Effects of Chronic Caffeine on Adenosine, Dopamine and Acetylcholine Systems in Mice

Abstract

Chronic ingestion of caffeine by male NIH Swiss strain mice leads in about 3 days to a significant increase in A₁-adenosine, nicotinic and muscarinic receptors, and a significant decrease of β₁-adrenoceptors in cerebral cortical membranes. Plasma levels of caffeine in the chronically treated mice range from 0.70 to 5.7 μg/ml. The changes in receptors reverse after withdrawal of caffeine within 7 days. An increase in nitrendipine binding sites, associated with L-type calcium channels, also occurs within 4 days and has reversed in 7 days after withdrawal. There is no change in the levels of striatal nicotinic receptors or D₂-dopamine receptors, nor of [³H]cocaine binding to dopamine uptake sites. Levels of opioid receptors are either increased (δ) or unaltered (μ, κ). σ-Receptors are unaltered. Stimulations of striatal adenylate cyclase by forskolin, dopamine and NECA are not significantly affected after chronic caffeine ingestion.

The adenosine agonist, NECA, reverses the amphetamine-elicited increases in locomotor activity and partly reverses the cocaine-elicited increases. The NECA dose-response curve is multiphasic (depression, stimulation and then depression) versus amphetamine in control mice, but only depressant versus amphetamine in chronic caffeine mice, while being multiphasic versus cocaine in both control and chronic caffeine mice.

NECA reverses the stimulation of locomotor activity elicited by the muscarinic antagonist, scopolamine, and is more effective in the chronic caffeine mice. The behavioral depressant effects of the muscarinic agonist, oxotremorine, are not markedly altered after chronic caffeine ingestion.

The behavioral depressant effects of nicotine are abolished after chronic caffeine ingestion, while the behavioral depressant effects of the nicotinic antagonist, mecamylamine, are not markedly altered after chronic caffeine ingestion. In combination with caffeine, nicotine has depressant effects in control mice, while having biphasic effects, consisting of initial stimulation followed by depression, in chronic caffeine mice.

Introduction

Caffeine is well known as a behavioral stimulant in man although its effects show extreme, individual variations. Because of such variations, conclusions as to development of tolerance, dependence and withdrawal in man have been inconclusive (Nehlig et al., 1992). The basis for tolerance and withdrawal has been investigated extensively in animal models, where chronic administration of caffeine leads to alterations in the levels of various central receptors, and in behavioral responses to caffeine, adenosine analogues, cholinergic agents, convulsants and other centrally active agents (Daly, 1993). Tolerance to the behavioral effects of caffeine has been frequently reported, although the opposite, a sensitization, has also been observed (Meliska et al., 1990; Nikodijević et al., 1993a). In male NIH Swiss strain mice, where tolerance does not develop, both biochemical and behavioral alterations during chronic ingestion of caffeine and withdrawal have been extensively studied (Nikodijević et al., 1993a, 1993b, 1993c; Shi et al., 1993; Daly et al., 1994). A variety of central receptors and L-type calcium channels are upregulated in this mouse strain after chronic caffeine ingestion (Shi et al., 1993). Behaviorally, after an initial increase in open-field locomotor activity, chronic ingestion of caffeine leads to a significant decrease of activity (Nikodijević et al., 1993a). Tolerance to caffeine does not develop in this mouse strain. Effects of adenosine analogues and of cholinergic agents on open-field locomotor activity are altered after chronic caffeine ingestion, while effects of agents, such as cocaine and amphetamine, that are thought to act via dopaminergic pathways, are not markedly altered (Nikodijević et al., 1993a, 1993b).

The time course for alterations of the levels of receptors and L-type calcium channels has now been determined during chronic caffeine ingestion and withdrawal in male NIH Swiss strain mice. Further biochemical and behavioral studies have been conducted on the function of the adenosine, dopamine and cholinergic systems.

Methods

Animals

Adult male mice of the NIH Swiss strain, weighing 25–30 g, were used. Mice were 5 weeks old when received. They were kept for 5–7 days before the start of any experiment. The groups of 10–11 mice per cage were housed and kept on a 12 hr dim lighting: 12 hr dark cycle in the animal holding room. Mice were given free access to standard food pellets and water. In chronic caffeine-treated groups, a caffeine solution (1 g/l unless otherwise noted) was provided instead of drinking water during the time prescribed by the experiment. Before behavioral testing, mice were habituated for 24 hr in the laboratory monitoring room on a 12 hr dim lighting: 12 hr dark cycle. Each mouse was used only once in the activity monitor, except for the time-dependency experiments, were the same animals were monitored for locomotor activity on successive days. Each caffeine group had a control group for behavioral testing. Liquid intake was similar for control and caffeine groups, as was weight gain and general appearance (Nikodijević et al., 1993a).

Drugs

[³H]N⁶-Cyclohexyladenosine (Specific activity: 30.2 Ci/mmol), [³H]8-cyclopentyl-1,3-dipropylxanthine (108 Ci/mmol), [³H]CGS 21680 (47 Ci/mmol), [³H]dihydroalprenolol (58 Ci/mmol), [³H]nicotine (75 Ci/mmol), [³H]quinuclidinylbenzilate (46 Ci/mmol), [³H]nitrendipine (71 Ci/mmol), [³H]raclopride (71 Ci/mmol), [³H]DAMGO (53 Ci/mmol), [³H]DPDPE (46 Ci/mmol), [³H]U69593 (153 Ci/mmol), [³H]DTG (38 Ci/mmol) and [³H]cocaine (30 Ci/mmol) were from New England Nuclear (Boston, MA, U.S.A.). Caffeine (free base) was from Matheson, Coleman and Bell (Cincinnati, OH, U.S.A.). The remaining compounds were from Research Biochemicals Inc. (Natick, MA, U.S.A.) and other commercial sources.

Membrane preparation

The mice were killed by cervical fracture and the brains rapidly removed into ice-cold 50 mM Tris-HCl buffer (pH 7.4). Cerebral cortex (5 mice), cerebellum (4 mice) or striatum (10 mice) were dissected from mice of either control or caffeine groups. Tissue was homogenized in 3 ml of ice-cold 50 mM Tris-HCl buffer (pH 7.4), using a polytron at setting number 6 for 10 sec. The homogenate was adjusted in volume to 7 ml using ice-cold 50 mM Tris-HCl (pH 7.4), and centrifuged at 35,000 × g for 15 min at 4° C. The pellet was suspended with polytron and recentrifuged in the same volume of buffer. The final pellet was resuspended in the appropriate incubation buffer at a protein concentration of 2–3 mg/ml. Aliquots were used for receptor-binding assays. Protein concentrations were determined by the BCA protein assay reagents (Pierce Chemical Co., Rockford, IL, U.S.A.) using bovine albumin as standard.

Binding assays

Radioligand binding to the following receptors or ion channel was as described previously (Shi et al., 1993): A₁-adenosine ([³H]N⁶-cyclohexyladenosine), A_2a-adenosine ([H³]CGS 21680), β₁-adrenergic ([³H]dihydroalprenolol), nicotinic ([³H] nicotine), muscarinic ([³H]-quinuclidinyl benzilate), and L-type calcium channel ([³H]nitrendipine). In all assays, including those described below, binding reactions were terminated by filtration through Whatman GF/B filters using a M24R Cell Harvester (Brandel, Gaithersburg, MD, U.S.A.). Filters were washed 2 times with 5 ml ice-cold buffer and placed in vials with 5 ml hydrofluor scintillation fluid, followed by counting for tritium. All assays were performed in duplicate.

Binding of [³H]8-cyclopentyl-1,3-dipropylxanthine ([³H]CPX) to A₁-adenosine receptors in cerebral cortical membranes was determined essentially as described by Bruns et al. (1980). [³H]CPX (0.25–1.6 nM) was incubated for 120 min at 25° C in 50 mM Tris-HCl (pH 7.4), containing 100 μl membrane suspension with adenosine deaminase (1.5 IU/ml) in a final volume of 1 ml. Competitive inhibition experiments with N⁶-cyclohexyladenosine (0.001–1 μM) were preformed with 0.2 nM [³H]CPX. Nonspecific binding was defined with 10 μM 2-chloroadenosine.

Binding of radioligands to μ-, δ- and κ-opioid receptors in cerebral cortical membranes was performed for 60–90 min at 25° C essentially as described by Goldstein and Naidu (1989) and Tiberi and Magnan (1990). [³H]DAMGO binding to μ-opiate receptors was determined using 0.5–16 nM [³H]DAMGO, 100 μl membrane suspension and 50 mM Tris buffer (pH7.4) in a final volume of 1 ml. Nonspecific binding was determined with 10 μM levallorphan. [³H]DPDPE binding to σ-opiate receptors was determined using 0.25–8 nM [³H]DPDPE, 100 μl membrane suspension and 50 mM Tris buffer (pH 7.4) in a final volume of 0.5 ml. Nonspecific binding was determined with 10μM DPDPE. [³H]U69593 binding to κ-opiate receptors was determined using 0.5–16 nM [³H]U69593, 100 μl membrane suspension and 50 mM Tris buffer (pH 7.4) in a final volume of 1 ml. Nonspecific binding was determined with 10 μM U50488.

Binding of [³H]DTG to σ-receptors in cerebral cortical membranes was determined essentially as described by Rothman et al. (1990). [³H]DTG (0.5–80 nM) was incubated for 120 min at 25° C in 5 mM HEPPSO buffer (pH 8.1), containing 100 μM EDTA, 100 μM EGTA, 25 μg/ml chymostatin, 25 μg/ml leupeptin and 100 μl membrane suspension in a final volume of 1 ml. Nonspecific binding was defined with 50 μM haloperidol. The Whatman GF/B filters for the assay had been presoaked in 0.3 % polyethylenimine.

Binding of [³H]raclopride to D₂-dopamine receptors in striatal membranes was determined essentially as described by Ferré and Fuxe (1992). [³H]Raclopride (0.5–16 nM) was incubated for 30 min at 37° C in 50 mM Tris-HCl (pH 7.4), containing 1 mM EDTA, 5 mM MgCl₂, 0.01 % ascorbic acid and 100 μl membrane suspension in a final volume of 1 ml for Scatchard analysis. Competitive inhibition experiments with dopamine (0.1 nM–10μM) were performed with 2 nM [³H]raclopride. Nonspecific binding was defined with 10 μM butaclamol.

Binding of [³H] cocaine to dopamine uptake sites in striatal membranes was determined essentially as described by Kennedy and Hanbauer (1983). Membrane preparations were prepared immediately before assay. [³H]Cocaine (20 nM) was incubated for 20 min at 21 ° C in 50 mM Tris-HCl (pH 7.4) and 100 μl membrane suspension in a final volume of 0.5 ml. Nonspecific binding was defined with 100 μM cocaine.

Linear fitting of Scatchard plots and fitting saturation isotherms to binding data were analyzed by nonlinear regression, all using computer program GraphPad InPlot (Version 4.0, San Diego, CA, U.S.A.). Both gave similar results for the determination of K_d and B_max values. Results are reported as means ± S.E.M. for 3 or 4 experiments. Statistical analysis was with the Student’s t-test using the computer program SYSTA (Version 5.0, Evanston, IL, U.S.A.).

Adenylate cyclase

The assay of activity of adenylate cyclase in striatal membranes was essentially as described by Hide et al. (1991). Assays were conducted in a total volume of 250 μl of 50 mM Tris-HCl (pH 7.4), containing 0.1 mM [α-³²P]ATP (2000 cpm/tube), 5 mM MgCl₂, 10 μM GTP, 100 μM ATP, 20 μM EGTA, 1 unit adenosine deaminase, 5 mM creatine phosphate, 0.3 mg/tube creatinine kinase and 30 μg/tube bovine serum albumin. Incubations were conducted for 10 min at 37° C and initiated by addition of striatal membranes to reaction mixtures that had been preincubated for 10 min at 37° C. Reactions were stopped by addition of 0.5 ml of trichloroacetic acid, containing 0.25 ml of 1 mM cAMP and [³H]cAMP (10,000 cpm/tube). Cyclic AMP was isolated by a two-step chromatographic procedure, using Dowex 50 and alumina columns as described by Salomon et al. (1974). The loss of [³²P]cAMP on the columns was corrected for by comparison with the losses of [³H]cAMP. Values are means ± S.E.M. for 3 experiments.

Caffeine levels in plasma

Mice were killed by decapitation. Blood of individual mice was collected from the severed neck and rapidly centrifuged at 1500 × g for 20 min. The plasma was removed, stored at − 70° C, and used within 2 weeks. The plasma was deproteinized with 1:1 volume of 0.4 M perchloric acid with 1 μg 7-β-hydroxyethyltheophylline as an internal standard. The solution was centrifuged at 20,000 × g for 15 min and the supernatants evaporated in vacuo to dryness. The residue was reconstituted in 100 μl of the mobile phase (see below) and filtered through a 0.25 μm Millipore filter. The HPLC chromatography was performed as described by Biaggoni et al. (1988), using a Hewlett Packard 1090 series II liquid chromatograph, equipped with diode array detector. A 20 μl aliquot of the final plasma extract was injected on a C₁₈ reversed-phase column (25 cm × 4.6 mm, 5 μm particle size) (Beckman, Columbia, MD, U.S.A.) with a 1.5 ml/min flow rate of the mobile phase (acetic acid:acetonitrile:water, 2:6:92 by volume). Absorbance was monitored at 280 nm. The retention times were 8.6 min and 13.8 min for β-hydroxyethyltheophylline and caffeine, respectively. The amount of caffeine in plasma was determined based on standard curves which were prepared using varying amounts of standard solutions of caf feine, mixed with an internal standard of 1 μg of β-hydroxyethyltheophylline. The ratio of peak areas of caffeine and of β-hydroxyethyltheophylline in plasma extracts was used to determine the recovery of plasma caffeine through the acid extraction procedure. The recovery was 80 ± 1 %. Plasma levels of caffeine were calculated based on the recovery factor.

Evaluation of open-field locomotor activity

The testing of mice was performed individually in a sound- and light-controlled area between 7:00 a.m. and 1:00 p.m. Chronic caffeine mice had ingested caffeine for 4 days. A 2-hr period without caffeine preceded each experiment. Drugs were dissolved in a 1:4 v/v mixture of Emulphor EL-620 (Rhône-Poulenc Chemicals Corp., Wayne, NJ, U.S.A.) and phosphate-buffered saline. The same mixture was used as a vehicle for injection in control animals. All drugs were administered i.p. in a volume of 5 ml/kg b.w. Immediately after injection, the mouse was placed in the middle of a Digiscan activity monitor (Omnitech Electronics Inc., Columbus, OH, U.S.A.), connected to a Digiscan analyzer and an IBM-compatible computer. The activity monitor cages (40 × 40 × 30.5 cm) were surrounded by horizontal and vertical sensors not detectable by mice. Data collection was begun 10 min after the mouse was placed in the middle of the open-field activity monitor. Multivariant locomotor data were collected for a 30-min period. Data for total distance traveled are presented as means ± S.E.M. The number of mice for each experimental point was 6 or greater. Statistical analysis was performed by Student’s t-test.

Results

A summary of present and prior studies on receptor and ion channel levels in control and chronic caffeine groups of male NIH Swiss strain mice is given in Table I. The time courses for onset of changes in levels of A₁-adenosine, β₁-adrenergic, nicotinic and muscarinic receptors during chronic caffeine ingestion are shown in Fig. 1. The alterations in the levels of these receptors reversed within 7 days after withdrawal of caffeine (Fig. 2). Alterations in the levels of L-type calcium channels, as assessed with [³H]nitrendipine, followed a similar time course to that for receptor alterations (Fig. 3).

Table I.

Effect of chronic caffeine ingestion on receptors and ion channels in brain membranes from male NIH Swiss strain mice

Receptor (ligand)	Kd(nM)		Bmax (fmol/mg protein)
	Control	Chronic caffeine	Contol	Chronic caffeine
A1-Adenosine
CHA, cortex	1.95 ± 0.33	2.00 ±0.45	911 ± 23	1089 ± 39**
CHA, striatum	1.70 ± 0.12	1.91 ± 0.18	668 ± 14	767 ± 28**
A2A-Adenosine
CGS 21680, striatum	10.26 ± 0.66	9.98 ± 0.19	872 ± 57	884 ± 44
α1-Adrenergic
Clonidine, cortex	0.049 ± 0.003	0.053 ± 0.012	175 ± 7	189 ± 12
α2-Adrenergic
Prazosin, cortex	1.59 ± 0.20	1.37 ± 0.20	200 ± 3	193 ± 2
β1-Adrenergic
DHA, cortex	1.15 ± 0.08	1.07 ±0.14	224 ± 9	167 ± 5*
β2-Adrenergic
DHA, cerebellum	1.57 ± 0.08	1.56 ± 0.10	158 ± 12	115 ± 11*
D1-Dopamine
SCH 23390, striatum	0.53 ± 0.11	0.48 ± 0.06	3097 ± 81	3165 ± 66
D2-Dopamine
Spiperone, striatum	0.108 ± 0.021	0.084 ± 0.018	729 ± 21	725 ± 55
5-HT1
Serotonin, cortex	2.93 ± 0.37	3.35 ± 0.58	361 ± 14	474 ± 48**
5-HT2
Ketanserin, cortex	0.33 ± 0.09	0.35 ± 0.06	275 ± 11	347 ± 11*
Nicotinic
Nicotine, cortex	0.77 ± 0.05	0.95 ± 0.05	34 ± 2	50 ± 3*
Nicotine, striatum	1.36 ± 0.17	1.59 ± 0.21	37 ± 1	39 ± 2
Muscarinic
Quinuclidinyl benzilate	0.17 ± 0.02	0.17 ± 0.03	1153 ± 56	1509 ± 47**
NMDA-Glutaminergic
MK-801	2.76+ 0.29	2.66 ± 0.17	2653 ± 97	2588 ± 46
GABAA
Diazepam	2.82 ± 0.26	4.69 ± 0.50**	1061 ± 69	1741 ± 100*
Opioid (cortex)
μ (DAMGO)	1.02 ± 0.11	1.13 ± 0.06	119 ± 6	134 ± 5
δ (DPDPE)	3.50 + 0.50	4.28 ± 0.48	83 ± 2	104 ± 7**
κ (U69593)	9.31 ± 2.67	8.96 ± 0.59	339 ± 65	230 ± 48
Sigma
DTG, cortex	32.3 ± 2.7	30.7 ± 3.1	2580 ± 90	2560 ± 170
Ca2+ channel
Nitrendipine	0.16 ± 0.03	0.17 ± 0.02	314 ± 6	369 ± 12**

Kinetic analysis of binding of radioligands to cortical cerebellar or striatal membranes from control mice and chronic caffeine mice was carried out as described in Methods.

Values for K_D and B_raax are means ± S.E.M. (n = 3 or greater).

^*p<0.01,

^**p<0.05.

Some values are from a prior publication (Shi et al., 1993).

Fig. 1. Time course for alterations in receptor levels during chronic caffeine ingestion.

A: Cortical A₁-adenosine receptors, determined with 8 nM [³H]N⁶-cyclohexyladenosine.

B: Cortical β₁-adrenoceptors, determined with 2 nM [³H]dihydroalprenolol.

C. Cortical nicotinic receptors, determined with 2 nM [³H]nicotine.

D. Cortical muscarinic receptors, determined with 0.4 nM [³H]quinuclidinylbenzilate.

*p<0.01, **p<0.02, ***p<0.05: Compared to the control group of mice.

Fig. 2. Effect of withdrawal of caffeine on alterations in receptor levels.

A: Cortical A₁-adenosine receptors.

B: Cortical β₁-adrenoceptors.

C: Cortical nicotinic receptors.

D: Cortical muscarinic receptors.

Mice ingested caffeine for 7 days followed by a 2 to 14 days withdrawal period.

Scatchard analysis (see Methods) was used to determine B_max values.

* p<0.01: Compared to the control group of mice.

Fig. 3. Time course for alterations in levels of L-type calcium channels.

A: During chronic caffeine ingestion; B: during withdrawal.

Levels in A were determined with 0.8 nM of [³H] nitrendipine. B_max values in B were determined by Scatchard analysis of binding of [³H]nitrendipine after 7 days of caffeine ingestion and after 2 or 7 days of caffeine withdrawal.

*p<0.01, **p<0.02, ***p<0.05: Compared to the control group of mice.

The levels of caffeine in plasma of mice ingesting 1 g/l, 0.5 g/l and 0.2 g/l caffeine were determined (see Methods). The values ranged from 0.7 to 5.7 μg/ml (2.1 ± 0.8, n = 6) for 1 g/l caffeine; 0–3.1 μg/ml (0.79 ± 0.49, n = 6) for 0.5 g/l caffeine and 0–0.37 μg/ml (n = 6) for 0.2 g/l caffeine. The limits of detection were estimated at <0.05 μg/ml. Based on the water consumption, the three groups were consuming about 100 mg/kg/day for 1 g/l caffeine, about 80 mg/kg/day for 0.5 g/l caffeine and about 40 mg/kg/day for 0.2 g/l caffeine. The B_max value for A₁-adenosine receptors, using [³H]N⁶-cyclohexyladenosine in these experiments, was increased by 22 % (p < 0.01) for the 1 g/l caffeine group, by 17 % (p < 0.02) for the 0.5 g/l caffeine group and by 9 % (NS, p>0.05) for the 0.2 g/l caffeine group. For comparison, the pharmacokinetics of a behavioral stimulant i.p. dose of 10 mg/kg of caffeine was determined. Peak plasma levels of about 8 μg/ml of caffeine were reached within 15 min and then declined with a half-time of 1.36 hr (data not shown). The volume of distribution was 1.18 l/g. Such acute treatment had no effect on the levels of A₁-adenosine receptors measured 4 hr later (data not shown).

The effect of caffeine ingestion on the levels of A₁-adenosine receptors was also assessed using the antagonist radioligand, [³H]CPX, rather than an agonist radioligand. For [³H]CPX, the K_D was 0.27 ± 0.02 nM in cerebral cortical membranes from control mice and 0.26 ± 0.02 nM in membranes from chronic caffeine mice. The B_max was increased from 950 ± 30 fmol/mg protein in controls to 1100 ± 30 fmol/mg protein in chronic caffeine mice (p<0.02). The apparent proportion of high (25 %) and low (75 %) affinity states for antagonism of binding of [³H]CPX by N⁶-cyclohexyladenosine was unaltered after chronic caffeine (data not shown). Similarly, the proportion of high (20 %) and low (80 %) affinity states for antagonism of binding of [³H]CGS 21680 by NECA was unaltered after chronic caffeine (data not shown).

Neither the K_D nor the B_max values for binding of the specific D₂-antagonist, [³H]raclopride, in striatal membranes were altered by chronic caffeine ingestion. The K_D and B_max values were 5.07 ± 0.53 nM and 267 ± 23 fmol/mg protein in control mice, and 4.78 ± 0.16 nM and 280 ± 27 fmol/mg protein in chronic caffeine mice. The inhibition of [³H]raclopride binding by the agonist, dopamine, was not significantly altered after chronic caffeine (Fig. 4). The Ki values for dopamine were 23 ± 2 nM (Hill coefficient: 0.77 ± 0.05) in control mice and 19 ± 2 nM (Hill coefficient: 0.87 ± 0.04) in caffeine mice. Binding of [³H] cocaine, presumably to dopamine uptake sites in striatal membranes, appeared to be slightly, but not significantly enhanced after chronic caffeine ingestion. The values at 20 nM [³H]cocaine were 13.8 ± 1.0 fmol/mg protein in controls and 18.6 ± 4.3 fmol/mg protein in chronic caffeine mice (p>0.1).

Fig. 4.

Inhibition of binding of [³H]raclopride to D₂-dopamine receptors by dopamine. Chronic caffeine mice ingested caffeine for 7 days prior to the experiment.

Basal levels of adenylate cyclase in striatal membranes decreased slightly in chronic caffeine mice, but the decrease was not statistically significant (p>0.05). The ability of dopamine, NECA and forskolin to stimulate adenylate cyclase appeared to be unaltered after chronic caffeine (Fig. 5).

Fig. 5.

Stimulation of adenylate cyclase in striatal membranes by dopamine (A), NECA (B) and forskolin (C).

Chronic caffeine mice ingested caffeine for 7 days prior to the experiment. Adenylate cyclase was assayed as described in Methods. The assay media for B included 100 nM 8-cyclopentyl-1,3-dipropylxanthine to block the A₁-adenosine receptors.

Values are means ± S.E.M. with A, B and C representing different sets of experiments (n = 3).

Amphetamine, at a submaximal dose of 1 mg/kg, has a significantly smaller stimulatory effect on locomotor activity in mice after chronic caffeine ingestion (Fig. 6; p<0.05). Antagonism of the stimulatory effect of amphetamine by NECA exhibited a multiphasic dose-response curve (depression, then stimulation and then depression) in control mice, but not in chronic caffeine mice (Fig. 6). Antagonism of the stimulatory effect of cocaine (20 mg/kg) by NECA exhibited a multiphasic dose-response curve in both control and chronic caffeine mice (Fig. 7).

Fig. 6.

Dose-dependent effects of NECA on the amphetamine-stimulated open-field locomotor activity of mice.

Amphetamine (1 mg/kg) was administered i.p., followed by i.p. NECA.

Values are means ± S.E.M. (n = 6 – 13).

The dose-dependent effects of NECA administered alone are from Nikodijević et al. (1993b).

Fig. 7.

Dose-dependent effects of NECA on the cocaine-stimulated open-field locomotor activity of mice.

Cocaine (20 mg/kg) was administered i.p., followed by i.p. NECA.

Values are means ± S.E.M. (n = 6 – 12).

Apomorphine, a directly acting D₁- and D₂-dopaminergic agonist, caused a depression of locomotor activity in both control and chronic caffeine mice (Fig. 8). Dose-response curves for the apomorphine-elicited depression of locomotor activity were similar in control and chronic caffeine mice (data not shown). Half-maximal depression of locomotor activity occurred at about 120 μg/kg of apomorphine in control mice and at about 50 μg/kg of apomorphine in chronic caffeine mice. Apomorphine (1 mg/kg) partly antagonized the cocaine-elicited stimulation of locomotor activity in both control and chronic caffeine mice (Fig. 8).

Fig. 8.

Effects of the dopamine agonist, apomorphine, on the open-field locomotor activity in the presence and absence of cocaine.

Apomorphine (120 μg/kg) was injected i.p. alone or after i.p. cocaine (20 mg/kg).

Values are means ± S.E.M. (n = 6).

Antagonism of the stimulation of locomotor activity, elicited by the muscarinic antagonist, scopolamine, by NECA, was more effective after chronic caffeine ingestion (Fig. 9). The shape of the NECA dose-response curve was similar in control and chronic caffeine mice. Oxotremorine, a muscarinic agonist, had similar depressant effects on open-field locomotor activity in control and chronic caffeine mice (Fig. 10).

Fig. 9.

Dose-dependent effects of NECA on the scopolamine-stimulated open-field locomotor activity in mice.

Scopolamine (1 mg/kg) was administered i.p., followed by i.p. NECA.

Values are means ± S.E.M. (n = 6 – 13).

Fig. 10.

Dose-dependent effects of the muscarinic agonist, oxotremorine, on the open-field locomotor activity of mice.

Oxotremorine was administered i.p.

Values are means ± S.E.M. (n = 6 – 12).

Nicotine caused a depression of open-field locomotor activity in control mice, while, at doses up to 3 mg/kg, it had no effect after chronic caffeine ingestion (Fig. 11). When co-administered with i.p. caffeine, nicotine appeared to be less effective as a depressant in control mice, while causing a biphasic (initial stimulation and then depression) effect in chronic caffeine mice (Fig. 11). Caffeine, at a dose of 35 mg/kg, caused pronounced choreiform movements in 13 % of control mice, which were increased to 70 % by co-administration of 0.1 mg/kg of nicotine (data not shown). Higher doses of nicotine (0.3, 1.0, 3.0 mg/kg) reduced the caffeine-elicited choreiform movements. Nicotine, at all doses, reduced the choreiform movements elicited by caffeine in chronic caffeine mice (data not shown).

Fig. 11.

Dose-dependent effects of nicotine on the basal and caffeine-stimulated open-field locomotor activity of control (A) and chronic caffeine mice (B).

Nicotine was administered either alone (○, ●) or after i.p. caffeine (35 mg/kg) (□, ■).

Values are means ± S.E.M. (n = 6 – 11).

The nicotinic antagonist, mecamylamine, had behavioral depressant effects in both control and chronic caffeine mice (Fig. 12). The depressant effects of mecamylamine appeared somewhat greater after chronic caffeine, but the increase was not statistically significant (p>0.05).

Fig. 12.

Effects of the nicotinic antagonist, mecamylamine, on the open-field locomotor activity of mice.

Mecamylamine was administered i.p.

Values are means ± S.E.M. (n = 6 – 12).

Discussion

Biochemical alterations

The effects of chronic treatment of mice and rats with caffeine have been extensively studied (Daly, 1993). A major focus has been on the upregulation of adenosine receptors, since blockade of such receptors represents a likely major site of action of caffeine at the dosages usually employed. Remarkably, only A₁-adenosine receptors appear upregulated with levels of striatal A_2a-adenosine receptors being unaltered, although there is one report on upregulation of A_2a-adenosine receptors after chronic caffeine in rats (Hawkins et al., 1988). β-Adrenoceptors are downregulated in rats (Goldberg et al., 1982; Fredholm et al., 1984; Green and Stiles, 1986) and mice (Shi et al., 1993). Effects of chronic caffeine on other receptor systems have not been systematically studied, except for the results summarized in Table I for male NIH Swiss strain mice. The plasma levels of caffeine in individual mice have now been determined during caffeine and were found to vary considerably, probably dependent on the amount and time interval between the last consumption of caffeine-containing liquid and sacrifice for assay. The maximal levels of plasma caffeine of about 6 μg/ml, observed in mice after administration of 1 g/l of caffeine, are somewhat lesser than the maximal levels of plasma caffeine (8 μg/ml) that occur after i.p. injection of a markedly stimulatory dose of caffeine (10 mg/kg).

Time courses for changes in A₁-adenosine, β₁-adrenergic, nicotinic and muscarinic receptors and for L-type calcium channel densities during chronic caffeine ingestion, and after withdrawal, have been determined (Figs. 1 3). Since prior studies (Shi et al., 1993 and unpublished data) had shown no alteration in affinity (K_D) but only a change in total levels (B_max) of receptors, the use of a single concentration of radioligand for time course studies provides a valid measure of relative levels. In all cases, 3–4 days are required before significant changes in levels occur, and in all cases, the changes in levels have reversed at some time between 2 and 7 days after withdrawal of caffeine. Elevated levels of A₁-adenosine receptors were reported to persist in rats for at least 15–30 days after withdrawal of caffeine (Wu and Coffin, 1984; Boulenger and Marangos, 1989), perhaps representing a species difference.

The increase in density of cortical A₁-adenosine receptors was confirmed using as radioligand an antagonist ([³H]CPX) rather than an agonist ([³H]N⁶-cyclohexyladenosine). There was no significant change in the inhibition of [³H]CPX to A₁-receptors by the agonist N⁶-cyclo-hexyl-adenosine; i.e., the proportion of high and low affinity agonist binding states did not change after chronic caffeine ingestion. The proportion of high affinity agonist state of cortical A₁-adenosine receptors was reported to be increased in rats after chronic caffeine (Ramkumar et al., 1988), perhaps again reflecting a species difference.

The lack of effect of chronic caffeine ingestion on the levels of striatal dopamine receptors (Shi et al., 1993) was unexpected, since a large body of evidence implicates striatal dopamine systems as a major target affected by caffeine (see Ferré et al., 1992). Therefore, the effects of chronic caffeine ingestion on the levels of D₂-dopamine receptors were reinvestigated with the highly selective D₂-antagonist, [³H]raclopride. There was no alteration in the levels of [³H]raclopride binding sites after chronic caffeine ingestion (see Results). However, the B_max values for [³H]raclopride of about 280 fmol/mg protein were significantly lower than those of about 730 fmol/mg protein previously found for [³H]spiperone (Shi et al., 1993, see Table I). Thus, it appears likely that [³H] spiperone binds to other sites, in addition to the D₂-dopamine receptors detected by [³H]raclopride. Inhibition of binding of the D₂-antagonist, [³H]raclopride, by the agonist, dopamine, was not significantly altered after chronic caffeine ingestion, although a slight shift in the curves suggests an increase in the affinity of dopamine (Fig. 4).

In further efforts to detect alterations in striatal A_2a-adenosine and D₂-dopamine receptor function, the stimulation of adenylate cyclase in striatal membranes by the adenosine analogue, NECA, and by dopamine was determined. There was no apparent change in stimulation elicited by NECA or dopamine or by forskolin, a direct activator of adenylate cyclase (Fig. 5).

Behavioral alterations in adenosine systems

In prior studies with NIH Swiss strain mice, the effects of chronic caf feine ingestion on behavioral responses to caffeine and adenosine analogues, both alone and in combination, have been investigated (Nikodijević et al., 1993a, 1993b). Chronic caffeine ingestion led to a significant reduction of open-field locomotor activity. The maximal open-field locomotor activity, elicited by injected caffeine, was lower after chronic caffeine ingestion (Nikodijević et al., 1993a). The biphasic dose-response curve (stimulation followed by depression) for injected caffeine was shifted leftward and the maximal per cent stimulation of locomotor activity was similar in control and chronic caffeine mice. Seven days of withdrawal were required before the basal open-field locomotor activity and the caffeine dose-response curve returned to control levels (Nikodijević et al., 1993a). This is similar to the 7 days required for biochemical alterations to reverse (Fig. 2). Choreiform movements, elicited by high doses of caffeine, were significantly reduced after chronic caffeine ingestion (Nikodijević et al., 1993c). The depressant effects of adenosine analogues were enhanced after chronic caffeine ingestion (Nikodijević et al., 1993a), providing a possible behavioral correlate to the upregulation of A₁-adenosine receptors. The dose-response curves for adenosine analogues, assessed in mice concomitant with injected caffeine, were significantly altered after chronic caffeine ingestion. In control mice, adenosine analogues had phasic dose-response curves (depression, followed by stimulation, followed by depression). After chronic caffeine, the phasic character of the adenosine analogue dose-response curves was reduced or absent (Nikodijević et al., 1993b). The synergistic depressant effects of selective A₁-and A_2a-adenosine analogues also appeared to be reduced after chronic caffeine ingestion (Nikodijević et al., 1993b). The present study focuses on behavioral interactions between adenosine and dopamine systems and between adenosine and cholinergic systems.

Behavioral alterations in dopaminergic systems

Effects of dopaminergic agents on the open-field locomotor activity were, in a prior study, investigated after chronic caffeine ingestion in mice (Nikodijević et al., 1993a). The stimulatory effect of amphetamine, a releasing agent for dopamine, on open-field locomotor activity was only slightly altered by chronic caffeine ingestion. The alteration consisted of a slight rightward shift in the amphetamine dose-response curve, so that sub-maximal doses of amphetamine caused a lesser stimulation of activity in chronic caffeine mice (Nikodijević et al., 1993a; see also Fig. 6). In the present studies, the effects of NECA on open-field locomotor activity are significantly different in amphetamine-injected control and chronic caffeine mice (Fig. 6). In control mice, NECA has clear phasic (depression, followed by stimulation, followed by depression) effects on the locomotor activity, while in chronic caffeine mice, the phasic nature of the NECA dose-response curve is nearly absent. The results are reminiscent of the NECA dose-response curves in mice injected with a stimulant dose of caffeine, where the NECA curve is also phasic in control mice, but not in chronic caffeine mice (Nikodijević et al., 1993c).

The stimulatory effect of cocaine, a potent inhibitor of dopamine uptake on the open-field locomotor activity was, in a prior study, not significantly altered after chronic caffeine ingestion (Nikodijević et al., 1993a). In the present study, the phasic dose-response curves for NECA in cocaine-injected mice were similar in control and in chronic caffeine mice (Fig 7) However the threshold for the depressant effects of NECA appeared significantly lower in control compared to chronic caffeine mice. In both control and chronic caffeine mice, NECA caused only a partial blockade of the cocaine-elicited stimulation of the locomotor activity, even at 10 μg/kg, a dose that causes a complete cessation of locomotor activity when administered alone to control or chronic caffeine mice (Nikodijević et al., 1993a).

Behavioral alterations in cholinergic systems

In a prior study, dose-response curves for the stimulatory effects of the muscarinic antagonist, scopolamine, on the open-field locomotor activity were shifted rightward after chronic caffeine ingestion (Nikodijević et al., 1993a), providing an apparent functional correlate to the upregulation of muscarinic receptors reported after chronic caffeine ingestion (Shi et al., 1993). In the present study, the NECA dose-response curves in scopolamine-injected mice were only slightly altered after chronic caffeine-ingestion. In control mice injected with scopolamine, NECA was weaker as a depressant than in chronic caffeine mice (Fig. 9). The muscarinic agonist, oxotremorine, depressed the open-field locomotor activity in control and chrome caffeine mice (Fig. 10). In a prior study, it had appeared that the depressant effects of oxotremorine (0.02 mg/kg) were reduced after chronic caffeine treatment (Nikodijević et al., 1993a). However, in the present study, such a reduction was not evident from the dose-response curves for oxotremorine (Fig. 10). Thus, for a muscarinic agonist, a behavioral correlate to the upregulation of muscarinic receptors has not been obtained.

The loss of nicotine-elicited depression of the open-field locomotor activity, seen after chronic caffeine ingestion (Nikodijević et al., 1993a) would at first appear difficult to reconcile with the apparent upregulation of nicotinic receptors after chronic caffeine ingestion (Shi et al., 1993). However, chronic nicotine treatment leads to nicotine tolerance, accompanied by a similar apparent upregulation of central nicotinic receptors (Marks et al., 1983, 1993 and references therein). It has been proposed that such “upregulated” nicotinic receptors represent desensitized nonfunctional receptors and evidence supporting this has been obtained in studies on the nicotine-evoked release of acetylcholine in striatal preparations from rats treated chronically with nicotine (Marks et al., 1993). In the present study, the loss of depressant effects of nicotine on the open-field locomotor activity after chronic caffeine ingestion was confirmed (Fig. 11A, B). In addition, after chronic caffeine ingestion, nicotine now had dose-dependent phasic (stimulation followed by depression) effects when tested in mice coinjected with a stimulant dose of caffeine (Fig. 11B). Such stimulant effects of an agent, nicotine, that usually has depressant effects, is reminiscent of effects of NECA and other adenosine analogues that can have stimulant effects when investigated in mice injected with caffeine (Snyder et al., 1981; Katims et al., 1983; Coffin et al., 1984; Nikodijević et al, 1993b). Nicotine, at low doses, could increase caffeine-elicited choreiform movements in control mice, but, at all doses, nicotine reduced such movements in chronic caffeine mice (see Nikodijević et al., 1993c, for analysis of choreiform movements in control and chronic caffeine mice).

The centrally active nicotinic antagonist, mecamylamine, had, like the agonist, nicotine, behavioral depressant effects in mice. The depressant effects of mecamylamine appeared somewhat greater in chronic caffeine mice compared to control mice (Fig. 12). Thus, 0.25 mg/kg of mecamylamine causes only a 34 % reduction in locomotor activity in control mice, while causing a 55 % reduction in chronic caffeine mice.

In summary, chronic caffeine ingestion, resulting in plasma levels of caffeine ranging up to 6 μg/ml, has marked effects on the levels of many central receptors, including cortical A₁-adenosine, muscarinic and nicotinic receptors. No significant effects on dopamine receptors, dopamine uptake systems or dopamine-stimulated adenylate cyclase are manifest. Behaviorally, there are subtile changes in behavioral responses to adenosine analogues when tested in the presence of agents, such as amphetamine or cocaine, that affect the dopaminergic function. There are marked changes in behavioral responses to nicotine either alone or in combination with caffeine. The results suggest that the behavioral responses to adenosine analogues are dependent on the functional state of dopaminergic and cholinergic systems, as revealed by changes caused by chronic caffeine ingestion.