Hydrolysis of extracellular ATP by CD39⁺ Treg cells: context matters!

Response:

We thank our colleagues for their comments as they stress the complex role extracellular adenosine triphosphate (ATP) actually plays in immune regulation. This applies even more so as there are indeed apparent differences between mice and humans that may result in species-specific signaling pathways. These differences comprise not only the differential expression of P2Y11 receptors but also the cellular distribution of CD39. In the mouse, the ectoenzyme is expressed constitutively on all Foxp3⁺ T regulatory (Treg) cells, whereas in humans the expression is restricted only to a specific Treg subset.¹  While these differences may complicate the transfer of experimental results between species, for the immune regulation by CD39⁺ Treg cells we still regard ATP primarily as a ‘classical’ danger signal for both humans and mice.

In principle, we agree that the role of extracellular ATP hydrolysis is likely to be more complex than just ‘blunt immune suppression.’ Nothing is black or white; everything depends on the context. A prominent example is mouse transforming growth factor (TGF)-β, which normally drives the generation of immune-suppressive Treg cells but in the presence of IL-6 converts into a differentiation factor for pro-inflammatory Th17 cells.²  The same also applies for ATP. As quoted already in the report by Borsellino et al,¹  exposure of immature human dendritic cells (DCs) to ATP alone triggers only an incomplete maturation and produces cells unable to secrete pro-inflammatory cytokines. The cells are in fact tolerogenic as they release thrombospondin-1 and express indoleamine 2,3 dioxygenase.³  The same study,³  however, also revealed that ATP triggers the up-regulation of IL-23. This is in line with Schnurr et al,⁴  who showed that also in the presence of bacterial stimuli ATP potentiates the induction of IL-23, while blocking the generation of IL-12. Notably, IL-23 is the key-survival factor for Th17 cells, a recently discovered subset of effector/memory cells involved in autoaggressive immune attacks.²  Thus, depending on the environment, ATP acts on human immature DCs not only as an enhancer of tolerance-induction. During infection and ongoing immune reactions it may instead promote the expansion of pro-inflammatory Th17 cells.

Extracellular ATP plays a similar role also in the regulation of innate immunity. The generation of pro-IL-1β is induced by TLR-ligands such as LPS or CpG. Some reports indicate that ATP is involved in the release of the cytokine as it triggers the shedding of IL-1β-filled microvesicles from activated monocytes or mature DC.^5,6  More importantly, other studies have shown that extracellular ATP actually regulates the inflammasome-dependent conversion of pro-IL-1β into the mature cytokine.⁷  Thus, in the presence of TLR ligands, ATP acts as an immediate inducer of pro-inflammatory cytokines. On human Treg cells, CD39 is expressed on a subset of effector/memory-like cells (T_REM).¹  T_REM cells accumulate inside the inflamed tissue⁸  to which they have been attracted by factors such as the CCR6-ligand CCL20. The chemokine is released by activated leukocytes during infection and inflammation, so that they should encounter extracellular ATP usually only in a context where it acts as a pro-inflammatory danger signal.

In the absence of other signals indicating inflammation or infection, extracellular ATP may indeed exhibit some suppressive effects. Although the suggested inhibitory role of the P2Y11 receptor still needs to be established, it could directly inactivate human T cells by raising intracellular cAMP levels. However, high concentrations of ATP are toxic also for mouse T cells, where the effect is mediated by P2×7 receptors. Mouse Treg cells are even particularly sensitive to the nucleotide⁹  and it is namely the ecto-ATPase CD39 that allows them to cope with the elevated ATP levels in inflamed sites.¹  The close linkage of the ATPase-activity of CD39 to the activation status of the Treg cell in mice clearly points to an immune-suppressive function. Restricted expression of the enzyme by a subset of Treg cells acting inside inflamed tissue suggests that also in humans its primary function is the removal of an activating danger signal.¹  The generation of adenosine by CD39 and CD73, as demonstrated by Deaglio et al,¹⁰  may further add to the anti-inflammatory effect.