We employ the paradigm that the production of CCL18 as well as Interleukin-1 receptor antagonist (IL-1Ra, supplementary data Fig. 1A, B) on one side and IL-1β as well as TNFSF2 (supplementary data Fig. 2A, B) on the other side may mirror the extreme endpoints of the macrophage activation pattern between M1 and M2 macrophages. Unstimulated monocytes cultured for 24h do not release CCL18 but release high levels of IL-RA [9]. Culture of these cells for 72h, however, induces high levels of CCL18 and IL-1RA, indicating an M2 shift. Interestingly, early stimulation with the danger signal ATP prevented the M2 differentiation of monocytes, seen in a consistent and dose-dependent drop in CCL18 levels. On the other hand, IL-1b is only marginally produced possibly induced via danger signals released by dying cells in the culture. Stimulation with ATP did not induce or fortify additional IL-1b, which would be indicative for an M1 shift. This might be due to the well-known effect that ATP alone without co-stimulation through LPS has no significant effect on IL-1β secretion [35-37]. In addition, it has been shown that monocytes may produce ATP by their own which is sufficient to activate the IL-1b maturation here and do not require external ATP [38].
Early stimulation with ATP prevented M2 differentiation without affecting M1 differentiation in a model of inflammation using LPS as a pro-inflammatory stimulator. In this case, LPS alone showed only minimal influence on CCL18 release. The influence of LPS on CCL18 release has been discussed for alveolar macrophages and monocytes in the literature. In alveolar macrophages derived from bronchoalveolar lavage, Kollert et al. demonstrated that LPS downregulated CCL18 release in smokers [39], but not in non-smokers. A similar effect with only a marginal impact of LPS on the secretion of CCL18 was described before in vitro in monocyte derived dendritic cells [40]. In our study, ATP together with LPS profoundly downregulated CCL18 release and inhibited the M2-polarization. LPS lead to a significant release of IL-1β with no additional effect of the co-stimulation with ATP. As previous studies showed, ATP alone seemed to have no effect on the accumulation of intracellular pro-IL-1b without stimulation with LPS [37, 41]. Monocytes seem to be able to release IL-1b in response to LPS alone without addition of external ATP and without P2X7 receptor activation [42], whereas macrophages seem to require a secondary stimulus like ATP [43]. Ferrari et al. showed in their pioneering study that the IL-1b-levels produced by macrophages after stimulation with LPS alone were lower than after LPS plus ATP (with a peak at 1mM ATP) and that this effect was mediated by P2X7 receptors by different mechanisms [36, 37]. Furthermore different subsets of monocytes show different patterns of IL-1b secretion after stimulation with LPS with higher levels of IL-1b after LPS stimulation of classical monocytes compared to the non-classical subset in both an ATP-dependent and -independent manner although P2X7 activity was comparable [44].
The classic model of IL-1 release suggests that cells producing IL-1b require the activity of Caspase-containing multi-protein complexes (inflammasomes) to cleave pro-IL-1b to IL-1b. For their priming of the inflammasomes, the cells require a Toll-like receptor (for example LPS) or cytokine signal, and afterwards an additional activating signal (like ATP) in this model [45]. Vice versa, monocytes can release IL-1β P2X7-independently through Toll-like-receptor (TLR)-stimulation by LPS [42], very likely because monocytes are able to release ATP by themselves after stimulation of their pathogen-sensing receptors thereby inducing IL-1β secretion in an autocrine way [38].
Our data demonstrate that native monocytes are able to differentiate intrinsically and without further stimulation to CCL18-releasing M2 (alternatively activated)-like macrophages as over time the levels of CCL18 increased. This ability was significantly diminished in the presence of extracellular ATP.
The discrepancy between high extracellular ATP levels and a predominance in M2 macrophage that has been described in tumor and lung fibrosis [26, 33, 46] could be explained that intrinsically differentiated M2 macrophages migrate to the tissues. These macrophages and monocytes that have developed to M2 like macrophages without ATP stimulation would be inert to following ATP stimulation as our results show a temporal limitation of differentiation.
The shift preventing a M2 activation pattern is time dependent. ATP has to be present early at the initiation of the culture. Addition of ATP at later time points (here 48h), when in-vitro differentiation of macrophages is already initiated or even established, showed no effect on the release of CCL18 and IL-1b.
Immune cells express a large variation of purinergic P2X and P2Y receptors. In macrophages P2Y2-receptors are the predominant subtype, while previous investigations showed that there are also P2Y1-, P2Y4, P2Y11-, and P2Y12-receptors expressed on alveolar macrophages and the receptors P2Y1, P2Y2, P2Y4 and P2Y6 on monocytes [30]. Regarding ligand-gated ion-channels, P2X7 is the most important subtype in macrophages and monocytes beside P2X1 and P2X4 [31, 47]. ATP and the potent P2X7 agonist BzATP disclose a virtually indistinguishable pattern of down-regulation of CCL18, indicating that P2X7 might play an important role in the shift of differentiation considering that BzATP is not a specific agonist at the P2X7 receptor. BzATP is a P2X7 receptor agonist that exhibits 5 - 10 fold greater potency than ATP [48-50]. ATP activates P2X7 receptor leading to recruitment and activation of pro-inflammatory cells and production of the pro-inflammatory cytokine IL-1β and inflammatory mediators [34, 45, 51].
The expression of P2X7 receptors of monocytes is positively regulated by the proinflammatory cytokines TNFSF2 and IFNg and by LPS [52]. We demonstrated dropping levels of CCL18 with rising concentrations of ATP. A comparable decreasing effect on proinflammatory and antifibrotic cytokines like CCL2 and CCL3 was described for dendritic cells with an involvement of the receptors P2Y1 and P2Y11 [53]. The non-selective antagonists PPADS and Suramin (which block most P2X and P2Y receptors) and KN62 (which is a P2X7-antagonist) did not revert the ATP-induced reduction of the CCL18 release. The early ATP-induced drop of CCL18 release and thus the prevention of spontaneous M2 differentiation of the cultured monocytes seems to be independent of P2X and P2Y receptors as the levels of CCL18 and IL-1b remained unchanged after pre-incubation with these antagonists of purinergic receptors.
However, the ineffectiveness of Suramin and PPADS could also be explained by the mechanism that ATP promotes its effects via P2X7-receptor-activation, because these antagonists have no effect on P2X7. In contrast, the ineffectiveness of the specific potent P2X7-inhibitor KN62 is not completely understood. Our data imply that KN62 did not influence the down-regulatory effect of ATP on CCL18 release. This incapability regarding the blockade of the P2X7 receptor has to take into account that KN62 is not a specific P2X7 receptor antagonist but its affinity to P2X7 is 5 to 30 times greater than to other receptors [48-50]. Besides, KN62 has multiple pleiotropic effects. For example KN62 also binds to the Calmodulin-dependent protein-kinase CaMK II and inhibits its activation at a IC50 of 0.9µM. CaMK II is discussed to be involved in M1 activation [54] and its inhibition has been shown to reduce CCL3-release [55]. Thus, it is conceivable that CaMK II inhibition by KN62 might foster M2 activation leading to an unchanged CCL18 release.
Nonetheless, P2X7 seems to be an important receptor in the M2/M1 polarization as P2X7 was considerably downregulated during in-vitro macrophage generation whereas the other P2-receptors analyzed remained stable. The shift towards an anti-inflammatory phenotype was described before not due to a loss of P2X7 but to an uncoupling of the receptor from the activation of the inflammasome in macrophages that are already shifted towards a M2 phenotype [56]. The expression of P2X7 is positively regulated by proinflammatory cytokines TNFSF2, IFNg and by LPS and downregulated by stimulation with TGFb and cAMP [52, 57]. This implies that P2X7-receptors are elevated in M1 subsets and reduced in the M2 subpopulation. We showed that the loss of P2X7 gives rise to anti-inflammatory and pro-fibrotic M2 macrophages which fits into this picture.
A comparable observation to our findings with a loss of P2X7 was made before, where a monocyte-/macrophage-like phenotype was induced in HL60 promyelocytes to determine expression of purinergic receptors [58]. There was 10-fold downregulation of P2X7 at 24 hours after monocytic differentiation that was followed by a marked loss of P2X7 expression after 48h.