1,25(OH)2D3-DEX combination treatment suppresses TNFα production from CCR6+ memTh cells
As a first step to examine the potential use of 1,25(OH)2D3 and DEX to inhibit synovial inflammation, CCR6+ memTh cells were sorted from healthy controls and cultured for 3 days with 1,25(OH)2D3, DEX, or both. In line with the previous studies, treatment with 1,25(OH)2D3 reduced the percentage of cells producing IL-17A, IL-22, and IFNγ. On the other hand, treatment with DEX did not affect the percentage of IL-22- or IFNγ-producing cells. The percentage of IL-17A-producing cells was reduced by DEX, but significantly less than by 1,25(OH)2D3. Combining 1,25(OH)2D3 and DEX had a similar effect as 1,25(OH)2D3 single treatment (Fig. 1a, b). Notably, treatment with DEX significantly reduced the amount of IL-17A and IFNγ that was produced during the 3-day culture period (Fig. 1c). Since DEX is known to induce cell death, the treatment effect on apoptosis was investigated. However, the minimal increase in apoptosis on DEX or combination exposure could not explain the difference between flow cytometry and ELISA (Additional file 1: Figure S1). Although no additive effect was observed with 1,25(OH)2D3-DEX combination treatment on IL-17A, IL-22, or IFNγ, TNFα was significantly inhibited by the combined exposure (Fig. 1c). Since 1,25(OH)2D3 can induce an anti-inflammatory phenotype in the CCR6+ memTh cells (Dankers et al., submitted manuscript), the combinatory effects of 1,25(OH)2D3 and DEX on IL-10 were also assessed. However, DEX did not affect IL-10 production when used alone and reduced the induction of IL-10 by 1,25(OH)2D3 when used in combination (Fig. 1c). Overall, these data show that 1,25(OH)2D3 is the strongest modulator of IL-17A, IL-22, IFNγ, and IL-10 production by CCR6+ memTh cells, but the combination of 1,25(OH)2D3 and DEX provides additional value through enhanced TNFα inhibition.
1,25(OH)2D3 and DEX additively suppress the proinflammatory feedback loop between CCR6+ memTh cells and RASF
Since 1,25(OH)2D3 and DEX had additive effects in blocking TNFα and they were both capable of reducing the level of IL-17A that is produced by CCR6+ memTh cells, we hypothesized that the combination treatment would augment inhibition of the proinflammatory loop between CCR6+ memTh cells and RASF when compared with either compound alone. Similar to the effects of 1,25(OH)2D3 and DEX on CCR6+ memTh cell cultures, treatment of the CCR6+ memTh-RASF cocultures significantly reduced IL-17A, IL-22, and IFNγ production, especially with 1,25(OH)2D3. However, the production of TNFα was stimulated upon treatment with DEX, whereas it was still inhibited with the combination treatment. Furthermore, IL-10 was significantly upregulated in response to DEX and the combination of DEX and 1,25(OH)2D3, but not 1,25(OH)2D3 alone (Fig. 2).
Despite the differences in modulation of T cell-derived cytokines by DEX and 1,25(OH)2D3, they equally inhibited the RASF-derived factors IL-6, IL-8, and PGE2. For IL-6 and IL-8 there was also a trend towards additive inhibition in response to the combination of 1,25(OH)2D3 and DEX. Interestingly, DEX treatment led to a significantly greater inhibition of the tissue-destructive enzymes MMP1 and MMP3 than 1,25(OH)2D3 treatment (Fig. 2). These data suggest that, although no additive effect of 1,25(OH)2D3 and DEX is found on individual cytokines, the two compounds cooperate to reduce the proinflammatory milieu at an inflammatory synovial site by each targeting different players in the inflammation. Furthermore, the increased production of IL-10 potentially mediates anti-inflammatory effects on other immune cells.
DEX is a stronger inhibitor of proinflammatory factors from RASF than 1,25(OH)2D3
The finding that DEX is an equal inhibitor of IL-6 and IL-8 and stronger inhibitor of MMP1 and MMP3 than 1,25(OH)2D3 while being less efficient in inhibiting T cell-derived cytokines in the CCR6+ memTh-RASF cocultures suggests a direct effect of DEX on RASF. To investigate this, RASF were cultured with or without stimulation of TNFα, IL-17A, or both for 3 days and treated with 1,25(OH)2D3, DEX, or a combination (Fig. 3). Without stimulation, 1,25(OH)2D3 did not significantly affect cytokine production by RASF, whereas DEX reduced the levels of IL-6 and IL-8. Upon stimulation with TNFα, both 1,25(OH)2D3 and DEX inhibited IL-6 and IL-8, whereas MMP1 and MMP3 were only significantly inhibited by DEX or the combination treatment. Similar patterns were observed with stimulation using IL-17A or a combination of IL-17A and TNFα. Together with our previous findings, these data suggest that, although DEX and 1,25(OH)2D3 can affect both T cells and RASF, DEX acts most strongly on RASF whereas the main effect of 1,25(OH)2D3 is mediated via the T cells.
Low dose of 1,25(OH)2D3 and DEX improves the effects of TNFα blockade in CCR6+ memTh-RASF cocultures
Since 1,25(OH)2D3 and DEX additively inhibit the proinflammatory loop between CCR6+ memTh cells and RASF, we next studied whether the 1,25(OH)2D3-DEX combination can enhance the anti-inflammatory effect of TNFα blockade in this model system. To make these experiments more physiologically relevant, we used a dose-testing experiment to determine whether physiologically relevant dosages of 1,25(OH)2D3 and DEX still affected cytokine production in the CCR6+ memTh-RASF coculture system; 10 nM DEX still inhibited IL-6, IL-8, and MMP1 in CCR6+ memTh-RASF cocultures, but 0.1 nM 1,25(OH)2D3 only slightly affected these proinflammatory factors and did not appear different from the 1 nM dose (Additional file 1: Figure S2). Therefore, the value of adding DEX and 1,25(OH)2D3 to etanercept was assessed using 10 nM DEX and 0.1 or 10 nM 1,25(OH)2D3. CCR6+ memTh cells were sorted from healthy controls and cultured together with RASF while being exposed to various combinations of 1,25(OH)2D3, DEX, and the TNFα-blocking agent etanercept. Inhibition of IL-17A, IL-6, IL-8, MMP1, and MMP3 is shown as heatmaps in Fig. 4 and further detailed in Additional file 1 (Figure S3).
For IL-17A, etanercept dose-dependently inhibited cytokine expression, but this effect was stronger (darker shades) when the cells were also treated with 10 nM 1,25(OH)2D3. Combining DEX with this dose of 1,25(OH)2D3 further suppresses IL-17A and leaves little added effect for etanercept. IL-6 and IL-8 are more strongly inhibited by etanercept than IL-17A and show a trend towards a dose response (left column of the heatmap, top to bottom). Adding increasing dosages of 1,25(OH)2D3 augments the effects of etanercept, but the optimal effect is reached when 10 nM DEX, 10 nM 1,25(OH)2D3, and etanercept are combined. Notably, under these conditions there is no significant difference between 0.1 or 10 μg/ml etanercept and cytokine expression is more than 90% reduced. MMP1 and MMP3 show a similar pattern of inhibition as IL-6 and IL-8, except that the effect of DEX is stronger even without additional 1,25(OH)2D3 or etanercept. These data suggest that combining 1,25(OH)2D3 and especially DEX with etanercept has additive effects compared with etanercept alone.
Since the cells from the healthy controls that were used in Fig. 4 and Additional file 1 (Figure S3) may react differently to treatment than the cells from RA patients, the experiment was repeated using sorted CCR6+ memTh cells from treatment-naive early RA patients cultured with RASF serving as a proof of principle (Fig. 5 and Additional file 1: Figure S4). Due to large variation between patients (Additional file 1: Figure S4), there was a less clear dose-dependent inhibition of IL-17A in response to etanercept. However, 10 nM 1,25(OH)2D3 still enhanced the effects of etanercept, and IL-17A inhibition was again further increased when 10 nM DEX was added. The inhibition of IL-6 and IL-8 by etanercept was stronger in the CCR6+ memTh cells from RA patients than those from healthy individuals. Similar to the results for the healthy controls, both 1,25(OH)2D3 and DEX enhanced the effects of etanercept. MMP1 and MMP3 were also inhibited by etanercept and their expression was further suppressed by DEX (Fig. 5).
Altogether, the data from Figs. 4 and 5 show that combination with DEX, and to a lesser extent 1,25(OH)2D3, provide a beneficial effect over TNFα blockade alone in a model for synovial inflammation.