. 110, 1397–1414. synaptic reorganization in very specific circuits. It is Anagnostaras, S.G., Schallert, T., Robinson, T.E., 2002. Memory not known which drug-induced structural changes in processes governing amphetamine-induced psychomotor sensitiza- which brain regions are relatedto which actions and tion. Neuropsychopharmacology 26, 703–715. 44 T.E. Robinson, B. Kolb / Neuropharmacology 47 Supplement No. 1 (2004) 33–46

Andersen, P., Soleng, A.F., 1998. Long-term potentiation and spatial Geinisman, Y., Morrell, F., de Toledo-Morrell, L., 1989. Perforated training are both associatedwith the generation of new excitatory synapses on double-headed dendritic spines: a possible structural synapses. Brain Res. Rev. 26, 353–359. substrate of synaptic plasticity. Brain Res. 480, 326–329. Andersen, P., Trommald, M., 1995. Possible strategies for finding the Gonzalez, C.L.R., Gharbawie, O.A., Whishaw, I.Q., Kolb, B., 2004. substrate for learning-induced changes in the hippocampal cortex. Chronic low-dose administration of nicotine stimulates dendritic J. Neurobiol. 26, 396–402. arborization, enhances cortical plasticity, but impairs later Badiani, A., Browman, K.E., Robinson, T.E., 1995. Influence of novel acquisition of skilledmotor behavior in rats. (submittedfor versus home environments on sensitization to the psychomotor stimu- publication). lant effects of cocaine andamphetamine. Brain Res. 674, 291–298. Gray, E.G., 1959. Axo-somatic andaxo-dendriticsynapses of the Berke, J.D., Hyman, S.E., 2000. Addiction, dopamine, and the cerebral cortex: an electron microscope study. J. Anat. 93, molecular mechanisms of memory. Neuron 25, 515–532. 420–433. Biernaskie, J., Corbett, D., 2001. Enrichedrehabilitative training pro- Greenough, W.T., 1984. Structural correlates of information storage motes improved forelimb motor function and enhanced dendritic in the mammalian brain: a review andhypothesis. Trends growth after focal ischemic injury. J. Neurosci. 21, 5272–5280. Neurosci. 7, 229–233. Bolanos, C.A., Nestler, E.J., 2004. Neurotrophic mechanisms in drug Greenough, W.T., Bailey, C.H., 1988. The anatomy of a memory: addiction. Neuromol. Med. 5, 69–83. convergence of results across a diversity of tests. Trends Neurosci. Bolla, K.I., Cadet, J.L., London, E.D., 1998. The neuropsychiatry of 11, 142–147. chronic cocaine abuse. J. Neuropsychiatry Clin. Neurosci. 10, 280–289. Greenough, W.T., Chang, F.-L.F., 1985. Synaptic structural corre- Browman, K.E., Badiani, A., Robinson, T.E., 1998. The influence of lates of information storage in mammalian nervous systems. In: environment on the induction of sensitization to the psychomotor Cotman, C.W. (Ed.), Synaptic Plasticity. Guilford Press, New activating effects of intravenous cocaine in rats is dose-dependent. York, pp. 335–372. Psychopharmacology 137, 90–98. Greenough, W.T., Withers, G.S., Wallace, C.S., 1990. Morphological Brown, R.W., Kolb, B., 2001. Nicotine sensitization increases changes in the arising from behavioral experience: dendritic length and spine density in the nucleus accumbens and what is the evidence that they are involved in learning and mem- cingulate cortex. Brain Res. 899, 94–100. ory? In: Squire, L.R., Lindenlaub, E. (Eds.), The Biology of Cardinal, R.N., Everitt, B.J., 2004. Neural and psychological Memory, Symposia Medica Hoechst, vol. 23. F. K. Schattauder mechanisms underlying appetitive learning: links to drug addic- Verlag, New York, pp. 159–185. tion. Curr. Opin. Neurobiol 14, 156–162. Groenewegen, H.J., Berendse, H.W., Meredith, G.E., Haber, S.N., Cardinal, R.N., Parkinson, J.A., Hall, J., Everitt, B.J., 2002. Emotion Voorn, P., Wolters, J.G., Lohman, A.H.M., 1991. Functional andmotivation: the role of the amygdala,ventral , and anatomy of the ventral, limbic system-innervatedstriatum. In: prefrontal cortex. Neurosci. Biobehav. Rev. 26, 321–352. Willner, P., Scheel-Kru¨ger, J. (Eds.), The Mesolimbic Dopamine Chang, F.L., Greenough, W.T., 1982. Lateralizedeffects of monocu- System: From Motivation to Action. John Wiley & Sons, New lar training on dendritic branching in adult split-brain rats. Brain York, pp. 19–59. Res. 232, 283–292. Hallows, J.L., Chen, K., DePinho, R.A., Vincent, I., 2003. Decreased Chao, J., Nestler, E.J., 2004. Molecular neurobiology of drug addic- cyclin-dependent kinase 5 (cdk5) activity is accompanied by redis- tion. Annu. Rev. Med. 55, 113–132. tribution of cdk5andcytoskeletal proteins andincreasedcytoske- Coleman, P.D., Riesen, A.H., 1968. Environmental effects on cortical letal protein phosphorylation in p35 null mice. J. Neurosci. 23, dendritic fields. I. Rearing in the dark. J. Anat. 102, 363–374. 10633–10644. Comery, T.A., Stamoudis, C.X., Irwin, S.A., Greenough, W.T., 1996. Harris, K.M., Kater, S.B., 1994. Dendritic spines: cellular specializa- Increased density of multiple-head dendritic spines on medium- tions imparting both stability andflexibility to synaptic function. sizedspiny neurons of the striatum in rats rearedin a complex Annu. Rev. Neurosci. 17, 341–371. environment. Neurobiol. Learn. Mem. 66, 93–96. Hebb, D.O., 1949. The Organization of Behavior. John Wiley, New Crombag, H.S., Gorny, G., Li, Y., Kolb, B., Robinson, T.E., 2004. York. Opposite effects of amphetamine self-administration experience on Heijtz, R.D., Kolb, B., Forssberg, H., 2003. Can a therapeutic dose of dendritic spines in the medial and orbital prefrontal cortex. Cereb. amphetamine during pre-adolescence modify the pattern of synap- Cortex (in press). tic organization in the brain? Eur. J. Neurosci. 18, 3394–3399. Dworkin, S.I., Mirkis, S., Smith, J.E., 1995. Response-dependent versus Holland, P.C., Gallagher, M., 2004. Amygdala-frontal interactions response-independent presentation of cocaine: differences in the andrewardexpectancy. Curr. Opin. Neurobiol. 14, 148–155. lethal effects of the drug. Psychopharmacology (Berlin) 117, 262–266. Horner, C.H., 1993. Plasticity of the . Prog. Neuro- Ferrario, C., Crombag, H.S., Gorny, G., Li, Y., Kolb, B., Robinson, biol. 41, 281–321. T.E., 2003. Amphetamine or cocaine self-administration produces Hyman, S.E., Malenka, R.C., 2001. Addiction and the brain: the persistent regionally-specific changes in spine density in prefrontal neurobiology of compulsion andits persistence. Nat. Rev. Neu- cortex of rats. Soc. Neurosci. Absts. rosci. 2, 695–703. Fiala, J.C., Spacek, J., Harris, K.M., 2002. Dendritic spine pathol- Jones, T.A., Kleim, J.A., Greenough, W.T., 1996. Synaptogenesis ogy: cause or consequence of neurological disorders. Brain Res. and dendritic growth in the cortex opposite unilateral sensori- Rev. 39, 29–54. motor cortex damage in adult rats: a quantitative electron micro- Flores, C., Stewart, J., 2000. Basic fibroblast growth factor as a scopic examination. Brain Res. 733, 142–148. mediator of the effects of glutamate in the development of long- Kalivas, P.W., 1995. Interactions between dopamine and excitatory lasting sensitization to stimulant drugs: studies in the rat. Psycho- amino acids in behavioral sensitization to psychostimulants. Drug pharmacology (Berlin) 151, 152–165. Alcohol Depend. 37, 95–100. Flores, C., Rodaros, D., Stewart, J., 1998. Long-lasting induction of Kalivas, P.W., Stewart, J., 1991. Dopamine transmission in the astrocytic basic fibroblast growth factor by repeatedinjections of initiation and expression of drug- and stress-induced sensitization amphetamine: blockade by concurrent treatment with a glutamate of motor activity. Brain Res. Rev. 16, 223–244. antagonist. J. Neurosci. 18, 9547–9555. Kantor, L., Park, Y.H., Wang, K.K., Gnegy, M., 2002. Enhanced Flores, C., Samaha, A.N., Stewart, J., 2000. Requirement of amphetamine-mediated dopamine release develops in PC12 cells endogenous basic fibroblast growth factor for sensitization to after repeatedamphetamine treatment. Eur. J. Pharmacol. 451, amphetamine. J. Neurosci. 20, RC55. 27–35. T.E. Robinson, B. Kolb / Neuropharmacology 47 Supplement No. 1 (2004) 33–46 45

Kasai, H., Matsuzaki, M., Noguchi, J., Yasumatsu, N., Nakahara, ing spatial learning in adult rats suggests the formation of new H., 2003. Structure-stability-function relationships of dendritic synapses. Proc. Natl. Acad. Sci. USA 91, 12673–12675. spines. Trends Neurosci. 26, 360–368. Nestler, E.J., 2001. Molecular basis of long-term plasticity underlying Kelley, A.E., 2004. Ventral striatal control of appetitive motivation: addiction. Nat. Rev. Neurosci. 2, 119–128. role in ingestive behavior andreward-relatedlearning. Neurosci. Nimchinsky, E.A., Sabatini, B.L., Svoboda, K., 2002. Structure and Biobehav. Rev. 27, 765–776. function of dendritic spines. Ann. Rev. Physiol. 64, 313–353. Kelley, A.E., Smith-Roe, S.L., Holahan, M.R., 1997. Response- Norrholm, S.D., Bibb, J.A., Nestler, E.J., Ouimet, C.C., Taylor, J.R., reinforcement learning is dependent on N-methyl-d-aspartate Greengard, P., 2003. Cocaine-induced proliferation of dendritic receptor activation in the nucleus accumbens core. Proc. Natl. spines in nucleus accumbens is dependent on the activity of Acad. Sci. USA 94, 12174–12179. cyclin-dependent kinase-5. Neuroscience 116, 19–22. Kelz, M.B., Chen, J., Carlezon, Jr.., W.A., Whisler, K., Gilden, L., Park, Y.H., Kantor, L., Wang, K.K., Gnegy, M.E., 2002. Repeated, Beckmann, A.M., Steffen, C., Zhang, Y.J., Marotti, L., Self, D.W., intermittent treatment with amphetamine induces neurite outgrowth Tkatch, T., Baranauskas, G., Surmeier, D.J., Neve, R.L., in rat pheochromocytoma cells (PC12 cells). Brain Res. 951, 43–52. Duman, R.S., Picciotto, M.R., Nestler, E.J., 1999. Expression of the Park, Y.H., Kantor, L., Guptaroy, B., Zhang, M., Wang, K.K., transcription factor deltaFosB in the brain controls sensitivity to Gnegy, M.E., 2003. Repeatedamphetamine treatment induces cocaine. Nature 401, 272–276. neurite outgrowth andenhancedamphetamine-stimulateddopa- Killcross, S., Coutureau, E., 2003. Coordination of actions and mine release in rat pheochromocytoma cells (PC12 cells) via a habits in the medial prefrontal cortex of rats. Cereb. Cortex 13, protein kinase C- and mitogen activated protein kinase-dependent 400–408. mechanism. J. Neurochem. 87, 1546–1557. Kolb, B., 1990. Prefrontal cortex. In: Kolb, B., Tees, R.C. (Eds.), Paulson, P.E., Camp, D.M., Robinson, T.E., 1991. The time course The Cerebral Cortex of the Rat. MIT Press, Cambridge, MA, of transient behavioral depression and persistent behavioral sensi- pp. 437–458. tization in relation to regional brain monoamine concentrations Kolb, B., Gibb, R., 1991. Environmental enrichment andcortical during amphetamine withdrawal in rats. Psychopharmacology injury: behavioral andanatomical consequences of frontal cortex 103, 480–492. lesions. Cereb. Cortex 1, 189–198. Peters, A., Feldman, M.L., 1976. The projection of the lateral genicu- Kolb, B., Whishaw, I.Q., 1998. Brain plasticity andbehavior. Annu. late nucleus to area 17 of the rat cerebral cortex. I. General Rev. Psychol. 49, 43–64. description. J. Neurocytol. 5, 63–84. Kolb, B., Forgie, M., Gibb, R., Gorny, G., Rowntree, S., 1998. Age, Porrino, L.J., 1993. Functional consequences of acute cocaine treat- experience andthe changing brain. Neurosci. Biobehav. Rev. 22, ment depend on route of administration. Psychopharmacology 143–159. (Berlin) 112, 343–351. Kolb, B., Gibb, R., Gorny, G., 2003a. Experience-dependent changes Ramon y Cajal, S., 1928. Degeneration andRegeneration in the Ner- in dendritic arbor and spine density in neocortex vary qualitat- vous System. OxfordUniversity Press, London. ively with age andsex. Neurobiol. Learn. Mem. 79, 1–10. Rampon, C., Jiang, C.H., Dong, H., Tang, Y.P., Lockhart, D.J., Kolb, B., Gorny, G., Li, Y., Samaha, A.N., Robinson, T.E., 2003b. Schultz, P.G., Tsien, J.Z., Hu, Y., 2000. Effects of environmental Amphetamine or cocaine limits the ability of later experience to enrichment on gene expression in the brain. Proc. Natl. Acad. Sci. promote structural plasticity in the neocortex andnucleus accum- USA 97, 12880–12884. bens. Proc. Natl. Acad. Sci. USA 100, 10523–10528. Robinson, T.E., Becker, J.B., 1982. Behavioral sensitization is Lamprecht, R., LeDoux, J., 2004. Structural plasticity andmemory. accompaniedby an enhancement in amphetamine-stimulated Nat. Rev. Neurosci. 5, 45–54. dopamine release from striatal tissue in vitro. Eur. J. Pharmacol. Leranth, C., Petnehazy, O., MacLusky, N.J., 2003. Gonadal hor- 85, 253–254. mones affect spine synaptic density in the CA1 hippocampal sub- Robinson, T.E., Becker, J.B., 1986. Enduring changes in brain and fieldof male rats. J. Neurosci. 23, 1588–1592. behavior produced by chronic amphetamine administration: a Leuner, B., Falduto, J., Shors, T.J., 2003. Associative memory forma- review andevaluation of animal modelsof amphetamine psy- tion increases the observation of dendritic spines in the hippocam- chosis. Brain Res. Rev. 11, 157–198. pus. J. Neurosci. 23, 659–665. Robinson, T.E., Berridge, K.C., 1993. The neural basis of drug crav- Li, Y., Kauer, J.A., 2004. Repeatedexposure to amphetamine dis- ing: an incentive-sensitization theory of addiction. Brain Res. rupts dopaminergic modulation of excitatory synaptic plasticity Rev. 18, 247–291. andneurotransmission in nucleus accumbens. Synapse 51, 1–10. Robinson, T.E., Berridge, K.C., 2000. The psychology and neurobiol- Li, Y., Kolb, B., Robinson, T.E., 2003. The location of persistent ogy of addiction: an incentive-sensitization view. Addiction 95 amphetamine-induced changes in the density of dendritic spines (Suppl. 2), S91–S117. on medium spiny neurons in the nucleus accumbens and caudate- Robinson, T.E., Berridge, K.C., 2003. Addiction. Annu. Rev. Psy- putamen. Neuropsychopharmacology 238, 1082–1085. chol. 54, 25–53. Li, Y., Acerbo, M.J., Robinson, T.E., 2004. The induction of beha- Robinson, T.E., Kolb, B., 1997. Persistent structural modifications in vioral sensitization is associatedwith cocaine-inducedstructural nucleus accumbens andprefrontal cortex neurons producedby plasticity in the core (but not shell) of the nucleus accumbens. previous experience with amphetamine. J. Neurosci. 17, Eur. J. Neurosci. (in press). 8491–8497. Mark, G.P., Hajnal, A., Kinney, A.E., Keys, A.S., 1999. Self-admin- Robinson, T.E., Kolb, B., 1999a. Alterations in the morphology of istration of cocaine increases the release of acetylcholine to a dendrites and dendritic spines in the nucleus accumbens and pre- greater extent than response-independent cocaine in the nucleus frontal cortex following repeatedtreatment with amphetamine or accumbens of rats. Psychopharmacology (Berlin) 143, 47–53. cocaine. Eur. J. Neurosci. 11, 1598–1604. McEwen, B.S., 1999. Stress andhippocampal plasticity. Annu. Rev. Robinson, T.E., Kolb, B., 1999b. Morphine alters the structure of Neurosci. 22, 105–122. neurons in the nucleus accumbens andneocortex of rats. Synapse McEwen, B.S., 2000. Effects of adverse experiences for brain struc- 33, 160–162. ture andfunction. Biol. Psychiatry 48, 721–731. Robinson, T.E., Jurson, P.A., Bennett, J.A., Bentgen, K.M., 1988. Moser, M.B., Trommald, M., Andersen, P., 1994. An increase in den- Persistent sensitization of dopamine neurotransmission in ventral dritic spine density on hippocampal CA1 pyramidal cells follow- striatum (nucleus accumbens) produced by past experience with 46 T.E. Robinson, B. Kolb / Neuropharmacology 47 Supplement No. 1 (2004) 33–46

(+)-amphetamine: a microdialysis study in freely moving rats. tive periods of brain development. Curr. Opin. Pharmacol. 4, Brain Res. 462, 211–222. 65–71. Robinson, T.E., Gorny, G., Mitton, E., Kolb, B., 2001. Cocaine self- Stefanski, R., Ladenheim, B., Lee, S.H., Cadet, J.L., Goldberg, S.R., administration alters the morphology of dendrites and dendritic 1999. Neuroadaptations in the dopaminergic system after active spines in the nucleus accumbens andneocortex. Synapse 39, 257–266. self-administration but not after passive administration of Robinson, T.E., Gorny, G., Savage, V.R., Kolb, B., 2002. Widespread . Eur. J. Pharmacol. 371, 123–135. but regionally specific effects of experimenter- versus self-adminis- Stewart, J., 1992. Conditioned stimulus control of the expression of teredmorphine on dendritic spines in the nucleus accumbens, hippo- sensitization of the behavioral activating effects of opiate and campus, andneocortex of adult rats. Synapse 46, 271–279. stimulant drugs. In: Gormezano, I., Wasserman, E.A. (Eds.), Rogers, R.D., Robbins, T.W., 2001. Investigating the neurocognitive Learning andMemory: The Behavioral andBiological Substrates. deficits associated with chronic drug misuse. Curr. Opin. Neuro- Erlbaum, Hillsdale, NJ, pp. 129–151. biol. 11, 250–257. Stewart, J., Badiani, A., 1993. Tolerance and sensitization to the Roitman, M.F., Na, E., Anderson, G., Jones, T.A., Bernstein, I.L., behavioral effects of drugs. Behav. Pharmacol. 4, 289–312. 2002. Induction of a salt appetite alters dendritic morphology in Stewart, M.G., Rusakov, D.A., 1995. Morphological changes asso- nucleus accumbens andsensitizes rats to amphetamine. J. Neu- ciatedwith stages of memory formation in the chick following rosci. 22, RC225. passive avoidance training. Behav. Brain Res. 66, 21–28. Rusakov, D.A., Stewart, M.G., Korogod, S.M., 1996. Branching of Thomas, M.J., Beurrier, C., Bonci, A., Malenka, R.C., 2001. Long- active dendritic spines as a mechanism for controlling synaptic term depression in the nucleus accumbens: a neural correlate of efficacy. Neuroscience 75, 315–323. behavioral sensitization to cocaine. Nat. Neurosci. 4, 1217–1223. Samaha, A.-N., Mallet, N., Ferguson, S.M., Gonon, F., Robinson, Trommald, M., Vaaland, J.L., Blackstad, T.W., Andersen, P., T.E., 2004. The rate of cocaine administration alters gene regu- 1990. Dendritic spine changes in rat dentate granule cells asso- lation and behavioural plasticity: implications for addiction. J. ciated with long-term potentiation. In: Guidotti, A. (Ed.), Neu- Neurosci. 24, 6362–6370. rotoxicity of Excitatory Amino Acids. Raven Press, New York, Sasaki, Y., Cheng, C., Uchida, Y., Nakajima, O., Ohshima, T., pp. 163–174. Yagi, T., Taniguchi, M., Nakayama, T., Kishida, R., Kudo, Y., Trommald, M., Hulleberg, G., Andersen, P., 1996. Long-term poten- Ohno, S., Nakamura, F., Goshima, Y., 2002. Fyn andCdk5 tiation is associatedwith new excitatory spine synapses on rat mediate semaphorin-3A signaling, which is involved in regulation dentate granule cells. Learn. Mem. 3, 218–228. of dendrite orientation in cerebral cortex. Neuron 35, 907–920. Tsay, D., Yuste, R., 2004. On the electrical function of dendritic Sesack, S.R., Pickel, V.M., 1990. In the rat medial nucleus accum- spines. TINS 27, 77–83. bens, hippocampal andcatecholaminergic terminals converge on Ujike, H., Takaki, M., Kodama, M., Kuroda, S., 2002. Gene spiny neurons andare in apposition to each other. Brain Res. expression relatedto synaptogenesis, neuritogenesis andMAP 527, 266–279. kinase in behavioral sensitization to psychostimulants. Ann. NY Shankaranarayana Rao, B.S., Lakshmana, M.K., Meti, B.L., Acad. Sci. 965, 55–67. Raju, T.R., 1999. Chronic () deprenyl administration alters den- van Praag, H., Kempermann, G., Gage, F.H., 2000. Neural con- dritic morphology of layer III pyramidal neurons in the prefrontal sequences of environmental enrichment. Nat. Rev. Neurosci. 1, cortex of adult Bonnett monkeys. Brain Res. 821, 218–223. 191–198. Shepherd, G.M., 1996. The dendritic spine: a multifunctional inte- Vezina, P., 2004. Sensitization of midbrain dopamine neuron reac- grative unit. J. Neurophysiol. 75, 2197–2210. tivity and the self-administration of psychomotor stimulant drugs. Sholl, D.A., 1981. The Organization of the Cerebral Cortex. Neurosci. Biobehav. Rev. 27, 827–839. Methuen, London. Vyas, A., Mitra, R., Shankaranarayana Rao, B.S., Chattarji, S., Sklair-Tavron, L., Shi, W.-X., Lane, S.B., Harris, H.W., Bunney, 2002. Chronic stress induces contrasting patterns of dendritic B.S., Nestler, E.J., 1996. Chronic morphine induces visible chan- remodeling in hippocampal and amygdaloid neurons. J. Neurosci. ges in the morphology of mesolimbic dopamine neurons. Proc. 22, 6810–6818. Natl. Acad. Sci. USA 93, 11202–11207. Wilson, C.J., Groves, P.M., Kitai, S.T., Linder, J.C., 1983. Three- Smith, A.D., Bolam, J.P., 1990. The neural network of the basal dimensional structure of dendritic spines in the rat neostriatum. ganglia as revealed by the study of synaptic connections of ident- J. Neurosci. 3, 383–398. ifiedneurons. TrendsNeurosci. 13, 259–265. Wilson, J.M., Nobrega, J.N., Corrigall, W.A., Coen, K.M., Shannak, Smith, J.E., Dworkin, S.I., 1990. Behavioral contingencies determine K., Kish, S.J., 1994. Amygdala dopamine levels are markedly changes in drug-induced transmitter turnover. Drug Develop. elevatedafter self- but not passive-administrationof cocaine. Res. 20, 337–348. Brain Res. 668, 39–45. Sorra, K.E., Fiala, J.C., Harris, K.M., 1998. Critical assessment Wolf, M.E., 1998. The role of excitatory amino acids in behavioral sen- of the involvement of perforations, spinules, andspine branching in sitization to psychomotor stimulants. Prog. Neurobiol. 54, 679–720. hippocampal synapse formation. J. Comp. Neurol. 398, 225–240. Woolley, C.S., 1999. Electrophysiological andcellular effects of Spiga, S., Serra, G.P., Puddu, M.C., Foddai, M., Diana, M., 2003. estrogen on neuronal function. Crit. Rev. Neurobiol. 13, 1–20. Morphine withdrawal-induced abnormalities in the VTA: con- Yuste, R., Bonhoeffer, T., 2001. Morphological changes in dendritic focal laser scanning microscopy. Eur. J. Neurosci. 17, 605–612. spines associatedwith long-term synaptic plasticity. Annu. Rev. Stanwood, G.D., Levitt, P., 2004. Drug exposure early in life: func- Neurosci. 24, 1071–1089. tional repercussions of changing neuropharmacology during sensi- Zilles, K., 1985. The Cortex of the Rat. Springer-Verlag, Berlin.