Introduction Regulatory T lymphocytes (Treg) play a key role in the control of autoimmunity. However, under inflammatory conditions like rheumatoid arthritis (RA), Treg become less suppressive and may shift toward a Th17 profile. On the contrary, polymorphonuclear neutrophils (PMN) are typical inflammatory cells activated in RA, although regulatory PMN sub-populations have been described. Treg and PMN are thus supposed to have opposite functions, but in both cases these functions can be reverted. Very few data are available on Treg-PMN communication in normal conditions, and even less during inflammation, especially in RA.

Objectives The aim was to analyse the interplay between Treg and PMN and their reciprocal modulation, both in physiological and pathological inflammatory environments.

Methods Splenic Treg and bone marrow PMN (C57BL/6 mice) were purified by magnetic sorting. Treg and PMN of healthy donors were freshly isolated by dextran sedimentation and magnetic sorting from peripheral blood. Co-cultures (1:1 ratio) were unstimulated or exposed to anti-CD3/anti-CD28 antibodies and/or LPS to stimulate Treg and/or PMN. CD4⁺FoxP3⁺ Treg (mouse and human), Ly6G⁺ (mouse) and CD66b⁺ (human) PMN were identified by flow cytometry. Cell activation was studied using antibodies against CD39, CD25, CTLA-4 (Treg) and CD11b (PMN). Treg maintenance was evaluated as the frequency of FoxP3 expression among CD4⁺ cells, and their proliferation by CFSE staining followed by flow cytometry analysis. Co-cultures were performed using Transwell (0.4 µm) to determine the respective involvement of soluble mediators/cell-contacts. Cytokine levels were quantified in culture supernatants by ELISA. Collagen-induced arthritis (CIA) was induced in C57BL/6 mice by immunisation with chicken type II collagen in complete Freund’s adjuvant at days 0 and 21.

Results Without stimulation of both Treg and PMN from normal mice, no effect on any cell type was observed in co-culture. In contrast, co-culture of activated Treg with activated PMN resulted in increased maintenance of Treg with higher CTLA4 but lower CD39 expression, sustained PMN activation evidenced by CD11b up-regulation, and higher secretion of MIP-2, IL-6 and IL-17 but not IFN-γ in normal mice. All these effects were lost in transwell experiments. Nevertheless, transfer of supernatants from LPS-activated PMN also partly increases Treg maintenance. In addition, co-cultures led to Treg proliferation. Similar results were observed in co-cultures of activated Treg/PMN isolated from CIA mice. Likewise, human Treg-PMN co-cultures led to enhanced Treg maintenance with higher expression of CTLA4/CD25 in a cell contact-dependent manner. Secretion of both IL-8 and IL-10 was enhanced in co-cultures.

Conclusions Our results show that cross-talk between Treg and PMN mainly leads to an activation of both cell types in normal or inflammatory conditions. Although cell contacts are clearly required, soluble mediators are also involved. Whether these synergistic interactions lead to a global suppressive or inflammatory milieu with functional modulation of either partner needs to be determined.

Disclosure of interest None declared