Our findings show that incubation of microglia or astrocytes with exogenous 2′,3′-cAMP induces an increase in the levels of 3′-AMP, 2′-AMP and adenosine in the medium, whereas incubation of microglia or astrocytes with 3′,5′-cAMP has little affect on extracellular levels of 5′-AMP or adenosine. We conclude that both microglia and astrocytes express the extracellular 2′,3′-cAMP-adenosine pathway, but not the extracellular 3′,5′-cAMP-adenosine pathway.
Previous studies demonstrate that the extracellular 3′,5′-cAMP-adenosine pathway is mediated by an ecto-phosphodiesterase that is sensitive to inhibition by both IBMX and DPSPX (Jackson & Raghvendra 2004
). The current study shows that IBMX has little effect on the conversion of 2′,3′-cAMP to 3′-AMP or 2′-AMP by either microglia or astrocytes. Thus the 2′,3′-cAMP-phosphodiesterase that mediates the 2′,3′-cAMP-adenosine pathway is distinct from the 3′,5′-cAMP-phosphodiesterase that mediates the 3′,5′-cAMP-adenosine pathway. 2′,3′-cAMP is hydrophilic and would not be expected to penetrate cell membranes, so enzymes that metabolize 2′,3′-cAMP must be ecto-enyzmes, i.e., ecto-2′,3′-cAMP-phosphodiesterases.
In both microglia and astrocytes, DPSPX reduces the conversion of 2′,3′-cAMP to 2′-AMP, yet has no effect on the metabolism of 2′,3′-cAMP to 3′-AMP. This suggests that there are two separate ecto-2′,3′-cAMP-phosphodiesterases in microglia and astrocytes. One metabolizes 2′,3′-cAMP to 3′-AMP (i.e., ecto-2′,3′-cAMP-2′-phosphodiesterase) and is resistant to inhibition by both IBMX and DPSPX. A second metabolizes 2′,3′-cAMP to 2′-AMP (i.e., ecto-2′,3′-cAMP-3′-phosphodiesterase) and is resistant to inhibition by IBMX, but is modestly inhibited by DPSPX. These findings are consistent with observations in preglomerular vascular smooth muscle and glomerular mesangial cells where degradation of 2′,3′-cAMP to 2′-AMP is partially inhibited by DPSPX, but that to 3′-AMP is not (Jackson et al. 2010
That different enzymes mediate the metabolism of 2′,3′-cAMP to 2′-AMP versus 3′-AMP has implications. If 3′-AMP and 2′-AMP have different biological effects apart from their conversion to adenosine, the relative expression of ecto-2′,3′-phosphodiesterases and their relative kinetic properties would determine whether extracellular 2′,3′-cAMP is converted predominantly to 3′-AMP or to 2′-AMP. Indeed, the current study demonstrates that both microglia and astrocytes convert 2′,3′-cAMP preferentially to 2′-AMP, rather than 3′-AMP.
Incubating primary microglia or astrocytes with exogenous 3′-AMP or 2′-AMP (hydrophilic compounds) induces an increase in extracellular adenosine. This suggests that these AMPs can be metabolized to adenosine in the extracellular biophase and supports the existence of the extracellular 2′,3′-cAMP-adenosine pathway (2′,3′-cAMP → 2′-AMP/3′-AMP → adenosine) in these cell types. This conclusion is confirmed by the increase in extracellular adenosine levels following incubation of microglia or astrocytes with 2′,3′-cAMP.
Astrocytes convert 2′,3′-cAMP to 2′-AMP and 3′-AMP more efficiently than microglia, yet microglia convert 2′-AMP and 3′-AMP to adenosine more efficiently than astrocytes. This suggests that astrocytes collaborate with microglia to provide a 2′,3′-cAMP-adenosine pathway that affects microglia function via adenosine formation. In this regard, astrocytes may function to metabolize 2′,3′-cAMP to 2′-AMP and 3′-AMP which can then diffuse to nearby microglia and be converted to adenosine in the biophase of adenosine receptors on microglia.
The present study shows that in both microglia and astrocytes, AMPCP inhibits the conversion of extracellular 5′-AMP to extracellular adenosine, but has little effect on the metabolism of extracellular 3′-AMP or 2′-AMP to extracellular adenosine. This suggests that the ecto-2′/3′-nucleotidase that mediates the conversion of 3′-AMP or 2′-AMP to adenosine is distinct from the ecto-5′-nucleotidase (CD73) that mediates the conversion of extracellular 5′-AMP to adenosine. Conceivably there exists distinct ecto-2′-nucleotidases and ecto-3′-nucleotidases and ecto-nucleotidases that recognize both 2′-AMP and 3′-AMP.
Cultures of nearly pure primary microglia or astrocytes do not express the 3′,5′-cAMP-adenosine pathway. However, there is ample evidence for the extracellular 3′,5′-cAMP-adenosine pathway in the brain. For example, in rat cerebral cortex in dissociated cell culture, 3′,5′-cAMP is converted to 5′-AMP and adenosine (Rosenberg & Dichter 1989
). Also, in cerebral cortical cultures, isoproterenol (Rosenberg et al. 1994
), norepinephrine (Rosenberg & Li 1995
), epinephrine (Rosenberg & Li 1995
) and forskolin (Rosenberg & Li 1996
) increase extracellular 3′,5′-cAMP and adenosine. The extracellular 3′,5′-cAMP pathway also exists in the hippocampus (Brundege et al. 1997
). In superfused hippocampal slices, both forskolin and exogenous 3′,5′-cAMP increase extracellular adenosine levels and induce adenosine-mediated electrophysiological effects (Brundege et al. 1997
). Moreover, studies indicate that the 3′,5′-cAMP-adenosine pathway limits D1 receptor expression in CAD catecholaminergic cells (Do et al. 2007
). Therefore, it is likely that the 3′,5′-cAMP-adenosine pathway exists in the brain, either in neurons or as a result of interactions among several cell types. Because the present study demonstrates robust metabolism of 5′-AMP to adenosine by microglia, if neurons metabolize 3′,5′-cAMP to 5′-AMP, microglia would provide ecto-5-nucleotidase activity to metabolize 5′-AMP to adenosine. Consequently, microglia may be involved in a functioning 3′,5′-cAMP-adenosine pathway when in collaboration with other CNS cell types.
Adenosine activates four subtypes of cell-surface adenosine receptors (A1
). Our recent study shows that A1
receptor stimulation reduces microglial activation after experimental traumatic brain injury (Haselkorn et al. 2009
). Seven days after controlled cortical impact injury, Iba-1 staining identifies marked microglial proliferation that is significantly enhanced in nearly all brain regions in A1
receptor knockout mice. Two prior studies using A1
-receptor-null mice in other unrelated models of CNS inflammation report similar magnitudes of increases in microglial proliferation in knockout mice versus wild-type (Tsutsui et al. 2004
, Synowitz et al. 2006
). Moreover, our previous work demonstrates that in BV-2 cells (immortalized mouse microglia cell line) in culture, cell proliferation, as assessed by 3
H-thymidine incorporation, is inhibited by either A1
receptor or A2B
receptor activation (Haselkorn et al. 2009
). Thus injury-induced formation of adenosine resulting from activation of the extracellular 2′,3′-cAMP-adenosine pathway in microglia may be critical with regard to limiting the aggressiveness of the microglial response to injury. This would be analogous to the now well established role of adenosine as an anti-inflammatory autocoid that helps resolve inflammation in peripheral tissues (Rajakariar et al. 2009
). Finally, a paracrine role for microglia-derived adenosine may protect neurons from glutamate-induced excitotoxicity (Lauro et al. 2008
Recent studies show that 2′,3′-cAMP enhances opening of the mitochondrial permeability transition pore (Azarashvili et al. 2009
), a process that is linked to apoptosis and necrosis. Thus the cellular transport of 2′,3′-cAMP by injured neurons into the brain interstitial space following an insult may be important in removing this potentially lethal intracellular compound. Rapid metabolism of extracellular 2′,3′-cAMP to 2′-AMP and 3′-AMP by astrocytes and microglia may facilitate the disposal of neuronal intracellular 2′,3′-cAMP in the area of injury by maintaining a low extracellular concentration of 2′,3′-cAMP and thereby providing a more favorable intracellular-to-extracellular gradient for efflux of intracellular 2′,3′-cAMP. In summary, the characterization of extracellular 2′,3′-cAMP metabolism presented here reveals a novel function of microglia and astrocytes which may be important in their response to brain injury.