Address correspondence and reprint requests to Dr Mairena Martín, Área de Bioquímica. Facultad de Ciencias Químicas. Avenida Camilo José Cela, 10. 13071 Ciudad Real, Spain. E-mail: Mairena.Martin@uclm.es
Pregnant rats were treated daily with 1 g/L of caffeine or theophylline in their drinking water during pregnancy and the effect of these methylxanthines on adenosine A1 receptor was assayed using binding and reverse transcription polymerase chain reaction (RT–PCR) assays in brains from both mothers and full-term fetuses. In plasma membranes from pregnant rat brain, caffeine and theophylline caused a significant decrease in total receptor numbers, of the same order in both cases (30%), with no significant changes on receptor affinity. The effect of these adenosine receptor antagonists on plasma membranes from fetal brains was more marked, being detected at approximately 50% of the total receptors detected in control conditions. However, in this tissue, a significant increase in the receptor affinity, of the same order in both cases, was also detected after antagonist administration. No significant variation on the potency of caffeine and theophylline as antagonists was detected after treatments in mothers; however, higher affinities were detected in fetuses. A decrease in the total receptor numbers in fetal brain was associated with an increase in the mRNA coding A1 receptor, as determined by RT–PCR assays, not having detected any mRNA difference in maternal brain. No variation in the levels of mRNA coding A2A receptor was detected in any case. These results suggest that maternal caffeine or theophylline intake modulates adenosine A1 receptor, causing a down-regulation of adenosine A1 receptor in brain in both mothers and fetuses.
Caffeine and theophylline are the most widely consumed psychoactive drugs in the world. Men and women of all ages, including children and pregnant women, intake daily doses of these stimulants which are present in many beverages and foods, and mainly found in coffee, tea, chocolate bars and cola drinks (Gilbert and Scott 2000). Scientific studies about the effects of on physiology and health often use 250–600 mg of caffeine as being representative of the average daily caffeine consumption. This quantity approximately represents ingesting between three and seven cups of coffee. In addition, caffeine is used for the treatment of apnoea in the premature newborn and as an additive in several analgesics and migraine remedies (Somani and Gupta 1988; Etzel and Guillet 1994). Although several effects of methylxanthines have been described, including phosphodiesterase inhibition, the caffeine- and theophylline-stimulated actions in the CNS in concentrations relevant to the daily intake of coffee are mediated by adenosine receptors blockade (Snyder et al. 1981; Fredholm 1980, 1995; Choi et al. 1988; Fredholm et al. 1999), which have been classified in four types: named A1, A2A, A2B and A3 receptors (Olah and Stiles 1995; Ralevic and Burnstock 1998). A1 and A3 receptors are coupled through a Gi/o protein to adenylyl cyclase inhibition, whereas A2A and A2B receptors are coupled to adenylyl cyclase stimulation through a Gs protein (Zhou et al. 1992; Palmer and Stiles 1997). Of these receptor subtypes, A3 and A2B have been shown to be little affected by many methylxanthines, including caffeine. Therefore, A1 and A2A receptors are likely to be the main target for caffeine and theophylline (Fredholm et al. 1999; von Lubitz 1999).
The study of caffeine and theophylline effects is especially interesting in pregnant women who usually consume these substances during pregnancy for two main reasons. First, there is neither a blood–brain nor a placental barrier to caffeine or theophylline (Ikeda et al. 1982; Kimmel et al. 1984; Tanaka et al. 1984). Therefore when pregnant rats are treated with these substances in their drinking water, fetuses also absorb the methylxanthines into the blood via the placenta. Caffeine is metabolized in the liver to form dimethyl- and monomethylxanthines. One of these metabolites is 1,3-dimethylxanthine or theophylline; therefore, in animals treated with caffeine, high levels of theophylline have also been detected (Arnaud 1987). During pregnancy, some caffeine and theophylline metabolites seem to be concentrated in fetal brain at a higher concentration than in the serum of the mothers (Wilkinson and Pollard 1993). Second, the caffeine half-life depends on age, being longer in fetuses than in full-term newborn infants as a result of lower enzymatic activities and the relative immaturity of the caffeine metabolic pathway, and decreases exponentially with post-natal age (Carrier et al. 1988; Pearlman et al. 1989; Nehlig 1999). In addition, pregnancy tends to slow down the rate at which caffeine is broken down, particularly during the last months of gestation (Brazier et al. 1983; Nehlig and Debry 1994).
There is considerable evidence about the widespread use of caffeine during and after pregnancy. Although caffeine is widely consumed during pregnancy and its consumption has been associated with low birth weight or risk of premature birth, there are not enough data for the teratological consequences in humans. However, caffeine was shown to be a teratogen in several species, including rabbits, mice and rats, in which high caffeine levels (330 mg/kg day) are necessary to induce this seizure (Narod et al. 1991). Limited exposure to caffeine in the early neonatal period may result in up-regulation of the adenosine A1 receptor, which persists to young adulthood in the rat (Guillet and Kellog 1991). Neonatal caffeine exposure of pups during the first 7 days after birth at low (0.3 g/L) or high dose (0.8 g/L) in the drinking water of their mothers shows a modulation of adenosine A1 receptor depending on the caffeine dose. Thus, A1 receptor density was not significantly affected after low doses of caffeine, but it was increased in the brain of rat pups at high doses (Bona et al. 1995). Neonatal caffeine exposure during the period in which premature human infants are administered caffeine as a treatment for apnoea altered the subsequent expression of adenosine A1 receptors in several brain regions (Etzel and Guillet 1994). Recently, it has been suggested that maternal intake of low caffeine doses during gestation and post-natal life had only minor effects on the development of adenosine A1 and A2A receptors in the rat brain (Ådén et al. 2000). However, little data are available on fetal and maternal brain after consumption of very high quantities of caffeine during pregnancy.
The aim of the present work was to determine the effect of caffeine or theophylline administration on adenosine A1 receptor in whole brain in pregnant rats during all gestation periods of both the mothers and the full-term fetuses. According to results presented herein, we show that maternal chronic caffeine or theophylline intake causes a down-regulation of adenosine A1 receptors in both maternal and fetal brain.
Materials and methods
Pregnant Wistar rats, kept on a 12-h light/12-h dark cycle (lights on at 07.00 h) and with free access to food and drinking water, were treated with caffeine or theophylline (1 g/L) in the drinking water from gestation day 2 (GD2) onwards during the entire gestation period. Control pregnant rats received drug-free tap water. At the end of this period, rats were killed and fetuses were delivered surgically. Maternal and fetal brains were then removed, frozen in liquid N2 and stored at − 70°C until experiments were performed. Daily water intake and rat weight were measured in all rats in the group. All experiments followed European Community regulations for the care and use of laboratory animals.
Plasma membranes isolation
Brain plasma membranes from mothers and full-term fetuses were isolated as described by Kessler et al. (1989), with some modifications. Brains were homogenized in 20 volumes of isolation buffer (50 mm Tris–HCl, pH 7.4 containing 10 mm MgCl2 and protease inhibitors) in Dounce (10 × A, 10 × B). After homogenization, brain preparations were centrifuged for 5 min at 1000 g in a Beckman (Coulter Espańa, Madrid, Spain) JA 21 centrifuge. Supernatant was centrifuged for 20 min at 27 000 g and the pellet was finally resuspended in isolation buffer. Protein concentration was measured by the Lowry method.
[3H]DPCPX binding assays to plasma membranes
Binding assays to plasma membranes from mothers and fetuses were performed as described by Bruns et al. (1980), with some modifications. Plasma membranes were incubated with 5 U/mg adenosine deaminase (ADA) in 50 mm Tris, 2 mm MgCl2, pH 7.4, for 30 min at 25°C, in order to eliminate endogenous adenosine from membrane preparations. Then, plasma membranes (30–75 µg of protein) were incubated with [3H]DPCPX for 2 h at 25°C. Saturation assays were carried out at different [3H]DPCPX concentrations (0.5–20 nm) using N6-cyclopentyladenosine (CPA) at a concentration 104 times that of the radioligand, in order to obtain non-specific binding. Competition curves were performed using 10 nm[3H]DPCPX and different concentrations (10 nm−10 mm) of caffeine or theophylline. Binding assays were stopped by rapid filtration through Whatman GF/B filters, pre-incubated overnight with 0.3% polyethylenimine, which were immediately washed three times with 4 mL ice-cold buffer. Filters were then transferred to vials and scintillation liquid was added to count the radioactivity.
Total RNA was isolated by guanidium thiocyanate/phenol/cloroform extraction following the method of Chomczynski and Sacchi (1987). Reverse transcription polymerase chain reaction (RT–PCR) assays of adenosine A1 receptor was performed as described by Vendite et al. (1998) using the adenosine A1 receptor primers 5′-ATCCCACTGGCCATCCTTATG-3′ and 5′-TGGCGATGTAGATCAGAATGC-3′. PCR products were analysed by electrophoresis in 2% agarose gels and stained with ethidium bromide. The PCR product size expected for adenosine A1 receptor was 630 bp. In all cases, amplification of a fragment corresponding to the β-actin sequence was carried out in parallel using the same cDNA samples in order to correct possible variations in the quantities of cDNA used for the process. The primers used for β-actin were 5′-GGTATGGAATCCTGTCGCATCCATGAAA-3′ and 5′-GTGTAAAACGCAGCTCAGTAACAGTCCG-3′. The size of the PCR product for the β-actin was 320 bp. Bands corresponding to PCR products were quantified by densitometry in a Bio-Rad GS-690 densitometer (Bio-Rad Laboratories, Hercules, CA, USA).
Protein concentration was measured by the method of Lowry, using bovine serum albumin as standard.
Statistical and data analysis
Statistical analysis was performed using the Student's t-test. Differences between mean values were considered statistically significant at p < 0.05. Saturation (Bmax, Kd) and competition (IC50) binding curves were analysed performing Scatchard and non-linear regression analysis, respectively, of binding data with the GraphPad Prism 3.02 program (GraphPad Software, San Diego, CA, USA). The IC50 values for [3H]DPCPX were converted to Ki values according to the Cheng–Prusoff equation Ki = IC50/(1 + [L]/Kd) (Cheng and Prusoff 1973) using the Kd values stated in Table 1 for the different treatments.
Table 1. Kinetic parameters of [ 3 H]DPCPX binding to plasma membranes from maternal and fetal brain
Bmax (fmol/mg prot)
Bmax (fmol/mg prot)
Data represent Bmax and Kd from maternal and fetal membranes determined by Scatchard analysis of binding data shown in Figs 1 and 2. Data are means ±SEM from at least four independent experiments performed with different plasma membranes isolations. ap < 0.005 and bp < 0.001, significantly different from control.
Daily water intake in all groups of rats (control, caffeine-treated and theophylline-treated) was not significantly different during the period of study. Caffeine or theophylline consumption was estimated from the loss of water from the drinking bottles. Caffeine and theophylline intake was 83.2 ± 5.3 and 83.8 ± 2.2 mg/kg day, respectively. Tap water intake of the control group was 84.41 ± 5.5 mL/kg day. Caffeine and theophylline did not significantly alter the weight of mothers at the end of the gestation period (data not shown). The average daily caffeine and theophylline consumption was in a dose range previously reported (Johansson et al. 1993, 1997; Svenningsson et al. 1999b). If we assume that there is approximately 80–180 mg caffeine in a cup of coffee, caffeine intake in our study could correspond to one cup of coffee. However, considering the metabolic body weight and the different half-life of methylxanthine in rats and humans, being much shorter in rats (0.7–1.2 h) than in humans (2.5–4.5 h; Morgan et al. 1982), it is generally assumed that 10 mg/kg in a rat represents about 250 mg of caffeine in a human weighing 70 kg (3.5 mg/kg), and this would correspond to about 2–3 cups of coffee (for a review see Fredholm et al. 1999). Therefore, caffeine consumption of rats in the present study would correspond to 28 mg/kg per day in humans.
Caffeine and theophylline intake effect on kinetic parameters of adenosine A1 receptors from maternal and fetal brain
In order to study the methylxanthine treatment effects on adenosine A1 receptor, we performed radioligand binding assays using [3H]DPCPX, specific A1 receptor antagonist, as radioligand. Obtained data could be best fitted to one binding site model in both treated and untreated mothers and fetuses. When we performed these assays in maternal brain, as shown in Fig. 1 and Table 1, a decrease in the total receptor number was detected in both caffeine- and theophylline-treated brains. This decrease was of the same order in caffeine-treated (30.2%) and theophylline-treated (30.6%) animals, showing that both antagonists cause the same effect at the receptor level. Although not significant, there was a 34% and 41% increase in the Kd value for caffeine and theophylline, respectively, suggesting a decrease in receptor affinity. When these assays were performed using membrane preparations isolated from fetal brain, we also detected a significant and similar decrease in total receptor numbers in caffeine-treated (49.1%) and theophylline-treated (41.0%) fetuses (see Fig. 2 and Table 1). However, the receptor decrease was associated with a significant decrease in Kd values in caffeine-treated (61%) and theophylline-treated (66%) fetuses, suggesting that treatments caused an increase in the receptor affinity in this young tissue. All these results suggest that a down-regulation mechanism of A1 receptors is operating after methylxanthine treatments.
Caffeine and theophylline intake effects on the potency order of caffeine and theophylline on adenosine A1 receptors from maternal and fetal brain
As has been previously reported, theophylline affinity for adenosine A1 receptors is higher than caffeine (for a review see Fredholm et al. 1999). On the other hand, theophylline and caffeine metabolism is quite slow in fetal tissue because of the low enzymatic activities compared with adult tissue (Nehlig 1999). Therefore, high levels of caffeine and theophylline can be detected in this immature tissue for longer periods of time. In order to study whether caffeine or theophylline treatments affect its methylxanthine affinities for adenosine A1 receptors, we performed competition curves in membranes from control and treated mothers and fetuses using different concentrations of these methylxanthines. Theophylline IC50 values were lower than those of caffeine in both mothers (Fig. 3) and fetuses (Fig. 4), suggesting that theophylline exhibits higher affinity than caffeine in both tissues assayed, as expected (Fredholm and Lindstrom 1999; Svenningsson et al. 1999b). On the other hand, Ki values (Table 2) obtained from IC50 data (Figs 3 and 4) using the Cheng–Prusoff equation revealed that methylxanthine treatment increased caffeine and theophylline affinities for A1 receptor in fetal brain.
Table 2. Caffeine and theophylline Ki values from maternal and fetal brain plasma membranes
Ki (µ m )
Data represent caffeine and theophylline Ki values calculated with the Cheng–Prusoff equation from IC50 values determined by non-linear regression analysis in both maternal (Fig. 3) and fetal (Fig. 4) plasma membranes. Data are means ±SEM from at least three experiments performed with different plasma membranes isolations. ap < 0.05, bp < 0.001, significantly different from control. cp < 0.005 and dp < 0.001, significantly different from corresponding caffeine Ki value.
Caffeine and theophylline intake effect on mRNA coding adenosine A1 receptors from maternal and fetal brain
Several reports show that adaptive changes to the effect of methylxanthines do not include molecular changes related to the adenosine A1 receptor (Johansson et al. 1993, 1997; Bona et al. 1995). However, some others show an increase in A1 receptor mRNA in several brain regions after long-term caffeine treatment (Svenningsson et al. 1999b). In order to verify whether caffeine or theophylline intake causes modulation on A1 receptor mRNA, we performed RT–PCR assays and we detected a slight but significant increase in mRNA from caffeine-treated (116.5 ± 2.9% of control) and theophylline-treated (116.8 ± 6.9% of control) fetal brain, suggesting a modulation at the molecular level (Fig. 5b), again of the same order in both treatments. However, no significant differences in maternal mRNA coding A1 receptor were observed after treatments (Fig. 5a).
Of the adenosine receptor subtypes, A3 and A2B have been shown to be little affected by many methylxanthines, including caffeine. Therefore, A1 and A2A receptors are likely to be the main target for caffeine and theophylline (Fredholm et al. 1999; von Lubitz 1999). Thus, we performed RT–PCR assays to determine whether treatments modulated the A2A receptor mRNA in both maternal and fetal brain. When we performed these assays, data did not reveal significant variation on A2A mRNA in any case, as it can be observed in Fig. 6.
It is usually considered that prolonged agonist exposure of G protein-coupled receptors results in a progressive loss of receptor (down-regulation), whereas antagonist exposure causes increased receptor level (up-regulation; Böhm et al. 1997). Most studies of the effects of chronic antagonist treatment report an increase in total receptor number as an adaptive response to high antagonist exposure of receptor. In the case of adenosine A1 receptors, up-regulation after chronic caffeine (Fredholm 1982; Guillet and Kellog 1991; Bona et al. 1995; Hettinger-Smith et al. 1996; Johansson et al. 1997) and theophylline (Murray 1982; Sanders and Murray 1988; Fastbom and Fredholm 1990; Lupica et al. 1991) treatment has been reported. However, several authors have also postulated that caffeine and theophylline do not cause any effect on total receptor levels; therefore, the role of adenosine A1 receptor up-regulation in mediating adaptive changes to long-term xanthines treatment remains controversial (Holtzman et al. 1991; Kaplan et al. 1993; Svenningsson et al. 1999b). The main finding in the present paper is that caffeine as well as theophylline oral intake during pregnancy causes a significant decrease in total A1 receptors, not only in fetal but also in maternal brain. This decrease was associated with a significant increase in A1 mRNA level in fetal tissue.
There are a few examples of adenosine A1 receptor antagonist-induced desensitization. One is reported by Navarro et al. (1999) in GH4 cells, where not only desensitization but also an internalization process takes place after DPCPX treatment. Moreover, caffeine infusion significantly decreased [3H]DPCPX-specific binding to adenosine A1 receptor in vivo (Kaplan et al. 1993), and [3H]CHA binding was also decreased in the molecular layer of cerebellum from neonatal rats exposed to caffeine within the first post-natal week (Etzel and Guillet 1994). These results agree with data reported in the present work in which adenosine A1 receptors were significantly decreased in fetuses from mothers treated with caffeine (49.1%) or theophylline (41.0%). In contrast, several authors have reported that caffeine administrated in the first week of life causes an increase in the A1 receptor density in rat brain (Bona et al. 1995) or in cortex, hippocampus and cerebellum, where the increase persists to young adulthood (Guillet and Kellog 1991). However, these studies were performed with neonatally caffeine-exposed animals and not after fetal exposure. In this way, specific adenosine A1 binding in fetal rat brain after maternal caffeine intake have been analysed, but no significant differences were reported in caffeine-treated versus control fetuses on GD 21 and 2 h post-natally (Ådén et al. 2000).
The fetal adenosine A1 receptor Bmax value (56.8 ± 4.2 fmol/mg protein in control animals) was significantly lower than the maternal value (139.9 ± 5.9 fmol/mg protein), which means that only 40% of A1 receptors detected in mothers are present in fetuses, corresponding to immature tissue. These data agree with those reported by Rivkees (1995) in which there was a two-fold increase in A1 receptor from birth to adulthood, and by Ådén et al. (2001) who reported that adenosine A1 receptor levels in 7-day-old rats were 23% of the adult levels.
In agreement with previously reported data (Fredholm 1982; Fredholm and Lindstrom 1999; Svenningsson et al. 1999b), the potency of caffeine or theophylline as adenosine A1 receptor antagonists was not altered in maternal brain by either caffeine or theophylline treatment, as Ki values revealed. However, Ki for caffeine was slightly lower (p < 0.05) in caffeine-treated than control brain membranes, suggesting an increase in affinity, whereas slightly higher Kd measured in these membranes suggests a decrease. Therefore, we think that this mathematical difference does not represent a biological difference in affinity for caffeine after caffeine treatment in mothers. On the other hand, lower Ki values (p < 0.001) detected in fetal brain after treatments suggest an increase of caffeine and theophylline affinity, in agreement with lower Kd values obtained from saturation binding assays. However, this difference was not statistically significant in theophylline Ki values after theophylline treatment, probably because of the high SEM obtained. This could also be the reason why there was no statistically significant difference between caffeine and theophylline Ki values in theophylline-treated membranes, although theophylline Ki values were always lower in all cases assayed, suggesting that theophylline was more potent than caffeine as an antagonist of A1 receptors, as previously reported (Fredholm and Lindstrom 1999; Svenningsson et al. 1999b).
Results presented herein have shown that a significant increase of mRNA coding A1 receptor was associated with the decrease in total receptor number in fetal brain after caffeine (116.5 ± 2.9% of control) or theophylline (116.9 ± 6.9% of control) treatment; however, no differences were observed in maternal tissue. Most studies of the effects of caffeine or theophylline on adenosine A1 receptor do not reveal any significant change on the mRNA level associated with an increase of A1 receptor (Johansson et al. 1993, 1997), including studies on the neonate brain (Bona et al. 1995). However, it has been reported recently that long-term oral caffeine treatment increases A1 receptor mRNA in the lateral amygdala, and also tends to increase this mRNA in hippocampal areas of adult rats (Svenningsson et al. 1999b). The different results obtained from maternal and fetal mRNA could be caused by a compensation mechanism to the important loss of receptor in developing fetal brain (approx. 50% higher than in mothers), revealing the hypersensitivity of this immature tissue to the action of methylxanthine. On the other hand, as has been suggested by Weaver (1996), the quantitative relationship between receptor binding and A1 mRNA levels is less certain. In humans (Ren and Stiles 1994), as in guinea pigs (Meng et al. 1994), there are splice variants of the A1 receptor with differing translational activities, which could also be the case in the rat.
One question is: why do high levels of caffeine or theophylline consumption during pregnancy cause a down-regulation of adenosine A1 receptors in both maternal and fetal brain? A possible explanation could be the enhanced endogenous adenosine release after treatments. Recently, it has been reported that caffeine increases plasma adenosine concentration by an unknown receptor-mediated mechanism (Conlay et al. 1997). Adenosine concentration increase was dependent on the dose and administration method, which was higher after chronic caffeine administration. When caffeinated solution was withdrawn and replaced with tap water, the plasma adenosine concentration declined to the control level. In Conlay's report, average caffeine (0.1% in drinking water) consumption was similar to that consumed by our pregnant rats. Accordingly, adenosine A1 receptor antagonists increase extracellular adenosine levels in cardiovascular cells from rats and humans (Andresen et al. 1999), suggesting that adenosine A1 receptor tightly modulates extracellular adenosine concentration. Moreover, in pregnant women and mainly in the third trimester, there is an important increase in plasma adenosine level possibly attributed to the enhanced adenosine release from activated platelets (Yoneyama et al. 2000a,b). Thus, it is possible that an adenosine A1 receptor down-regulation after caffeine or theophylline chronic intake is a consequence of increased adenosine release.
Of all adenosine receptor subtypes, A1 has the highest affinity for adenosine (Ralevic and Burnstock 1998; Fredholm et al. 1999; von Lubitz 1999). This affinity is high enough to permit A1 receptor activation at physiological concentrations of extracellular adenosine. However, adenosine A2A receptors could be activated for higher adenosine levels, released after caffeine or theophylline treatments, as we have postulated. In this regard, different effects on A2A receptors have also been reported after chronic caffeine treatment (Johansson et al. 1997; Svenningsson et al. 1999b). Nevertheless, A1 receptors are present in many brain regions, whereas A2A receptors are mainly found in the striatum (Ribeiro 1999; Svenningsson et al. 1999a). Therefore, we have not studied the effect of caffeine or theophylline treatment on this receptor type, although we have analysed A2A mRNA levels in caffeine- and theophylline-treated maternal and fetal brain and we did not detect a significant variation. Obviously, this absence of mRNA variation cannot exclude the possible modulation on adenosine A2A receptors. Additional experiments will be necessary in order to clarify this point.
In conclusion, we have shown that chronic caffeine or theophylline exposure during pregnancy promoted a decrease in adenosine A1 receptors in both maternal and fetal whole brain. This could reflect an increase in the stimulatory activities exhibited by caffeine and theophylline and, in consequence, an increase in the vulnerability of brain and probably other tissues to the harmful effects of both substances, which would be especially important in developing fetuses. Therefore, particular attention should be paid to caffeine consumption during pregnancy, mainly in pregnant women that consume high quantities of caffeine.
This work has been supported by grants FIU99–00 and FIU00–01 from UCLM. DL is a recipient of a predoctoral fellowship from JCCM.