Adenosine Axis — CD39, CD73 and A2A Receptors

One molecule, two opposite meanings, separated by two enzymes. Extracellular ATP is a danger signal. Released by necrosis, by hypoxic stress and through pannexin‑1 and connexin‑43 channels, it recruits and activates myeloid cells through P2X7, and it is one of the signals that primes the inflammasome. Adenosine — ATP minus three phosphates — does the reverse of all of that. The entire immunological meaning of the purine pool therefore rests on the hydrolysis rate, and that rate is set by two surface enzymes: CD39 (ENTPD1), which takes ATP and ADP to AMP, and CD73 (NT5E), which takes AMP to adenosine. ENPP1 provides a parallel route from ATP to AMP, and the CD38–CD157 pair generates AMP from NAD⁺ instead — which is why CD38 blockade has an adenosinergic rationale on top of its depleting one. Because CD39 and CD73 are co-expressed on Tregs, on tumour cells and on tumour endothelium, the tissue that most needs an inflammatory response is the tissue most efficient at abolishing it.

Four receptors, two directions. Adenosine has four G-protein-coupled receptors and they do not agree. A2AR (ADORA2A) and A2BR (ADORA2B) couple to G_s, activate adenylate cyclase, and raise cAMP. A1R and A3R couple to G_i and lower it. A2AR is the high-affinity receptor and the dominant one on T cells and NK cells; A2BR is low-affinity, so it only engages at the adenosine concentrations found in tumours and ischaemic tissue, and it is the receptor most associated with myeloid suppression and VEGF release. Which receptor a cell reads therefore depends on how much adenosine is present — a genuine concentration-dependent switch, not a redundancy. Any experiment that measures "adenosine signalling" without specifying the receptor is measuring an average of opposing effects.

cAMP is the suppression, and PKA is how it lands. Raised cAMP activates PKA, which phosphorylates CREB and, critically, activates Csk. Csk phosphorylates the inhibitory tyrosine of Lck and thereby blunts proximal TCR signalling — the mechanistic reason adenosine reduces IL-2, IFN-γ, perforin and granzyme B output and downregulates NKG2D. Epac1 carries the cAMP-dependent effects that are independent of PKA. On the other side of the same signal, CREB-driven transcription favours FoxP3 and CD25, so the suppression is not merely a brake on effectors but an active push towards regulatory phenotypes, and towards PD-1 and CTLA-4 expression — which is the rationale for combining adenosine-axis blockade with checkpoint blockade rather than substituting one for the other.

The loop is the point. HIF-1α and HIF-2α transcriptionally upregulate CD39 and CD73 under hypoxia. The adenosine they generate promotes Tregs, myeloid Arg1 and IL-10, and A2BR-driven VEGF, which produces the disorganised, leaky vasculature that sustains the hypoxia. TGF-β1 closes a second loop by inducing CD39 and CD73 directly. This is why single-point intervention disappoints and why the axis is usually attacked at two levels at once. The In Vivo tie-in: functional-grade anti-CD73 (clone TY/23) is the tool for testing whether adenosine generation is rate-limiting in your model, and pairing it with anti-CD25, anti-CD4/CD8 depletion and anti-PD-1 or anti-CTLA-4 lets you separate "less adenosine" from "fewer Tregs" from "better effectors" — three explanations that look identical in a tumour-growth curve. For research use only; not for use in diagnostic or therapeutic procedures.