The potencies of caffeine and related methylxanthines as adenosine antagonists were assessed with respect to three apparent subtypes of adenosine receptors in rat brain preparations: (i) the A₁-adenosine receptor which binds with a very high affinity the ligand [³H]cyclohexyladenosine (K _D, 1 nM) in rat brain membranes; (ii) a ubiquitous low-affinity A₂-adenosine receptor which activates cyclic AMP accumulation in rat brain slices—this A₂-adenosine system exhibits an EC₅₀ for 2-chloroadenosine of about 20µM; and (iii) a relatively high-affinity A₂-adenosine receptor which activates adenylate cyclase in rat striatal membranes—this A₂-adenosine system exhibits an EC₅₀ for 2-chloroadenosine of about 0.5µM and is present in striatal but not in cerebral cortical membranes.

The rank order of potency for methylxanthines versus binding of 1 nM [³H]cyclohexyladenosine in membranes from eight rat brain regions is theophylline (IC₅₀, 20–30µM) > paraxanthine (IC₅₀, 40–65µM) > caffeine (IC₅₀, 90–110µM) > theobromine (IC₅₀, 210–280µM). There thus appears to be little difference in A₁-receptors in different brain regions in terms of interaction with these methylxanthines. 1-Methylxanthine is more potent than caffeine in rat cerebral cortical membranes, while 3-methylxanthine and 7-methylxanthine are less potent than caffeine.

The rank order of potency for methylxanthines versus activation of cyclic AMP accumulation by 50µM 2-chloroadenosine in rat striatal slices is theophylline (IC₅₀, 60µM) > paraxanthine (IC₅₀, 90µM) > caffeine (IC₅₀, 120µM) » theobromine (IC₅₀, > 1000µM). Similar potencies pertain in cerebral cortical slices.

The rank order of potency of methylxanthines versus activation of adenylate cyclase by 1µM 2-chloroadenosine in rat striatal membranes is theophylline (IC₅₀, 20µM) > paraxanthine (IC₅₀, 40µM) > caffeine (IC₅₀, 80µM) » theobromine (IC₅₀, > 1000µM).

Caffeine and other methylxanthines, thus, antagonize effectively both A₁- and A₂-adenosine receptors in brain perparations. Theobromine appears less effective versus A₂-receptors than versus A₁-receptors. Caffeine exhibits aK _i value of about 50µM at the very high-affinity A₁-binding sites, aK _i value of about 30µM at the low-affinity A₂-adenosine site in brain slices, and aK _i value of about 27µM at the high-affinity A₂-adenosine site in striatal membranes. The functional significance of antagonism of such adenosine receptors by caffeinein situ will depend both on the local levels of adenosine and on the affinity for adenosine for the receptor, since antagonism by xanthines is competitive in nature. In addition, the functional significance of xanthine action will depend on the degree of inhibition of adenosine input which is required to alter the output signal. For a stimulatory input to adenylate cyclase via an A₂-adenosine receptor, profound antagonism by methylxanthines is probably required to alter the cyclic AMP-mediated output signal, while for inhibitory input to adenylate cyclase via an A₁-adenosine receptor, presumably a lesser degree of antagonism by methylxanthines may be required to alter the cyclic AMP-mediated output signal.