The Journal of Neuroscience, May 1995, 15(5): 3583-3593
Increased Expression of c-jun, junB, AP-1, and Preproenkephalin mRNA in Rat Striatum Following a Single Injection of Caffeine
Per Svenningsson,’ Anders Str6m,2 Bjijrn Johansson,’ and Bertil B. Fredholm’ l Department of Physiology and Pharmacology, Center for Pharmacology, Karolinska Institutet, S-l 71 77 Stockholm, Sweden and 2Department of Medical Nutrition, Huddinge Hospital, S-141 86 Huddinge, Sweden
The effect of a single injection of caffeine on the expres- lig et al., 1992; Daly, 1993). An important role is apparently sion of c-fos, c-jun, junB, and junD, on activator protein 1 played by adenosinereceptors which are inhibited by caffeine (AP-1) and on the levels of preproenkephalin mRNA in rat in concentrationssimilar to those encounteredin humansafter striatum was studied. Male rats were given caffeine (25 normal caffeine use (Fredholm, 1980). Only at higher concen- mglkg, 50 mg/kg, or 100 mglkg, i.p.) and sacrificed at dif- trations does caffeine inhibit CAMP phosphodiesterasesor re- ferent times (0.5, 1, 2, 4, or 8 hr) after administration. By leaseintracellular calcium. Of the four adenosinereceptors that using in situ hybridization of adjacent sections we found a have been cloned, the adenosineA, and AZa receptors are the rapid, transient, and dose-dependent increase of c-fos, c- most likely sitesof action of caffeine (seeFredholm et al., 1994). jun, and junB by caffeine in striatum, especially in the lat- AdenosineA,, receptorsare enriched in striatum (Parkinsonand eral part. The induction peaked after 1 hr, but persisted for Fredholm, 1990) whereas adenosineA, receptors have a wide 2 hr, and in the case of junB for 4 hr. No induction of junD distribution in the CNS and are present in striatum (Fastbomet was found. A strong induction of jun6, a weak induction of al., 1987). c-fos and c-jun, but not of junD, was seen in nucleus ac- In order to examine the neuronal basisof caffeine action we cumbens. Furthermore, by using gel shift assay we found have used the approach pioneered by Sagar and coworkers an induction of AP-1 by caffeine (100 mglkg) in striatum, (1988), namely to map the induction of immediateearly genes which peaked 2 hr after administration and was clearly in- following drug treatment. Caffeine induces C--X in the lateral creased after 4 hr. c-Fos, c-Jun, and JunB proteins were part of striatum, in both substanceP- and enkephalin-containing components of the AP-1. There was also a dose-dependent GABAergic medium-sized spiny neurons (Johanssonet al., induction of preproenkephalin mRNA, which was most pro- 1994).This is in contrast to c-fos induction by cocaineand am- nounced in the lateral and caudal part of striatum; the level phetaminewhich is restricted to substance P-containing neurons peaked 4 hr after injection and was still significantly in- (Young et al., 1991; Beretta et al., 1992). creased after 8 hr. In a complementary study we could not c-fos is a memberof a family of immediateearly genesthat find increased binding to the AP-l-like site in the 5’-flank- also includes, for example, c-&n andjunB. Recently it was re- ing sequence of proenkephalin following caffeine treat- ported that junB was induced parallel with c-fos following co- ment. caine administration, whereasc-&n was not altered (Moratalla The data show that a single dose of caffeine induces a et al., 1993). A different pattern of expressionwas found in the temporally and spatially characteristic pattern of c-fos, spinal cord after peripheral stimulation and in brain after sei- c-jun, and junB induction, followed by changes in AP-1 and zures (Dragunow and Robertson, 1987; Morgan et al., 1987; preproenkephalin mRNA. Thus, a single dose of caffeine Saffen et al., 1988; Sonnenberget al., 1989b; Wisden et al., causes changes in gene transcription in the brain that may 1990; Herdegenet al., 1991a,b;Uhl et al., 1991).These findings be related to the adaptive changes that occur after caffeine suggestthe possibility that dependingon the stimulus there is a administration. However, a direct causal link between the characteristicpattern of expressionof immediateearly genes. immediate early genes and enkephalin could not be prov- The different membersof the Fos/Jun family of immediate en. early genescombine to form homodimersand heterodimersthat [Key words: caffeine, immediate early genes, AP-1, en- kephalin, gel shift assay, in situ hybridization] bind to the activator protein 1 (AP-I) consensussequence (Nak- abeppuet al., 1988).The basalexpression of AP-1 in rat striatum Caffeine is the most widely consumedpsychomotor stimulant in viva is weak, but is inducible, for example, following am- (Nehlig et al., 1992). The mechanism(s)underlying the central phetamine administration (Nguyen et al., 1992; Wang et al., actions of caffeine are probably complex (for reviews, see Neh- 1992). AP-1 may be involved in the regulation of the neuropeptide Received May I I, 1994; revised Nov. 28, 1994; accepted Dec. I, 1994. enkephalin in vitro (Comb et al., 1988; Sonnenberg et al., These studies were supported by the Swedish Medical Research Council 1989a). However, it has recently beenreported that AP-1 is not (Project 2553), by the Swedish Cancer Fund, by the Institute for Scientific Information on Coffee, and by Karolinska Institutet. B.J. was a recipient of a important for this regulation in vivo (Konradi et al., 1993). En- Swedish Medical Research Council graduate fellowship. kephalin and preproenkephalinmRNA are found almost exclu- Correspondence should be addressed to Per Svenningsson, Department of sively in GABAergic medium-sizedspiny neuronsprojecting to Physiology and Pharmacology, Karolinska Institutet, S-171 77 Stockholm, Sweden. the globus pallidus(Gerfen and Young, 1988). The level of pre- Copyright 0 1995 Society for Neuroscience 0270.6474/95/153583-l l$OS.OO/O proenkephalinmRNA increases,in the lesionedside, in rats with 3584 Svenningsson et al. * Caffeine and Immediate Early Genes unilateral 6-hydroxydopamine lesions of the nigrostriatal dopa- dATP (Du Pont New England Nuclear, Stockholm, Sweden) to a spe- mine pathway (Young et al., 1986; Gerfen et al., 1990, 1991) cific activity of about lo” cpm/gm. Gel shift assay. Striatum was immediately dissected from the brain following acute or chronic treatment with dopamine DZ receptor and frozen at -sO’C. Each striatum was then manually ground in an antagonists (Tang et al., 1983; deceballos et al., 1986; Taylor et Eouendorf tube. Thereafter the cells were lvsed with 50 ml of buffer al., I99 I ; Angulo, 1992). Recently it has also been reported that [id mM HEPES-KOH (pH 7.9), 0.4 M NaC< 0.1 mM EDTA, 5% glyc- chronic treatment with caffeine causes increased levels of pre- erol, I mM dithiothreitol, 2 mM phenylmethane sulfonylfluoride]. The samples were centrifuged at 10,000X g for I5 min, the supernatants proenkephalin mRNA (Schiffmann and Vanderhaeghen, 1993a). were collected, aliquoted in small volumes, and frozen at - 135°C until Here we report that a single injection of caffeine causes not used. The protein content was measured as described by Bradford only the previously demonstrated induction of c+s, but also an (1976) and diluted to the same concentration. We used whole-cell ex- induction of c-jun and junB, and also that the levels of the tran- tracts in this experiment for two reasons. First, the association of Fos and Jun proteins occurs in the nucleus (Curran et al., 1990), so all of scription factor AP-I and preproenkephalin mRNA are in- the measured AP-I is probably located to nucleus and can act as tran- creased. We also present data indicating that the AP-l-like site scription factor. Second, the preparation of proteins from whole-cell in the proenkephalin enhancer does not bind AP-I. extracts is rapid and the activity of proteases and phosphatases is lim- ited. Materials and Methods The AP- I binding sites were double-stranded synthetic oligonucleo- tides, CCATGGTTGCTGACTAATTGAGATCT, deduced from the SV Anirml.s artrl drug trrcrrment. Male Sprague-Dawley rats (ALAB, Stock- 40 enhancer (Jones et al., 1988) and, AAGCATGAGTCAGACAC, holm, Sweden) weighing 200-220 gm were injected intraperitoneally from the collagenase enhancer (Lee et al., 1987). The ENKCRE-2 (i.p.) with saline or with caffeine (Sigma, Labkemi, Stockholm, Swe- binding site was a double-stranded synthetic oligonucleotide, den) 25 mglkg, 50 mglkg, or 100 mg/kg. Thirty minutes, I hr, 2 hr, 4 GATCGGCCTGCGTCAGCTG, deduced from the proenkephalin en- hr, or 8 hr after injection rats were briefly anesthetized with CO? and hancer (Comb et al., 1988; Konradi et al., 1993). The NF-1 binding site killed immediately by decapitation. There is no difference in caffeine- was a double-stranded synthetic oligonucleotide, ATTTTGGCTTGA- induced ~~f0.sin rats briefly anesthetized with CO? before sacrifice and AGCCAATATG, deduced from the adenovirus origin of replication rats sacrificed without being CO,-anesthetized (P Svenningsson, B. Jo- (Chodosh et al., 1988). The oligos were end-labeled by using polynu- hansson, and B. B. Fredholm, unpublished observations). The brain was cleotide kinase and -Q’P-ATP (Amersham, Solna, Sweden). rapidly dissected out, frozen on dry ice, and stored at -20°C for less Twenty micrograms of protein extract was incubated on ice water in than a week until sectioning (see below). 30 mM HEPES (oH 7.9). 6% Ficoll (400). 80 mM KCI. 0.015% NP 40. In situ hyhridimtion. Coronal sections (14 pm) were made through 20 mM NaCI, Oyl mg/ml poly-dIdC‘ for’ IO min. Thereafter 0. I ng of striatum, thalamus, hippocampus, and cerebellum and thaw-mounted on labeled probe was added to a final volume of 20 ml. For competition poly-t-lysine (50 mg/ml)-coated slides. For cellular colocalization stud- experiments a IOO-fold excess of unlabeled probe was added to the ies consecutive 5 pm sections were made. binding reaction. The samples were incubated for I hr on ice-water. The Sections were thereafter hybridized in a cocktail containing 50% resulting protein-DNA complexes were resolved on a preelectrop- deionized formamide (Fluka, Buchs, Switzerland), 4X SSC (3 M NaCI, horesed 5% nondenaturating polyacrylamide gel in a buffer containing 0.3 M trisodium citrate), IX Denhardt’s solution, 1% sarcosyl, 0.02 M 25 mM Tris-borate, 0.5 mM EDTA, 0.02% NP 40. The gel was dried NaPO, (pH 7.0) 10% dextran sulfate, 0.5 mg/ml yeast tRNA (Sigma, and autoradiographed on XAR film (Kodak, Jlrfalla, Sweden) for I d. Labkemi, Stockholm, Sweden), 0.06 M dithiothreitol, 0. I mglml sheared Antibodies used in gel shift analysis. Specific antibodies to c-Fos, salmon sperm DNA, and IO7 cpm/ml of probe. After hybridization for c-Jun, JunB, and JunD (Thierry et al., 1992, or those from Santa Cruz IS hr at 42°C the sections were washed four times, for 15 min each, Biotechnology Inc., Santa Cruz, CA) were used. The antibodies were in IX SSC at 55°C dipped briefly in water, 70%, 95%, and 99.5% added I5 min after addition of the probe to the binding reaction. ethanol and air dried; I4 pm sections were apposed to Hyperfilm-B,,,, Statistics. The differences between means + SEM were analyzed by (Amersham, Solna, Sweden) for 2-4 weeks. For the colocalization stud- single-factor analysis of variance (ANOVA) followed by pairwise group ies consecutive 5 pm sections were incubated with different probes and comparisons by the Tukey-Kramer method (SYSTAT). dipped in NTB-3 emulsion (Kodak, Jlrfalla, Sweden) and exposed for 4-7 weeks. After the emulsion had been developed, sections were light- Results ly counterstained in cresyl violet (0.5%). Films were analyzed with a Microcomputer Imaging Device (Imaging I!#ect of a single injection qf’ cufeine on the expression of Research). Results are presented as relative optical density, and film c-jun, junB, c-fos, and junD in striatum background was subtracted from all measured values. Emulsion auto- In situ hybridization against the four different oncogenes was radiograms were evaluated by inspection in a light microscope or by analyzing photographs. Cells that showed a grain density at least four done in adjacent sections from each animal. There was a clear- times the background were considered to be labeled, the remaining cells cut induction of c-jun following administration of caffeine (100 unlabeled. To identify neuron-like cells in two consecutive sections, mg/kg), which was most pronounced in the lateral part of cau- landmark structures (blood vessels, etc.) were identified in both sections date-putamen (Figs. I, 2, 3~). The expression of c-jun was de- and the two sections virtually superimposed using the image analysis system. tected by two different probes, one identical with the probe used Oligonucleotidr probes. The following oligonucleotide probes were by Moratalla and coworkers (1993). The two probes showed used in this study: c+.s, 48 bases long, complementary to rat c-fos identical results. No induction was found in rats treated with the mRNA encoding amino acids 137-152 of the c-Fos protein (Curran et lower doses of caffeine (50 mg/kg and 25 mg/kg). The c-jun al., 1987); c-&n, 45 and 48 bases long, complementary to mouse c-&n level was highest at one hour after treatment and gradually re- mRNA encoding amino acids 123-137 and 2655280 of the c-Jun pro- tein (Lamph et ai., 1988);junB, 48 bases long, complementary to mouse turned towards control, which was reached at 8 hr after caffeine iunB mRNA encoding amino acids 195-2 IO of the JunB orotein (Rvder administration. No induction of c-jun was observed in nucleus “et al., 1988); junD, 4i bases long, complementary to &D mRNA en- accumbens following administration of caffeine (100 mg/kg) coding amino acids 200-215 of the JunD protein (Hirai et al., 1989); (Figs. 1, 4a). preproenkephalin, 48 bases long, complementary to nucleotides 388- 435 of rat preproenkephalin gene (Yoshikawa et al., 1984); preprota- The expression of junB following caffeine administration was chykinin, 48 bases long, complementary to nucleotides 124-l 7 I of the more marked and long-lasting than that of c-fos and c-jun. Thus, rat preprotachkinin A gene (Krause et al., 1987). All probes were syn- there was a rapid and prominent induction both throughout the thesized by Scandinavian Gene Synthesis AB, Koping, Sweden. The caudate-putamen and in the nucleus accumbens following treat- specificity of each probe-has been tested earlier (Moratalla et al., 1993; Johansson et al., 1994; A. Dagerlind, personal communication). ment with caffeine (100 mg/kg) (Figs. 1, 2, 3b, 46). In both Oligodeoxyribonucleotides were radiolabeled using terminal deoxy- areas the highest levels of induction were observed I hr after nucleotidyl transferase (Pharmacia LKB, Uppsala, Sweden) and a-??- administration. The junB in caudate-putamen was significantly The Journal of Neuroscience, May 1995, 15(5) 3585
8hr.r
Figure 1. In situ hybridization autoradiograms showing the distribution of c-jun, junB, and c&s expression in coronal sections through the rostra1 part of striatum of rats treated with caffeine (100 mg/kg) and sacrificed after 30 min, 1 hr, 2 hr, 4 hr, or 8 hr. elevated at 4 hr after caffeine administration, but the induction at 1 hr after treatment and remained elevated for 2 hr. There was in the nucleus accumbens was less persistent. There was a ten- a tendency to a weak induction in rats treated with 50 mg of dency to induction ofjunB in rats treated with 50 mg of caffeine caffeine per kilogram and sacrificed 1 hr after administration, per kilogram and sacrificed 1 or 2 hr after administration, but whereas the c-fos signal in rats treated with the lowest dose of no induction was seen in rats treated with the lowest dose of caffeine (25 mg/kg) was indistinguishable from control. There caffeine (25 mg/kg). was only a weak and statistically insignificant induction of C--OS In agreement with previous studies (Nakajima et al., 1989; in nucleus accumbens 1 hr after administration of caffeine (100 Johansson et al., 1994) there was a rapid and prominent induc- mg/kg) (Figs. 1, 4~). tion of c-fos in striatum following administration of caffeine At the cellular level in the lateral part of rostra1 caudate-pu- (100 mg/kg), which was most pronounced in the lateral part of tamen, more neuron-like cells were labeled against junB and caudate-putamen (Figs. 1, 2, 3~). The c-fos signal was strongest c-fos than against c-&n (Fig. 5). No obviously glial-like cells 3586 Svenningsson et al. l Caffeine and immediate Early Genes
0.6 C-jZll? 0.5 3 2 0.4 1 a. 2 0.3
$ 0.2 0 0.1
0.0
30 min 1 hr 2 hrs 4 hrs 0 hrs junB 0.6 0.5 2 2 0.4
b. z 0.3 2. 3 0.2 .- 0.1
0.0 30 min 1 hr 2 hrs 4 hrs 0 hrs
0.6 r ***
0.5 3 $ 0.4
c. 6 0.3 9 $ 0.2 0 0.1
0.0 30 min 1 hr 2 hrs 4 hrs 0 hrs
Figure 3. Histograms showing the time-course of c-jun (a.),junB (b.), Figure 2. In situ hybridization autoradiograms showing the distribu- and c-fos (c.) expression in the lateral part of rostra1 caudate-putamen tion of c-jun, junB, and c-fos expression in coronal sections through the following treatment with saline (O), caffeine 25 mg/kg (m), caffeine caudal part of striatum of rats treated with saline (left hemisphere) or 50 mgkg (o), or caffeine 100 mg/kg (W). The results are given as the caffeine (100 mgkg) (right hemisphere) and sacrificed I hr after ad- mean + SEM, from three individuals in each group. *, P < 0.05; **, ministration. P < 0.01; ***, P < 0.001; each versus corresponding saline treated.
were labeledwith any of the probes.Only some25% of the cells and preprotachykinin A (SP) expressingcells. Thus, c-jun was in the study illustrated in Figure 5c showedlabeling with c-jun. found in 27% of the ENK positive cells, 22% of the ENK neg- The degreeof labeling was somewhathigher (40%) in a separate ative cells, 27% of the SP positive cells, and 21% of the SP experiment.By contrast, thejunB and c-fos probeswere detected negative cells. For junB the correspondingfigures were 80, 70, in a muchlarger percentageof cells 80% (84%) and 67% (77%), 84, and 86%, and for c-fos they were 79, 65, 68, and 64% (20- respectively. The difference in the degree of labeling could not 88 cells were counted in each category-the largestnumbers in be attributed to a difference in the population of cells labeled. caseof c-jun). Instead all three probes labeled both preproenkephalin(ENK) None of the treatmentsin this experimentcaused induction of The Journal of Neuroscience, May 1995, 15(5) 3587 junD. The probe did, however, reveal a clear expression of junD in hippocampus (not shown) and cerebellum (Fig. 6), but this was not different from that seen in saline-treated animals. 0.18 3 Effect of a single injection of caffeine on expressionof c-jun 3 0.15 in ventral part of thalarnus,junB in cerebral cortex, and c-fos in globus pallidus a. 4 0.12 An increased level of c-jun in the ventral part of thalamus, es- 2, pecially in ventral posteromedial and ventral posterolateral tha- 2 0.09 lamic nucleus, was observed following treatment with caffeine ‘7 (100 mg/kg) (Fig. 7), but no similar induction was found for the o 0.06 1 other oncogenes examined. An increased level of junB and c-fos, 0.03 but not of c-jun, in the cerebral cortex was seen following ad- ministration of caffeine (100 mg/kg) (Fig. 2). The junB and c-fos increase was more clearcut in posterior parts of the cortex (not 30 min 1 hr 2 hrs 4 hrs 8 hrs shown). Interestingly, c-fos tended to increase in globus pallidus following treatment with caffeine (100 mg/kg) (Fig. 2), but no increase in the other immediate early gene products was ob- served in this structure. *** EfSectof a single injection of caffeine on expressionof AP-I in striatum Caffeine (100 mg/kg) administration caused an induction of AP- 1, which was clearly detectable already after 1 hr, peaked after 2 hr, and was well maintained 4 hr after administration (Figs. 2 0.09 Sa, 9). By contrast, caffeine did not affect binding to CRE-2 at any of these time-points (Fig. Sd, results from rats treated 4 hr ” 0.06 not shown). The experiment was repeated five times with new extracts from different rats each time and the results were iden- 0.03 tical. Furthermore, identical results were obtained with two dif- ferent probes against the AP-1 binding site. The results were not 30 min 1 hr 2 hrs 4 hrs 8 hrs due to unspecific DNA-binding, since an excess of cold NF-1 (Fig. SC) or CRE-2 (not shown) did not alter the AP-1 binding following caffeine treatment. Using super-shift assay it was dem- 0.21 onstrated that at least some of the AP-1 complexes after 2 hr contained JunB, c-Jun and c-Fos (Fig. Sb). The presence of JunD 0.18 c in AP-1 was also suggested. The same results were found in striatal extracts from four different rats injected with caffeine (100 mg/kg) and sacrificed 2 hr after administration. The AP- 1 complex obtained from animals 4 hr after caffeine was shifted by antibodies against Jun B and c-Fos (Fig. 86). Effect of a single injection of caffeine on preproenkephalin mRNA 1 0.06 In agreement with previous results (Schiffmann et al., 1991) 0.03 preproenkephalin mRNA was expressed throughout the caudate putamen and the nucleus accumbens (Fig. 10d). There was a clear-cut increase of preproenkephalin mRNA in the lateral part 30 min 1 hr 2 hrs 4 hrs 8 hrs of caudate-putamen following administration of caffeine (100 Figure 4. Histograms showing the time-course of c-&n (a.),junB (b.), mg/kg) (Figs. 10, 11, 12a). A small induction was seen already and c-fos (c.) expression in nucleus accumbens following treatment with after 1 hr, but it peaked after 4 hr, and remained significantly saline (U), caffeine 25 mg/kg (@l), caffeine 50 mg/kg (m), or caffeine increased even after 8 hr. The expression of preproenkephalin 100 mgfl
Figure 6. In situ hybridization autoradiograms showing the expression of junD in coronal sections from the rostra1 part of striatum (upper panel) and cerebellum (lower panel) of a caffeine-treated rat (100 mg/kg), sacrificed 1 hr after treatment.
anything causea motor depression.This could mean that our results are not immediately relevant for an understandingof the neuronal basisof the excitatory actions of caffeine. Also other drugs, for example, cocaine and amphetamine,are behaviorally active in much lower dosesthan those required to stimulateim- mediate early gene expression (e.g., Moratalla et al., 1992, 1993). Whereasinformation about changesin gene expressionafter caffeine treatment is limited, the ability of addictive compounds Figure 5. Emulsion autoradiograms from the lateral part of rostra1 caudate-putamen of a caffeine-treated rat (100 mg/kg), sacrificed 1 hr suchas cocaine,amphetamine, and opiatesto alter geneexpres- after treatment, showing neuron-like cells labeled in thin, consecutive sion has beenextensively studied(see Nestler et al., 1993).Since sections against preproenkephalin mRNA (ENK, fop leff) or preprota- caffeine sharesat least somebehavioral characteristicswith am- chykinin A mRNA (SP, fop right) and, for comparison, with probes for c-jun (panel a, bottom two parts), junB (panel b, bottom two parts) and c-fos (panel c, bottom two parfs). Arrows indicate cells that coexpress mRNA for the respective peptides and immediate early genes. Of the neuron-like cells examined 52% were positive for ENK, 53% for SR 25% for c-jun, 80% for junB, and 67% c-jos. Between119 and 282 cellswere counted to obtaineach percentage value. genesare preferentially induced in some areas;this could pro- vide an indicator of the neural systemsthat are activated. In agreementwith previous findings (Nakajima et al., 1989; Johanssonet al., 1994)the induction of C--OSwas seenonly after rather high dosesof caffeine. Thesedoses of caffeine are much (more than lo-fold) higher than thosethat causeinitial activation Figure 7. In situ hybridization autoradiograms showing the expression of c-jun in the ventral part of thalamus of rats treated with saline (left of motor behavior (see Daly, 1993). In fact, the dosesrequired hemisphere) or caffeine (100 mglkg) (right hemisphere) and sacrificed to induce a clearcut induction of the immediate early genesif after 1 hr. The Journal of Neuroscience, May 1995, 75(5) 3589
, 2%. s b . Figure 8. a, Gel-shift assay showing the expression of AP-1 in striatum of rats treated with saline or caffeine (100 mg/kg). The AP-I binding site was de- duced from the collagenase enhancer. To every second lane, a lOO-fold ex- cess of unlabeled AP-1 binding site was added. The lowermost part of the gel shows free probe. b, Gel-shift assay showing the expression of AP-1 in striatum of rats treated with caffeine (100 mg/kg) and sacrificed after 2 (lanes 1-6) or 4 (lanes 7-12) hr. An- tibodies against JunB, JunD, cJun, or qlJbLI)r -1 c-Fos proteins were added to the indi- cated lanes. c, Control experiment showing the effect of increasing amounts of unlabeled NF-I binding sites, unlabeled AP- 1 (derived from SV 40 promotor or from the collagenase promotor) on the binding of labeled AP-I from collagenase. d, Gel-shift as- say showing the binding to ENKCRE-2 of extracts prepared from striatum of rats treated with saline or caffeine (100 mg/kg). The extracts from caffeine- treated rats were known to include AP- 1; this was shown by a parallel exper- iment where we used AP-I site from the collagenase enhancer. Competition with unlabeled NF-1 and ENKCRE-2 were made with extracts from rats treated with caffeine for I hr. phetamine and cocaine (Daly, 1993) it may be of interest to as neither amphetaminenor cocaineinduces c-&n (Moratalla et comparetheir actions on immediateearly geneexpression. There al., 1990, 1993).This qualitative difference betweencocaine and are somesimilarities, for example, in that c-fis is induced in the caffeine may be related to the quantitative difference between caudate-putamen(Nakajima et al., 1989), but also important dif- the ability of the substancesto induce C--OS.It is known that ferences. For example, we show, using two different oligonu- cocaineinduces c-fos almost exclusively in dopamineD, recep- cleotide probes,that caffeine is a potent inducer of c-jun, where- tor and substanceP-expressing medium-sized spiny neurons 3590 Svenningsson et al. l Caffeine and Immediate Early Genes
0.5
0.4 3 ‘2 3 0.3 Ctrl 6 z 0.2
4 0.1 0.0 -i---M r- r 1 hr 2 hrs 4 hrs
Figure 9. Histogram showing the time course of AP-1 expression in striatum following saline (0) or caffeine (m) treatment (100 mg/kg). The AP-1 expression was quantitated using image analysis. The findings were replicated twice using the same probe (derived from SV 40 prom- otor) and twice with a different probe (derived from the collagenase 2 hrs promotor) with qualitatively identical results.
(Graybiel et al., 1990; Young et al., 1991; Beretta et al., 1992; Cole et al., 1992), whereascaffeine inducesc-fos also in neurons that express met-enkephalin(Johansson et al., 1994). Approxi- mately half of the medium-sizedspiny neurons in striatum be- long to the class that express large amounts of dopamine D, receptor mRNA, whereas the other half of the cells predomi- nantly express dopamine D, receptor and preproenkephalin . mRNA (Gerfen et al., 1990). The present results show that both c-f& andjunB increasein considerably more than half the cells 4hrs of the caudate-putamen,and not only in the subpopulationof cells expressingabundant dopamine D, receptor mRNA but also in the other, equally large, subpopulationof cells. c-jun was also induced in both the subpopulationsof cells, albeit to a much more limited extent. The finding that the examined immediate early genesshow somewhatdifferent spatiotemporalcharacteristics might also be related to in vitro studiesthat have shown that fos and jun on- cogenesare differently regulated(e.g., Bartel et al., 1989; Gubits et al., 1990; Worley et al., 1990). For example, the 5’-flanking sequencesof c-fos and junB genespossess elements responsive to CAMP and protein kinaseA, whereasthat of c-jun lacks these elements(Angel et al., 1988; Lamph et al., 1988; Sassone-Corsi 8hrs et al., 1988; Chiu et al., 1989; de Groot et al., 1991). It has been proposedthat the coordinate expressionof c-fos and junB fol- lowing cocaine administration could be the result of an activa- tion of a calcium- and CAMP responseelement secondary to the activation of dopamineD, receptors(Graybiel et al., 1990; Mor- atalla et al., 1993). The lack of effect of cocaine on c-jun ex- Figure 10. In situ hybridization autoradiograms showing the distri- bution of preproenkephalin mRNA expression in coronal sections pressionwas suggestedperhaps to be due to active repression through the rostra1 part of striatum of rats treated with saline and sac- by CAMP or by junB (Moratalla et al., 1993). Since caffeine rificed after 2 hr (Ctrl) or treated with caffeine (100 mg/kg) and sac- could induce c-jun, c-fos, and junB in the sametypes of cells, rificed after 2, 4, or 8 hr. it could perhapsbe concludedthat the caffeine effect is unrelated to CAMP Indeed, despitethe fact that caffeine is inter alia an inhibitor of CAMP breakdown, its effects on C--OSinduction can- Effects of caffeine on immediateearly genesin brain areas not be mimicked by selective phosphodiesteraseinhibitors related to caudate-putamen (Svenningsson,Johansson, and Fredholm, unpublished obser- Nucleus accumbensreceives an important dopaminergicinner- vations). vation from the ventral tegmental area, which is proposedto The Journal of Neuroscience, May 1995, S(5) 3591
* z 0.7 c T
1 hr 2 hrs 4 hrs 0 hrs
0.42 r-
3 0.39 yz 1 0.36
b . % 0.33 2 “E 0.30
!$ 0.27 Figure II. In situ hybridization autoradiograrns showing the distri- bution of preproenkephalin mRNA expressionfrom coronalsections 0.24 through the caudalpart of striatumof rats treatedwith caffeine(100 mg/kg) (upperpanel) or saline(lower panel) andsacrificed after 4 hr. 1 hr 2 hrs 4 hrs 0 hrs
Figure 12. Histograms showing preproenkephalin (WE) mRNA ex- play key role in the development of drug abuse(Koob, 1992). pression in the lateral part of rostra1 caudate-putamen (a) and the nu- We previously found that caffeine, in contrast to cocaine and cleus accumbens (b) following treatment with saline (Cl), caffeine 25 amphetamine,gives little or no induction of c-j& in this nucleus mg/kg (LZl), caffeine 50 mg/kg (8). or caffeine 100 mg/kg (W). The (Johanssonet al., 1994). Acute cocaine was found to causea results are given as the mean f. SEM from three individuals in each group. *, P < 0.05, each versus corresponding saline treated. marked increasein i.a. junB and c-jun in nucleusaccumbens as judged by Northern blotting (Hope et al., 1992), whereas the induction asjudged by in situ hybridization was modestin mag- that decreasethe releaseof glutamate(Fredholm and Dunwiddie, nitude or absent(Moratalla et al., 1993). The presentfinding that 1988). administration of caffeine (100 mg/kg) caused a clearcut in- creaseof junB in at least part of nucleusaccumbens indicates Caffeine induces c-Fos-, c-Jun-, and JunB-containingAP-1 that this region is affected by caffeine. Actions in the nucleus and preproenkephalinmRNA accumbensmay be of significance since caffeine has been de- The productsof the immediateearly genesof thefosljun family scribed as a weak reinforcing drug (Griffiths and Woodson, are known to associatein a variety of dimers that bind to the 1988; Holtzman and Finn, 1988). AP-1 site known to regulatethe expressionof many genes.Spe- A possibleexplanation for the ability of caffeine to produce cifically, it has been shown that induction of c-fos, c-jun, junB a quantitatively and qualitatively different increasein immediate mRNA precedes induction of proenkephalin mRNA in hippo- early genesthan doescocaine might be that cocaineacts directly campus(Sonnenberg et al., 1989a). These authors also showed in striatum, whereascaffeine activates it largely indirectly. Un- that synthetic c-Fos/c-Junand c-Fos/JunB complexesbind to the publisheddata indicate that the induction of c-fos is alteredwhen AP-l-like site called ENKCRE-2 in the 5’-flanking sequenceof rats are pretreated with glutamate receptor antagonists(Sven- the preproenkephalin gene. In another paper from the same ningsson, Johansson,and Fredholm, unpublished data). One group (Sonnenberget al., 1989b) it was shown that AP-1 bind- probable explanation is that a componentof the effects of caf- ing activity was induced, and it was suggestedthat the compo- feine is due to actions on glutamatergic afferent pathways from sition of the transcription factor may vary with time. Given this striatum. In this study we report an increaseof c-jun in the background, which was available to us at the start of our study, ventral part of thalamus and of junB in cerebral cortex by caf- the data presented in the present study imply that c-fos, c-jun, feine; both theseareas interact intimately with striatum (for re- andjunB, via formation of AP-1, contribute to the regulation of views, see Albin et al., 1989; Gerfen et al., 1992). The ventral enkephalinexpression in the caudateputamen following caffeine part of thalamusprovides an important striatal output and pro- exposure. Thus, AP-1 binding activity was detected somewhat vides input to cerebral cortex. The mechanismis probably re- after the first induction of the immediateearly gene mRNA. It lated to the inhibition, by caffeine, of adenosineA, receptors was also shown, by super-shift assay, that at least part of the 3592 Svenningsson et al. * Caffeine and Immediate Early Genes
AP-1 binding activity was composedof JunB, c-Fos, and c-Jun. (1988) Proteins bound at adjacent DNA elements act synergistically Finally, and with a still greater time lag, there was an increased to regulate human proenkephalin CAMP inducible transcription. EMBO J 7:3793-3805. expressionof preproenkephalinmRNA. However, in a careful Curran T, Gordon MB, Rubino KL, Sambucetti LC (1987) Isolation study, Fos-containing AP-1 was found to play little or no role and characterization of the c-fos (rat) cDNA and analysis of post- in the induction of proenkephalinby haloperidoltreatment (Kon- transcriptional modification in vitro. Oncogene 2:79-84. radi et al., 1993). Instead, the study provided evidence that Curran T, Abate C, Cohen DR, Macgregor PF, Rauscher FJ III, Son- nenberg JL, Connor JA, Morgan JI (1990) Inducible proto-oncogene CREB binds and regulates via phosphorylation of the ENK- transcription factors: third messengers in the brain? Cold Spring Har- CRE-2 site. We similarly found that caffeine did not alter bind- bor Symp Quant Biol 55:225-234. ing to ENKCRE-2, using the same sequenceas that used by Daly JW (1993) Mechanism of action of caffeine. In: Caffeine, coffee Konradi et al. (1993). Therefore the causal link between the and health (Garattini S, ed), pp 977150. New York: Raven. induction of AP-1 and the induction of preproenkephalinmRNA deceballos ML, Boyce S, Jenner P Marsden CD (1986) Alterations in [metsI- and [leu5]-enkephalin and neurotensin content in the basal is highly uncertain. In the caseof two other neuropeptides,neu- ganglia induced by the long-term administration of dopamine agonist rotensin and substanceP, the role of Fos and AP- 1 may be bigger and antagonist drugs to rats. Eur J Pharmacol 130:305-309. (Merchant and Miller, 1994). These two peptides are also af- de Groot RP, Auwerx J, Karperien M, Staels B, Kruijer W (1991) Ac- fected by caffeine (Schiffmann and Vanderhaegen 1993b; P tivation of junB24 by PKC and PKA signal transduction through a Svenningsson,U. Aden, and B. B. Fredholm, unpublisheddata). novel cis-acting element. Nucleic Acids Res 19:7755781. Dragunow M, Robertson HA (I 987) Kindling stimulation induces c-fos Interestingly, in the 5’-flanking sequenceof substanceP, one protein(s) in granule cells of the rat dentate gyrus. Nature 329:441- detects an AP-1 binding site identical to the AP-1 site in the 442. collegenaseenhancer, which we usedin this experiment (Carter Fastbom J, Pazos A, Palacios JM (1987) The distribution of adenosine and Krause, 1990). A, receptors and 5’-nucleotidase in the brain of some commonly used The present data thus show that a single doseof caffeine can experimental animals. Neuroscience 22:8 13-826. Fredholm BB (1980) Are methylxanthine effects due to antagonism of induce marked changesnot only in immediateearly gene ex- endogenous adenosine? Trends Pharmacol Sci 1:129-132. pressionbut also in the expressionof genesthat respondto the Fredholm BB, Dunwiddie TV (1988) How does adenosine inhibit transcription factors that thesegenes elaborate. It is tempting to transmitter release? Trends Pharmacol Sci 9: 130-l 34. speculatethat such changesmay underlie the well-known adap- Fredholm BB, Abbracchio MP Burnstock G, Daly JW, Harden TK, tive changesthat occur in brain following caffeine treatment. Jacobson KA, Leff P Williams M (1994) Nomenclature and classi- fication of purinoceptors: a report from the IUPHAR subcommittee. For example, in naive subjectsa dose of caffeine activates the Pharmacol Rev 46: 143- 156. sympathoadrenalsystem resulting, for example, in increasesin Georgiev V, Johansson B, Fredholm BB (1993) Long-term caffeine blood pressure. However, subsequentdoses produce progres- treatment leads to a decreased suspectibility to NMDA-induced clo- sively smaller effects and within a few days no effect is ob- nic seizures in mice without changes in adenosine A, receptor num- ber. Brain Res 6121271-277. served (Robertson et al., 1981). Furthermore, there is evidence Gerfen GR (1992) The neostriatal mosaic: multiole levels of comuart- for behavioral tolerance (see Holtzman and Finn, 1988) and a mental organization. Trends Neurosci 15:133-i39. few days of caffeine treatment in rodents significantly alterstheir Gerfen CR, Young WS III (1988) Distribution of striatonigral and stria- susceptibility to drug-induced seizures (Georgiev et al., 1993) topallidal peptidergic neurons in both patch and matrix compart- and to damagefollowing global or regional ischemia(Rudolphi ments: an in situ hybridization histochemistry and fluorescent retro- grade tracing study. Brain Res 460: 161-l 67. et al., 1989). Gerfen CR, Engber TM, Mahan LC, Susel Z, Chase TN, Monsma FJ Jr, Sibley DR (1990) Dl and D2 dopamine receptor-regulated gene References expression of striatonigral and striatopallidal neurons. Science 250: Albin RL, Young AB, Penney JB (1989) The functional anatomy of 1429-1432. basal ganglia disorders. Trends Neurosci 12:366-375. Gerfen CR, McGinty JE Young WS III (1991) Dopamine differentially Angel P Hattori K, Smeal T, Karin M (1988) The jun proto-oncogene regulates dynorphin, substance P and enkephalin expression in stri- is positively autoregulated by its product, Jun/AP-1. Cell 55:875-885. atal neurons: in situ hybridization histochemical analysis. J Neurosci Angulo JA (1992) Involvement of dopamine D, and D, receptors in 11:1016-1031. the regulation of pro enkephalin mRNA abundance in the striatum Graybiel AM, Moratalla R, Robertson HA (1990) Amphetamine and and accumbens of the rat brain. J Neurochem 58:1104-l 109. cocaine induce drug-specific activation of the c-fos gene in striosome- Bartel DP, Sheng M, Lau LE Greenberg ME (1989) Growth factors matrix compartments and limbic subdivisions of the striatum. Proc and membrane depolarization activate distinct programs of early re- Nat1 Acad Sci USA 87:6912-6916. sponse gene expression: dissociation of& and jun induction. Genes Griffiths RR, Woodson PP (1988) Reinforcing properties of caffeine: Dev 3:304-313. studies in humans and laboratory animals. Pharmacol Biochem Behav Berretta S, Robertson HA, Graybiel AM (1992) Dopamine and gluta- 29:419-427. mate agonists stimulate neuron-specific expression of Fos-like protein Gubits RM, Wollack JB, Yu H, Liu W-K (1990) Activation of aden- in the striatum. J Neurophysiol 68:767-777. osine receptors induces c-fos, but not c-jun, expression in neuron- Bradford MM (1976) A raoid and sensitive method for the auantitation glia hvbrids and fibroblasts. Mol Brain Res 8:275-281. of microgram quantities of protein utilizing the principle of protein- HeTdegen T, Tolle TR, Bravo R, Zieglgansberger W, Zimmermann M dye binding. Anal Biochem 72:248%254. (1991a) Sequential expression of JunB, JunD and FosB proteins in Carter MS, Krause JE (1990) Structure, expression, and some regula- rat spinal neurons: cascade of transcriptional operations during no- tory mechanisms of the rat preprotachykinin gene encoding substance ciception. Neurosci Lett 129:221-224. P neurokinin A, neuropeptide K and neuropeptide y. J Neurosci 10: Herdegen T, Leah JD, Manisali A, Bravo R, Zimmermann M (1991b) 2203-2214. c-Jun-like immunoreactivity in the CNS of the adult rat: basal and Chiu R, Angel P Karin M (1989) Jun-B differs in its biological prop- transynaptically induced expression of an immediately-early gene. erties from, and is a negative regulator of, c-jun. Cell 59:979-986. Neuroscience 41:643-654. Chodosh LA, Baldwin AS, Carthew RW, Sharp PA (1988) Human Hirai S-I, Ryseck RP Mechta F, Bravo R, Yaniv M (1989) Character- CCAAT-binding proteins have heterologous subunits. Cell 53: 1 l-24. ization of junD: a new member of the jun proto-oncogene family. Cole AJ, Bhat RV, Patt C, Worley PF, Baraban JM (1992) D, dopamine EMBO J 8:1433-1439. receptor activation of multiple transcription factor genes in the rat Holtzman SC, Finn IB (1988) Tolerance to behavioral effects of caf- striatum. J Neurochem 58:1420-1426. feine in rats. Pharmacol Biochem Behav 29:41 l-418. Comb M, Mermod N, Hyman SE, Pearlberg J, Ross ME, Goodman HM Hope B, Kosofsky B, Hyman SE, Nestler EJ (1992) Regulation of The Journal of Neuroscience, May 1995, 15(5) 3593
immediate early gene expression and AP-1 binding in the rat nucleus damage in gerbil hippocampus is reduced following upregulation of accumbens by chronic cocaine. Proc Nat1 Acad Sci USA 89:5764- adenosine (A,) receptor by caffeine treatment. Neurosci Lett 103: 5768. 275-280. Johansson B, Lindstrom K, Fredholm BB (1994) Differences in the Ryder K, Lau LE Nathans D (1988) A gene activated by growth factors regional and cellular localization of c-fos mRNA induced by am- is related to the oncogene v-jun. Proc Nat1 Acad Sci USA 85:1487- phetamine, cocaine and caffeine. Neuroscience 59:837-849. 1491. Jones NC, Rigby PWJ, Ziff EB (I 988) Truns-acting protein factors and Saffen DW, Cole AJ, Worley PE Christy BA, Ryder K, Baraban JM the regulation of eukaryotic transcription: lessons from studies on (1988) Convulsant-induced increase in transcription factor messenger DNA tumor viruses. Genes Dev 2:267-28 1. RNAs in rat brain. Proc Nat1 Acad Sci USA 85:7795-7799. Konradi C, Kobierski LA, Nguyen TV, Heckers S, Hyman SE (1993) Sagar SM, Sharp FR, Curran T (1988) Expression of c-fos protein in The CAMP-response-element-binding protein interacts, but Fos pro- brain: metabolic mapping at the cellular level. Science 240:1328- tein does not interact, with the preproenkephalin enhancer in rat stria- 1331. turn. Proc Nat1 Acad Sci USA 90:7005-7009. Sassone-Corsi P Lamph WW, Verma IM (1988) Regulation of proto- Koob GF (1992) Drugs of abuse: anatomy, pharmacology and function oncogenefos: a paradigm for early response genes. Cold Spring Har- of reward oathwavs. Trends Pharmacol Sci 13: 177-184. bor Symp Quant Biol 53:749-760. Krause JE, Chirgwin JM, Carter MS, Xu S, Hershev AD (1987) Three Schiffmann SN, Vanderhaeghen J-J (1993a) Adenosine A, receptors rat preprotachykinin mRNAs encode the neuropeptides substance P regulate the gene expression of striatopallidal and striatonigral neu- and neurokinin A. Proc Nat1 Acad Sci USA 84:881-885. rons. J Neurosci 3:1080-1087. Lamph WW, Wamsley P, Sassone-Corsi P, Verma IM (1988) Induction Schiffmann SN, Vanderhaeghen J-J (199313) Caffeine regulates neu- of-proto-oncogene jun/AP- 1 by serum and TPA. Nature 334:629-63 1. rotensin and cholecystokinin messenger RNA expression in the rat Lee W. Mitchell P. Tiian R (1987) Purified transcriotion factor AP-1 striatum. Neuroscience 54:681-689. interacts with TPAIinducibie enhancer elements. &II 49:741-752. Schiffmann SN, Jacobs OP, Vanderhaeghen J-J (1991) Striatal restrict- Mellstrom B, Achaval M, Montore D, Naranjo JR, Sassone-Corsi P ed adenosine A, receptor (RDC8) is expressed by enkephahn but not (1991) Differential expression of thejun family members in rat brain. by substance P neurons: an in situ hybridization histochemistry study. Oncogene 6: 1959-1964. J Neurochem 57:1062-1067. Merchant KM, Miller MA (1994) Coexpression of neurotensin and Sonnenberg JL, Rauscher FJ III, Morgan JI, Curran T (1989a) Regu- c-fos mRNAs in rat neostriatal neurons following acute haloperidol. lation of proenkephalin by Fos and Jun. Science 246:1622-1625. Mol Brain Res 23:271-277. Sonnenberg JL, Macgregor-Leon PE Curran T, Morgan JI (I 989b) Dy- Moratalla R, Robertson HA, DiZio PA, Graybiel A (1990) Parallel namic alterations occur in the levels and composition of transcription induction of junB and c-fos evoked in the striatum by the psycho- factor AP-1 complexes after seizure. Neuron 3:359-365. motor stimulant drugs cocaine and amphetamine. Sot Neurosci Abstr Tang E Costa E, Schwartz J (1983) Increase of proenkephalin mRNA 16:953. and enkephalin content of rat striatum after daily injection of halo- Moratalla R, Robertson HA, Graybiel AM (1992) Dynamic regulation peridol for 2 to 3 weeks. Proc Nat1 Acad Sci USA 80:3841-3844. of NGFZ-A (~$268, egrl) gene expression in the striatum. J Neurosci Taylor MD, DeCeballos ML, Jenner P Marsden CD (1991) Acute ef- 12:2609-2622. fects of d-l and d-2 dopamine receptor agonist and antagonist drugs Moratalla R, Vickers EA, Robertson HA, Cochran BH, Graybiel AM on basal ganglia [met5]- and [leu5]-enkephalin and neurotensin con- (1993) Coordinate expression of c-fos and junB is induced in the rat tent in the rat. Biochem Pharmacol 41:1385-1391. striatum by cocaine. J Neurosci 13:423-433. Thierry E Spyrou G, Yaniv M, Howley P (1992) Two AP-1 sites bind- Morgan JI, Cohen DR, Hempstead JL, Curran T (1987) Mapping pat- ing JunB are essential for human papillomavirus type 18 transcription terns of clfos expression in the central nervous system after seizure. in keratinocytes. J Virol 66:3740-3748. Science 237: 192-197. Uhl GR, Walther D, Nishimoti T, Buzzi MG, Moskowitz MA (1991) Nakabeppu Y, Ryder K, Nathans D (1988) DNA binding activities of junB, c-jun, junD and c-fos mRNAs in nucleus caudalis neurons: three murine jun proteins: stimulation of Fos. Cell 55:907-915. rapid selective enhancement by afferent stimulation. Mol Brain Res Nakajima T, Daval J, Morgan PE Post RM, Marangos PJ (1989) Ad- 11:133-141. enosinergic modulation of caffeine-induced c-fos mRNA expression Wang X-B, Watanabe Y, Osugi T, Ikemoto M, Hirata M, Miki N (1992) in mouse brain. Brain Res 501:307-314. In situ DNA-protein binding: a novel method for detecting DNA- Nehlig A, Daval J-L, Debry G (1992) Caffeine and the central nervous binding activity of transcription factor in brain. Neurosci Lett 146: system: mechanisms of action, biochemical, metabolic and psycho- 25-28. stimulant effects. Brain Res Rev 17:139-170. Wisden W, Errington ML, Williams S, Dunnett SB, Waters C, Hitchcock Nestler EJ, Hope BT, Widnell KL (1993) Drug addiction: a model for D, Evan G, Bliss TVP, Hunt SP (1990) Differential expression of the molecular basis of neural plasticity. Neuron 11:995-1006. immediately early genes in the hippocampus and spinal cord. Neuron Nguyen TV, Kosofsky BE, Birnbaum R, Cohen BM, Hyman SE (1992) 4:603-614. Differential expression of c-Fos and Zif 268 in rat striatum after Worley PE Cole AJ, Murphy TH, Christy BA, Nakabeppu Y, Baraban haloperidol, clozapine, and amphetamine. Proc Nat1 Acad Sci USA JM (1990) Synaptic regulation of immediate-early genes in brain. 89:427&4274. Cold Spring Harbor Symp Quant Biol 55:213-223. Parkinson FE, Fredholm BB (1990) Autoradiographic evidence for Yoshikawa K, Williams C, Sabol SL (1984) Rat brain preproenkephalin G-protein coupled A,-receptors in rat neostriatum using [‘HI-CGS mRNA. J Biol Chem 259:14301-14308. 21680 as a ligand. Naunyn-Schmiedeberg’s Arch Pharmacol 342:85- Young ST, Porrino LJ, Iadanola MJ (1991) Cocaine induces striatal 89. c-Fos immunoreactive proteins via dopaminergic D, receptors. Proc Robertson D, Wade D, Workman R, Woosley RL, Oates JA (1981) Nat1 Acad Sci USA 88:1291-1295. Tolerance to the humoral and hemodynamic effects of caffeine in Young WS III, Bonner TI, Brann MR (1986) Mesencephalic dopamine man. J Clin Invest 67:lll l-l 117. neurons regulate the expression of neuropeptide mRNAs in the rat Rudolphi KA, Keil M, Fastbom J, Fredholm BB (1989) Ischaemic forebrain. Proc Nat1 Acad Sci USA 83:9827-9831.