Proc. Natl. Acad. Sci. USAVol. 93, pp. 11202-11207, October 1996Neurobiology

Chronic morphine induces visible changes in the morphology ofmesolimbic dopamine neurons

(opiate addiction/brain-derived neurotrophic factor/tyrosine hydroxylase/Lucifer yellow/mesolimbic dopamine system)

LioRA SKLAIR TAVRON*, WEI-XING SHI, SARAH B. LANE, HERBERT W. HARRISt, BENJAMIN S. BUNNEY,AND ERIC J. NESTLERtLaboratory of Molecular Psychiatry, Departments of Psychiatry and Pharmacology, Yale University School of Medicine and Connecticut Mental Health Center,34 Park Street, New Haven, CT 06508

Communicated by Paul Greengard, Rockefeller University, New York NY July 5, 1996 (received for review May 21, 1996)

ABSTRACT The mesolimbic dopamine system, whicharises in the ventral tegmental area (VTA), is an importantneural substrate for opiate reinforcement and addiction.Chronic exposure to opiates is known to produce biochemicaladaptations in this brain region. We now show that theseadaptations are associated with structural changes in VTAdopamine neurons. Individual VTA neurons in paraformal-dehyde-fixed brain sections from control or morphine-treatedrats were injected with the fluorescent dye Lucifer yellow. Theidentity of the injected cells as dopaminergic or nondopam-inergic was determined by immunohistochemical labeling ofthe sections for tyrosine hydroxylase. Chronic morphinetreatment resulted in a mean -25% reduction in the area andperimeter of VTA dopamine neurons. This reduction in cellsize was prevented by concomitant treatment of rats withnaltrexone, an opioid receptor antagonist, as well as byintra-VTA infusion of brain-derived neurotrophic factor. Incontrast, chronic morphine treatment did not alter the size ofnondopaminergic neurons in the VTA, nor did it affect thetotal number of dopaminergic neurons in this brain region.The results of these studies provide direct evidence for struc-tural alterations in VTA dopamine neurons as a consequenceof chronic opiate exposure, which could contribute to changesin mesolimbic dopamine function associated with addiction.

Dopaminergic neurons in the ventral tegmental area (VTA)and their projections to the limbic forebrain are importantneural substrates for the acute reinforcing properties of opiatesand other drugs of abuse (1-4). It has been proposed thatchronic exposure to a drug of abuse elicits long-lasting adap-tations within these same brain regions that underlie changesin drug reinforcement mechanisms, as well as drug craving,associated with drug addiction.Among the most consistent drug-induced adaptations in the

VTA is induction of tyrosine hydroxylase (TH; refs. 5-8), therate-limiting enzyme in the biosynthesis of dopamine. Inaddition, chronic morphine administration has been shown todecrease levels of the three major neurofilament proteins,NF200, NF160, and NF68 (9), and to increase levels of glialfibrillary acidic protein (10) in this brain region. Each of thesebiochemical adaptations occurs specifically in the VTA and isnot elicited by chronic exposure to other psychotropic drugsthat lack reinforcing properties. In addition, similar adapta-tions in the VTA are observed in rats after chronic heroinself-administration, indicating that they are induced in re-sponse to self-regulated drug intake (11).The lower levels of neurofilament proteins and the recip-

rocal increase in glial fibrillary acidic protein present in theVTA after chronic opiate administration raises the possibilitythat chronic exposure to opiates may result in prominent

structural changes within this brain region (12, 13). Indirectsupport for this possibility is provided by the observation thatchronic morphine treatment results in a 50% impairment inaxoplasmic transport from the VTA to a major forebraintarget, with no change in transport observed in other neuralpathways studied (14). Moreover, intra-VTA infusion of brain-derived neurotrophic factor (BDNF) or related neurotrophins,known to support the survival of dopaminergic neurons in cellculture (see Discussion), has been shown recently to preventthe morphine-induced biochemical adaptations in this brainregion (15).The goal of the present study was to assess directly whether

the various biochemical adaptations elicited in the VTA bychronic morphine administration are associated with morpho-logical changes in this brain region. We show here that chronicexposure to morphine elicits a selective reduction in the size ofVTA dopamine neurons, with no change observed in nondo-paminergic neurons in this brain region. Moreover, the reduc-tion in size of VTA dopamine neurons is prevented by intra-VTA infusion of BDNF.

MATERIALS AND METHODSIn Vivo Drug Treatments. Sprague-Dawley rats (initial

weight 100-120 g; Camm Research Institute, Wayne, NJ) wereused in these studies. Morphine was administered chronically,as described (5), by daily subcutaneous implantation of mor-phine pellets (containing 75 mg of morphine base; NationalInstitute on Drug Abuse) for 5 days under light halothaneanesthesia. Rats were used 24 hr after the last pellet implan-tation. This treatment paradigm produces morphine toleranceand dependence based on behavioral, electrophysiological,and biochemical criteria (see ref. 5). Control rats receivedequivalent implantation of placebo pellets.

Naltrexone (Sigma) was administered daily [50 mg/ml i.p. insaline and 50 mg/ml in an emulsion of saline and lipid]immediately before morphine or placebo pellet implantationas described (10). This treatment paradigm has been shown tocompletely block the development of morphine tolerance anddependence based on behavioral, electrophysiological, andbiochemical criteria (see ref. 10).BDNF was provided by Ronald Lindsay (Regeneron Phar-

maceuticals, Tarrytown, NY). It was infused at a dose of 2,ug/day (in 1QmM sodium phosphate, pH 7.4/0.9% NaCl/1%bovine serum albumin) directly into the VTA by osmoticminipumps via midline cannulae exactly as described (15). This

Abbreviations: VTA, ventral tegmental area; TH, tyrosine hydroxy-lase; BDNF, brain-derived neurotrophic factor.*Present address: Technion Institute of Technology, Bat-Galim, Haifa31096, Israel.

tPresent address: National Institute on Aging, National Institutes ofHealth, Bethesda, MD 20892.ITo whom reprint requests should be addressed. e-mail: eric_nestler@qm.yale.edu.

11202

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement" inaccordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

10, 2

021

Proc. Natl. Acad. Sci. USA 93 (1996) 11203

BDNF dose has been shown to prevent many of the biochem-ical effects elicited in the VTA by chronic morphine administra-tion (15). Control rats received intra-VTA infusions of vehicle.Measurements of VTA Neuronal Morphology. The mor-

phology ofVTA neurons was assessed by filling individual cellswith Lucifer yellow by use of a procedure based on previousstudies (16-19). Rats were deeply anesthetized with chloralhydrate (400 mg/kg i.p.) and then perfused transcardially with50 ml of ice-cold phosphate-buffered saline (PBS), followed by200 ml of4% paraformaldehyde in PBS at a rate of 15 ml/min.Brains were removed and 100- to 150-,um thick oblique slices(described below) were obtained through the VTA by use ofa vibratome. Neurons in the fixed slices were injected withLucifer yellow under an upright fluorescence microscope.Individual neurons were visualized with Nomarski optics,impaled with a glass microelectrode filled with 0.2% Luciferyellow in PBS, and injected by applying negative current pulses(2 Hz, 400 msec, 1 nA) for 5-10 min. After filling 4-8 cells perslice, the slices were fixed in 4% paraformaldehyde in PBSovernight at 4°C.

Oblique brain slices, illustrated in Fig. 1, were used in thesestudies according to published procedures (21). In theseoblique slices, VTA dopamine neurons can be readily distin-guished from substantia nigra dopamine neurons as beinglocated medial to the medial terminal nucleus of the accessoryoptic tract (Fig. 1). The VTA cells injected with Lucifer yellowwere located at -5.3 + 0.2 mm Bregma.The dopaminergic phenotype of the Lucifer yellow-injected

neurons was determined by TH immunohistochemistry. Thefixed slices were washed with PBS and permeabilized by 2-3cycles of freezing (at -70°C) and thawing. The slices wereincubated for 18 hr with a mouse monoclonal anti-TH antibody(1:1500; Incstar, Stillwater, MN) in PBS plus 0.3% TritonX-100 at 4°C. The slices were washed three times in PBS atroom temperature, after which time they were incubated for2.5 hr with Texas Red-conjugated anti-mouse IgG made inhorse (1:200; Vector Laboratories) in PBS plus 0.3% TritonX-100 and 10% normal horse serum at room temperature.Slices were washed three times in PBS, mounted on gelatin-coated slides, dried, dehydrated, cleared in methyl salicilate,and coverslipped with DPX (Merck).

Slices were examined under fluorescence microscopy (Ni-kon) to visualize Lucifer yellow-filled neurons and to deter-mine whether the filled neurons were TH+ or TH-. The filledneurons were recorded with a Xabion charged coupled device-

FIG. 1. Schematic illustration of the location of VTA neurons

sampled in the present study (box). The dotted lines indicate thepositions of oblique brain slices, 100- to 150-,um thick, obtained in a

plane perpendicular to the printed page. The medial terminal nucleusof the accessory optic tract (mt) was used as the lateral boundary ofthe VTA. VTA neurons were sampled at -5.3 ± 0.2 mm Bregma, thelocation of the coronal section shown in the figure. SNR, substantianigra reticulata; and SNC, substantia nigra compacta. This figure ismodified from Paxinos and Watson (20).

gated video camera. The morphological parameters of theneurons were measured by National Institutes of Health IMAGEversion 1.57 software; this was done with the experimenterblinded to the identity of the neurons as TH+ or TH- and ofthe slices as control or morphine-treated. Cell perimeter andarea were then calculated.

In pilot experiments, 400-gm thick oblique midbrain slicescontaining the VTA from control and morphine-treated ratswere prepared as described (21). Dopamine neurons, identi-fied by well-established electrophysiological criteria (for ex-ample, see refs. 16 and 21), were recorded in whole cell modewith electrodes filled with an intracellular solution containing1% Lucifer yellow. The electrodes were withdrawn after 15min in cell-attached mode. One to two neurons were injectedper slice. The slices were then fixed in 4% paraformaldehyde,cut into 50- to 100-gm thick sections, and mounted for analysisof cell morphology as described above.

Counting VTA Dopamine Neurons. Control and morphine-treated rats were perfused with paraformaldehyde, as de-scribed above. TH immunohistochemistry was then conductedon fixed 40-gm thick coronal brain sections, and the numberofTH+ neurons was counted at x400 magnification, exactly asdescribed (22). Briefly, TH+ neurons were counted in every40-,um section throughout the rostral-caudal extent of theventral midbrain. An immunoreactive area was counted as acell if a region of central pallor, consistent with the absence ofstaining of a nucleus, could be visualized. The slides werecounted in a blind fashion. Regional boundaries of the VTAwere determined by use of neuroanatomical landmarks ac-cording to published criteria (20, 23). The major landmarkdefining the lateral boundary of the VTA was the mediallemniscus. Other major landmarks used to establish level andorientation included the interpeduncular nucleus, the rednucleus, and the fasciculus retroflexus. Gross regions of THimmunoreactivity were compared and aligned with previouslypublished studies of the protein in the rat (23, 24).

RESULTSRegulation of VTA Neuronal Morphology by Chronic Mor-

phine Administration. In the present study, we assessed di-rectly the changes in cellular morphology in the VTA producedby chronic morphine administration. We used an approachbased on previous work (see refs. 16-19), which involved fillingindividual neurons with the fluorescent dye Lucifer yellow inparaformaldehyde-fixed brain slices. The injected cells werethen identified as dopaminergic or nondopaminergic by im-munohistochemically labeling the slices for TH.

This technique resulted in the reliable labeling of individualneurons and the proximal extents of their processes as visu-alized under fluorescence microscopy. In control rats (Fig.2A), Lucifer yellow-filled TH' neurons exhibited considerablevariability in their shape, although most cells were ellipticalwith several large caliber proximal processes, consistent withprevious descriptions of the morphology ofVTA dopaminergicneurons (see refs. 16, 17, and 25). In contrast, in morphine-treated rats (Fig. 2B), Lucifer yellow-filled TH' neuronsexhibited smaller cell bodies and tended to show a morespherical shape with narrowed proximal processes. Analysis of41 neurons from seven control and from seven morphine-treated rats revealed a 20-25% reduction in mean cell bodyarea and perimeter elicited by chronic morphine exposure(Table 1). An equivalent (-25%) reduction in the area andperimeter of VTA dopamine neurons was observed afterchronic morphine administration when the neurons wereinjected with Lucifer yellow via patch-recording electrodes inliving brain slices (data not shown).

In addition to a reduction in mean cell body size, there wasa dramatic shift in the distribution of cell body sizes betweenthe control and chronic morphine-treated conditions. This is

Neurobiology: Sklair-Tavron et al.

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

10, 2

021

11204 Neurobiology: Sklair-Tavron et al.

B40A 32-

_ _ median 22,<r- control E

Ib.

24-

§ 2~~~~L2&

a ~

* 12

4-

control morphine

I0 v

- - +- + mcdin; control

:I:so

control morphine

FIG. 2. Fluorescence photomicrographs of Lucifer yellow-filledneurons in the rat VTA. All neurons shown in the figure are dopa-minergic (i.e., TH+). (A) Neurons from control rats; (B) neurons frommorphine-treated rats; (C) neuron from BDNF-treated rat; and (D)neuron from BDNF- and morphine-treated rat. (Bar = 8 ,um.)

illustrated in Fig. 3A, which shows that the perimeter of all butone of the VTA dopamine neurons of morphine-treated ratswas below the median value exhibited by the neurons fromcontrol rats. A similar shift in distribution was observed for cellbody area (data not shown).

Chronic morphine exposure also was found to result in asignificant (=30%) reduction in the mean length of dopami-nergic neuronal processes. The average length of Luciferyellow-filled processes in TH+ neurons from control rats was80 ± 10 ,um (n = 25) compared with 56 ± 6 ,um (n = 20) inTH+ neurons from morphine-treated rats (P < 0.03 by Stu-dent's t test).The reduction in cell body area and perimeter elicited by

chronic morphine administration was blocked by concomitanttreatment of rats with the opioid receptor antagonist naltrex-one, which by itself had no effect on the size of the neurons

Table 1. Effect of drug treatments on the size of VTAdopamine neurons

Treatment* Perimeter, ,um Area, j.m2Control 27.5 ± 0.6 (41) 43.5 ± 1.7 (41)Morphine 22.1 ± 0.5 (41)t 33.2 ± 1.4 (41)tNaltrexone 26.8 ± 2.8 (9) 40.2 ± 8.0 (9)Morphine plus naltrexone 26.4 ± 1.5 (15) 41.6 ± 4.6 (15)Vehicle: 26.2 ± 2.1 (9) 43.3 ± 2.9 (9)BDNF 27.9 ± 0.8 (28) 49.2 ± 2.4 (28)Morphine plus BDNF 27.4 ± 1.1 (19) 48.1 ± 3.5 (19)Data are expressed in means ± SEM (n).

*All animals that were not treated with morphine received subcuta-neous implantation of placebo pellets. Number of animals used wereas follows: control, seven; morphine, seven; naltrexone, three; mor-phine plus naltrexone, four; vehicle, three; BDNF, five; and morphineplus BDNF, four.tP < 0.01 by Student's t test.tAnimals received intra-VTA infusions of vehicle solution (see ref. 15)as well as subcutaneous implantation of placebo pellets.

FIG. 3. Distribution of the perimeter of VTA dopamine andnondopamine neurons under control and morphine-treated condi-tions. (A) Dopaminergic (i.e., TH+) neurons; and (B) nondopamin-ergic (TH-) neurons. Bar and cross-hatch show mean ± SEM. *, P <0.01 by Student's t test.

(Table 1). These findings indicate that the effects of morphineon neuronal morphology represent specific effects achievedvia the activation of opioid receptors.

Interestingly, the chronic morphine-induced decrease inneuronal size was selective for the dopaminergic neurons in theVTA. Analysis of Lucifer yellow-filled TH- neurons, in thesame brain slices that were analyzed for TH+ neurons, re-

vealed no difference in cell body area or perimeter betweencontrol and morphine-treated rats [area, 38.1 ± 1.5 jim2 (n =32) after morphine treatment versus 32.9 ± 1.8 tLm2 (n = 28)for control; perimeter, 24.2 + 0.7 ,Am (n = 32) after morphinetreatment versus 23.4 ± 0.8 ,um (n = 28) for control]. Therealso was no difference in the distribution of cell sizes underthese treatment conditions, as illustrated in Fig. 3B.

Influence of BDNF on Morphine-Induced MorphologicalChanges in the VTA. Based on the ability of BDNF infuseddirectly into the VTA to prevent many of the biochemicaladaptations elicited in the VTA by systemic morphine admin-istration (15), we tested whether such BDNF infusions wouldalso prevent the morphine-induced structural changes in VTAdopamine neurons. As shown in Fig. 2C, intra-VTA infusionsof BDNF for 10 days did not alter the size of VTA dopamineneurons, nor did they appear to influence the shape orarborizations of these cells (see also Table 1). However, 5 daysof pretreatment with BDNF, followed by concomitant admin-istration of BDNF and morphine for another 5 days, com-pletely prevented the morphine-induced reductions in the sizeof VTA dopamine neurons (Fig. 2D and Table 1). BDNF alsoappeared to prevent the changes in cell body shape seen withchronic morphine, in that most cells were elliptical.Lack of Effect of Chronic Morphine Administration on the

Number of VTA Dopamine Neurons. An immunohistochemi-cal analysis of TH was carried out to obtain a more completeassessment of dopaminergic neurons in the VTA. As shown inFig. 4, TH immunohistochemical staining under low magnifi-cation revealed no gross abnormalities within the ventralmidbrain after chronic morphine administration. There alsowere no gross differences evident between the dopaminergicneurons in the VTA under control and morphine-treatedconditions at this magnification.Under higher magnification (Fig. 4 B and E), the number of

dopaminergic neurons present throughout the rostral-caudalextent of the VTA was counted to determine whether chronicmorphine treatment was associated with neuronal loss. Nodifference was found between the number of VTA dopami-nergic neurons in control and morphine-treated rats, whichwere, respectively, 7128 ± 120 and 6983 ± 169 (mean ± SEM;n = 6). These numbers are consistent with previous reports of

40--A-

3-

2-

24-24-20-

a 1I

I2

6-

Proc. Natl. Acad. Sci. USA 93 (1996)

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

10, 2

021

Proc. Natl. Acad. Sci. USA 93 (1996) 11205

F:. _~ _.

/l, '..

s. 40

.:..

X

-J.

D:itt '"~~~~~~~~~~~~~~~~.0:.:30.,: i;; .............................. 00040j;0.0;;.00r........... t

~~~~~~~~... .................................. ..........................,Ui4,5, v )g+...............................................................................,._...

~~~~~~~~~~~L

FIG. 4. Immunohistochemical analysis of TH in the rat VTA. (A and D) Low power (X25) bright field photomicrographs of coronal sectionsof the VTA and portions of adjacent substantia nigra. Photomicrographs of sections through the VTA at (B and E) X400 and (C and F) high power(x 1000). A, B, and C were derived from control rats; and D, E, and F were derived from morphine-treated rats. The photomicrographs showncorrespond to stereotaxic coordinate -5.4 mm Bregma and are representative of the analysis of numerous sections from six control and sixmorphine-treated rats. [Bar = 94 ,tm (A and D), 13 ,tm (B and E), and 6 ,um (C and F).]

VTA dopamine neuronal counts in various rodent strains (seerefs. 22 and 24).At the higher magnification, there did appear to be a small

decrease in the size of dopamine neurons under morphine-treated conditions (Fig. 4 C and F). However, this was difficultto quantify due to the high density of TH+ immunoreactivityin cell bodies and processes in the tissue sections and to thedifficulty in identifying complete (i.e., unsectioned) neurons.

Indeed, this limitation in the quantitative capability of immu-nohistochemistry necessitated the use of Lucifer yellow injec-tions originally.

DISCUSSIONThe results of the present study provide the first directdemonstration that dopaminergic neurons in the VTA exhibitstructural changes after chronic morphine administration. Theobserved reduction in cell size is consistent with earlierspeculation (12, 13), which was based on adaptations incytoskeletal proteins (9, 10) and on the impairment in axo-

plasmic transport (14) in this brain region associated withchronic morphine exposure. In the present study, we show thatthis reduction in cell size is specific to dopaminergic neurons

within this brain region and requires the specific action ofmorphine at opioid receptors. At the same time, this reductionin cell size was not associated with gross abnormalities in theVTA, nor was it associated with death of VTA dopamineneurons. Since other drugs of abuse are known to producesome similar biochemical adaptations in the VTA comparedwith morphine (see Introduction), it would be interesting infuture studies to determine whether the drugs also producesimilar effects on the morphology of VTA dopamine neurons.

We also show in the present study that the morphine-induced changes in the size of VTA dopamine neurons can beprevented by intra-VTA infusion of BDNF. VTA dopamineneurons are known to express TrkB, the protein tyrosinekinase receptor that exhibits high affinity for BDNF (26).Moreover, BDNF has been shown to enhance the survival ofmidbrain dopamine and other neurons in cell culture in vitroand to protect the neurons from neurotoxins (e.g., MPTP or

6-hydroxydopamine) administered in vivo (e.g., see refs. 27-35). The present study shows that BDNF can in addition exerta more subtle effect on these neurons: it cannot only rescue thecells and prevent cell death, it can also prevent neuroplasticchanges that are not associated with cell death in response tochronic drug exposure.The results of the present study are also of interest in light

of recent work, which suggests that some of the morphine-induced biochemical adaptations in the VTA may be achievedvia morphine perturbation of BDNF signaling cascades. Forexample, chronic morphine administration has been shown toincrease the phosphorylation state and activity of extracellularsignal regulated kinase (ERK), a major effector of BDNF andrelated neurotrophins, selectively in this brain region (36).Direct evidence has been obtained to implicate such regulationof ERK in chronic morphine induction of TH in the VTA. Inthis sense, the ability of BDNF to prevent some of morphine'seffects may represent not only a pharmacological intervention,but also an intervention at or near the pathophysiologicalmechanisms that mediate some of the long-term actions ofmorphine in this brain region. In addition, these studies suggestpossible approaches to the development of novel pharmaco-logical treatments of opiate addiction; they raise the possibilitythat agents that activate the BDNF signaling cascade may beuseful in treating certain sequelae of long-term opiate exposure.The morphine-induced reduction in the size of VTA dopa-

mine neurons did not appear to be due to a different rate ofdiffusion of the Lucifer yellow in the neurons between controland morphine-treated conditions. This is because the dye wasinjected into individual neurons visualized under fluorescentmicroscopy, and was observed to completely fill all of theneurons sampled during the injection process. However, wecannot exclude the possibility that a different rate of diffusioninfluenced the morphine-induced reduction in the meanlength of VTA neuronal processes observed in these studies.A major question raised by the present findings is: What are

the functional consequences of the morphine-induced struc-tural changes observed in VTA dopamine neurons? We knowthat opiates acutely activate VTA dopamine neurons, an effectmediated, at least in brain slices, via inhibition of inhibitory

Neurobiology: Sklair-Tavron et al.

611F:i..f..

.4

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

10, 2

021

11206 Neurobiology: Sklair-Tavron et al.

GABAergic interneurons (37). However, the electrophysio-logical effects of chronic opiate exposure on these neuronshave received virtually no attention to date. In addition, we donot know whether the structural changes in VTA dopamineneurons seen after chronic morphine administration are me-diated via transsynaptic effects of morphine through theGABAergic neurons or via direct effects of morphine on thedopaminergic neurons. Evidence for such direct effects is theapparent expression of opioid receptors by VTA dopamineneurons (38) as well as the synaptic specializations betweenthese neurons and enkephalin-containing nerve terminalsobserved by electron microscopy (39).One possible model of opiate regulation of the VTA is that

opiates acutely activate VTA dopamine neurons with anintensity and persistence not seen under normal conditions.After chronic exposure, compensatory adaptations occur inthe VTA to oppose this activation. In the absence of drug,these changes would contribute to deficient VTA neuronalfunction and possibly to an aversive state postulated to occurduring drug withdrawal (3, 4, 40-42). According to this model,the structural alterations in VTA neurons could reflect animpairment in the functional capacity of these cells, whichwould thereby contribute to motivational dependence.

It is also possible that the changes observed in the morphol-ogy of VTA dopamine neurons after chronic morphine expo-sure represent a form of neural injury. Indirect support for thispossibility comes from prior observations that chronic mor-phine administration, as well as heroin self-administration,results in reduced levels of neurofilaments and elevated levelsof glial fibrillary acidic protein selectively within the VTA (seeIntroduction). Glial fibrillary acidic protein is an astrocyte-enriched protein often induced in concert with astrocyteactivation or proliferation. While this would represent the firstevidence for neural injury in response to chronic opiateadministration, damage to midbrain dopamine neurons andconcomitant induction of glial fibrillary acidic protein in selectbrain regions have been well-documented in response tochronic exposure to stimulants, particularly amphetamine andits derivatives (for example, see refs. 43 and 44). However, itis important to emphasize that the altered levels of neurofila-ments and glial fibrillary acidic protein and the changes inVTA neuronal morphology seen after morphine administra-tion, while perhaps suggestive of neural injury, could insteadreflect neuroplastic changes associated with increased ordecreased functioning of these cells. Indeed, one must be verycautious in interpreting the observed changes as reflectingneural injury, given the current and successful practice oftreating opiate addiction with opiate agonists, such as meth-adone (see ref. 45).While more work is clearly needed to study the validity of

these and alternative interpretations, this study highlights thecomplex types of adaptations that can be elicited in the brainfollowing a chronic drug treatment. It also adds to the growingevidence that many examples of plasticity in the functioning ofneurons in the fully differentiated adult brain are associatedwith plasticity that can be visualized at the anatomical level(for example, see refs. 46-51). As the array of molecular andcellular adaptations induced in the VTA and other brainregions by chronic opiate exposure is delineated, these changescan be related to specific behavioral features of addiction andmay direct the development of fundamentally new approachesto the treatment of addictive disorders.

This work was supported by U.S. Public Health Service GrantsDA08227, DA10160, and DA00203 and by the Abraham RibicoffResearch Facilities of the Connecticut Mental Health Center, State ofConnecticut Department of Mental Health and Addiction Services.

1. Wise, R. A. (1990) in Psychotropic Drugs ofAbuse, ed. Balfour,D. J. K. (Pergamon, Oxford), pp. 23-57.

2. Kuhar, M. J., Ritz, M. C. & Boja, J. W. (1991) Trends Neurosci.14, 299-302.

3. Koob, G. F. (1992) Trends Pharmacol. Sci. 13, 177-184.4. Self, D. W. & Nestler, E. J. (1995) Annu. Rev. Neurosci. 18,

463-495.5. Beitner-Johnson, D. & Nestler, E. J. (1991) J. Neurochem. 57,

344-347.6. Hurd, Y. L., Linefors, N., Brodin, E., Brene, S., Persson, H.,

Ungerstedt, U. & Hokfelt, T. (1992) Brain Res. 578, 317-326.7. Sorg, B. A., Chen, S.-Y. & Kalivas, P. W. (1993) J. Pharmacol.

Exp. Ther. 266, 424-430.8. Vrana, S. L., Vrana, K. E., Koves, T. R., Smith, J. E. & Dworkin,

S. I. (1993) J. Neurochem. 61, 2262-2268.9. Beitner-Johnson, D., Guitart, X. & Nestler, E. J. (1992) J. Neu-

rosci. 12, 2165-2176.10. Beitner-Johnson, D., Guitart, X. & Nestler, E. J. (1993) J. Neu-

rochem. 61, 1766-1773.11. Self, D. W., McClenahan, A. W., Beitner-Johnson, D., Terwill-

iger, R. Z. & Nestler, E. J. (1995) Synapse 21, 312-318.12. Nestler, E. J. (1992) J. Neurosci. 12, 2439-2450.13. Nestler, E. J., Hope, B. T. & Widnell, K. L. (1993) Neuron 11,

995-1006.14. Beitner-Johnson, D. & Nestler, E. J. (1993) NeuroReport 5,

57-60.15. Berhow, M. T., Russell, D. S., Terwilliger, R. Z., Beitner-John-

son, D., Self, D. W., Lindsay, R. M. & Nestler, E. J. (1995)Neuroscience 68, 969-979.

16. Grace, A. A. & Bunney, B. S. (1983) Neuroscience 10, 317-331.17. Grace, A. A. & Onn, S.-P. (1989) J. Neurosci. 9, 3463-3481.18. Onn, S.-P. & Grace, A. A. (1994) J. Neurophysiol. 71, 1917-1934.19. Shi, W. X. & Rayport, S. (1994) J. Neurosci. 14, 4548-4560.20. Paxinos, G. & Watson, C. (1986) The Rat Brain in Stereotaxic

Coordinates (Academic, Sydney), 2nd Ed..21. Shi, W. X. & Bunney, B. S. (1991) Synapse 9, 157-164.22. Harris, H. W. & Nestler, E. J. (1996) Brain Res. 706, 1-12.23. Hokfelt, T., Mortensson, R., Bjorkland, A., Kleinau, S. & Gold-

stein, M. (1984) in Handbook of Chemical Neuroanatomy, eds.Bjorkland, A. & Hokfelt, T. (Elsevier Science, Amsterdam), pp.277-379.

24. German, D. C. & Manaye, K. F. (1993) J. Comp. Neurol. 331,297-309.

25. Oades, R. D. & Halliday, G. M. (1987) Brain Res. Rev. 12,117-165.

26. Seroogy, K. B. & Gall, C. M. (1993) Exp. Neurol. 124, 119-128.27. Altar, C. A., Boylan, C. B., Fritsche, M., Jones, B. E., Jackson, C.,

Wiegand, S. J., Lindsay, R. M. & Hyman, C. (1994) J. Neuro-chem. 63, 1021-1032.

28. Beck, K. D., Knusel, B. & Hefti, F. (1993) Neuroscience 52,855-866.

29. Fadda, E., Negro, A., Facci, L. & Skaper, S. D. (1993) Neurosci.Lett. 159, 147-150.

30. Hyman, C., Juhasz, M., Jackson, C., Wright, P., Ip, N. Y. &Lindsay, R. M. (1994) J. Neurosci. 14, 335-347.

31. Skaper, S. D., Negro, A., Facci, L. & Dal Toso, R. (1993)J. Neurosci. Res. 34, 478-487.

32. Spenger, C., Hyman, C., Studer, L., Egli, M., Evtouchenko, L.,Jackson, C., Dahl-Jorgensen, A., Lindsay, R. M. & Seiler, R. W.(1995) Exp. Neurol. 133, 50-63.

33. Spina, M. B., Squinto, S. P., Miller, J., Lindsay, R. M. & Hyman,C. (1992) J. Neurochem. 59, 99-106.

34. Studer, L., Spenger, C., Seiler, R. W., Altar, C. A., Lindsay, R. M.& Hyman, C. (1995) Eur. J. Neurosci. 7, 223-233.

35. Zhou, J., Bradford, H. F. & Stern, G. M. (1994) Brain Res. 656,147-156.

36. Berhow, M. T., Hiroi, N. & Nestler, E. J. (1996) J. Neurosci. 16,4707-4715.

37. Johnson, S. W. & North, R. A. (1992) J. Neurosci. 12, 483-488.38. Mansour, A., Watson, S. J. & Akil, H. (1995) Trends Neurosci. 18,

69-70.39. Sesack, S. R. & Pickel, V. M. (1995) Brain Res. 672, 261-275.40. Harris, G. C. & Aston-Jones, G. (1994) Nature (London) 371,

155-157.41. Koob, G. F. (1996) Neuron 16, 893-896.42. Nestler, E. J. (1996) Neuron 16, 897-900.43. Switzer, R. C. (1993) Ann. N.Y Acad. Sci. 679, 341-348.44. O'Callaghan, J. P. & Miller, D. B. (1994)J. Pharmacol. Exp. Ther.

270, 741-751.

Proc. Natl. Acad. Sci. USA 93 (1996)

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

10, 2

021

Neurobiology: Sklair-Tavron et al.

45. Novick, D. M., Richman, B. L., Friedman, J. M., Friedman, J. E.,Fried, C., Wilson, J. P., Townley, A. & Kreek, M. J. (1993) DrugAlcohol Depend. 33, 235-245.

46. Hosokawa, T., Rusakov, D. A., Bliss, T. V. P. & Fine, A. (1995)J. Neurosci. 15, 5560-5573.

47. Marty, S. & Peschanski, M. (1995) J. Comp. Neurol. 356, 523-536.

Proc. Natl. Acad. Sci. USA 93 (1996) 11207

48. Rocha, M. & Sur, M. (1995) Proc. Natl. Acad. Sci. USA 92,8026-8030.

49. Segal, M. (1995) Neurosci. Lett. 193, 73-76.50. Stein-Behrens, B., Mattson, M. P., Chang, I., Yeh, M. & Sapol-

sky, R. (1994) J. Neurosci. 14, 5373-5380.51. Woolley, C. S. & McEwen, B. S. (1994)J. Neurosci. 14, 7680-7687.

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

10, 2

021