GIRK channels in atrial myocytes (AM) can be activated by acetylcholine (ACh) and adenosine (Ado) via M₂ and A₁ receptors (A₁-R, M₂-R), respectively. In rat AM the amplitude of current that can be activated by a saturating concentration (10 mM) of Ado (I_K(Ado)) amounts to less than 40% of I_K(ACh). Adenovirus driven overexpression of A₁-R results in an increase in I_K(Ado), which is paralleled by a strong reduction in I_K(ACh), reflecting a loss in functional M₂-R (Wellner-Kienitz et al., Circ. Res. 31, 2000).

Knockdown of M₂-R by siRNA caused a reduction in I_K(ACh) by ~ 80%, but did not result in an increase in I_K(Ado), arguing against a competition between M₂-R and A₁-R for GIRK-containing signaling complexes. Reduction of I_K(ACh) in A₁-R overexpressing AM was prevented by incubation (>24 h) with the A₁-R antagonist DPCPX, whereas incubation with A₁-R agonists causes a reduction of both currents I_K(ACh) and I_K(Ado). These data suggest that downregulation of M₂-R is caused by activation of A₁-R by extracellular adenosine generated by atrial myocytes. Since the myocyte cultures are devoid of non-myocyte cells, other sources can be excluded.

In myocytes with normal expression levels of both receptors, density of functional M₂-R was reduced by incubation with AMP but not with ATP or ADP. The action of AMP was abolished by inhibiton of ecto-5-nucleotidase (CD73) using a,ß-methylene-adenosine 5'diphosphate. Moreover, inhibition of ecto-adenosine desaminase by EHNA caused significant reduction in densities of I_K(ACh) and I_K(Ado) without addition of AMP to the culture medium. These data demonstrate that (i) chronic activation of A₁-R in cultured atrial myocytes causes downregulation of both, A₁-R and M₂-R and (ii) adenosine is produced by atrial myocytes in vitro, which activates A₁-R in an autocrine fashion.