You are here

Selective Activation of Adenosine A2A Receptors on Immune Cells by a CD73-Dependent Prodrug Suppresses Joint Inflammation in Experimental Rheumatoid Arthritis.

Editors' Pick
Selective Activation of Adenosine A2A Receptors on Immune Cells by a CD73-Dependent Prodrug Suppresses Joint Inflammation in Experimental Rheumatoid Arthritis.
Flögel U, Burghoff S, van Lent PLEM, Temme S, Galbarz L, Ding Z, El-Tayeb A, Huels S, Bönner F, Borg N, Jacoby C, Müller CE, van den Berg WB, Schrader J
Sci Transl Med. 2012 Aug 8; 4(146):146ra108.
PMID: 22875828.


Aug 20 2012

Mark Zylka, University of North Carolina at Chapel Hill

Nucleotides like ATP and ADP can sensitize nociceptors and cause pain; however, little was known about the signaling potential of AMP—until now.  Several recent studies highlight the therapeutic potential of AMP analogs in the setting of inflammation and pain. 

Earlier this year, we found that AMP can directly activate the adenosine A1 receptor (Rittiner et al., 2012).  Activation did not require hydrolysis of AMP to adenosine by ectonucleotidases.  Moreover, we synthesized several AMP analogs that activated the adenosine A1 receptor (Korboukh et al., 2012).  One of these AMP analogs was orally active and had A1-dependent antinociceptive effects in mice with minimal to no cardiovascular side effects.  This analog thus bypassed a common side effect associated with adenosine receptor agonists.  Our research unexpectedly suggested that some of the physiological effects of AMP (and AMP analogs) might be due to direct activation of adenosine receptors and independent of ectonucleotidases.

Intriguingly, in this recent study by Flögel and colleagues, a different AMP analog (chet-AMP, first described in El-Tayeb et al., 2010 as 12g) had anti-inflammatory effects in a collagen-induced arthritis model and lacked cardiovascular side effects.  This work, along with our studies, raises the exciting prospect that AMP analogs might be effective for treating pain and inflammation.

Flögel and colleagues reported that the anti-inflammatory effects of chet-AMP were due to hydrolysis to chet-adenosine by ecto-5’-nucleotidase (NT5E, also known as CD73) followed by activation of the adenosine A2A receptor. 

While the anti-inflammatory effects of chet-AMP were significant, the mechanism underlying these anti-inflammatory effects was not rigorously pinned down in this study by Flögel and colleagues.  This includes uncertainty (at least from my perspective) as to whether chet-AMP acts via A2A receptors, as the title of their paper indicates, and uncertainty as to whether chet-AMP is a prodrug.  Also unclear is whether or not chet-AMP reduces the levels of inflammatory cytokines (data shown in Fig. 4c was not statistically significant and showed large measurement variability).

The authors infer that chet-AMP acts as a prodrug, that is, a drug that is only active after it is dephosphorylated to chet-adenosine by NT5E.  More studies are essential to confirm this mechanism in vivo.  The authors used a pharmacological inhibitor of NT5E in their study.  However, given issues related to inhibitor specificity in vivo (where multiple ectonucleotidases are present) and drug distribution, it is critical that the authors confirm a requirement for NT5E using Nt5e-/-mice. 

In fact, in their previous study (El-Tayeb et al., 2010; Fig. 3), chet-AMP (12g) hydrolysis was reduced only about 50% when incubated with T-cells from Nt5e-/- mice.  Clearly NT5E is not the only enzyme that dephosphorylates chet-AMP, at least in T-cells.

Given our finding that AMP and AMP analogs can signal directly through adenosine A1 receptors, this raises the question of whether chet-AMP is also an A1 agonist.  And related, are the anti-inflammatory effects seen in vivo A1 and/or A2A mediated? 

Indeed, in their previous study (El-Tayeb et al., 2010), chet-adenosine was equally effective at displacing radioligands at the A1 receptor and the A2A receptor (400 ± 3 nM vs 372 ±57 nM, respectively).  They did not examine binding of chet-AMP to either receptor, so it will be critical to determine if chet-AMP is an agonist of the A1 receptor and/or the A2A receptor.   

To ascertain receptor mechanism, Flögel and colleagues used only one A2A antagonist (KW-6002) at a single concentration in vivo (delivered continuously for 1-2 weeks using an osmotic minipump).  Whether this antagonist remains selective for A2A in vivo at this dose and route and for this extended duration is unclear.  Use of A1 and A2A knockout mice combined with use of additional antagonists (A1 and A2A selective) would provide stronger support that the anti-inflammatory effects of chet-AMP are due to A2A receptor activation.

From a therapeutic perspective, chet-AMP faces challenges if delivered via the oral route.  Chet-AMP has a free phosphate group, giving it a net negative charge at physiological pH.  Moreover, the free phosphate will likely make chet-AMP a substrate for alkaline phosphatases, as the authors acknowledged.  If delivered orally, chet-AMP will most likely be dephosphorylated by intestinal alkaline phosphatase before it reaches the blood stream.  Pharmacokinetic studies can address this possibility.  It may prove advantageous to protect the phosphate group, such as with methy-esters (Korboukh et al., 2012), to broaden the therapeutic potential of chet-AMP.

To conclude, chet-AMP and other AMP analogs show efficacy in preclinical models of pain and inflammation.  However, future studies will be needed to rigorously pin down mechanistically how chet-AMP and other AMP analogs act in vivo.  When planning these future studies, it should also be kept in mind that some AMP analogs can signal through P2Y receptors (Boyer et al., 1996).  Based on our experience at developing nucleotide “prodrugs” and the work of others, it is clear that NT5E (CD73) is not the only enzyme that hydrolyzes AMP in vivo and in physiological preparations (Street et al., 2011; Picher et al., 2003; Zhang et al., 2012).  These ectonucleotidases must be rigorously inhibited (genetically, when possible) before one can ascertain whether any nucleotide analog truly is a prodrug in vivo.