Background
Rheumatoid arthritis (RA) is a chronic inflammatory autoimmune disease characterized by inflamed synovial joints resulting in pain, fatigue, and disability in patients. Although treatment has improved over the last decades, many patients do not reach clinical remission, show progressive joint damage, or become resistant to their treatment [
1]. Furthermore, the current therapies include the use of expensive biological disease-modifying antirheumatic drugs (DMARDs) which pose a burden on the healthcare budget. Therefore, there is still a need to find new therapeutic options and to improve currently available treatments.
The main therapeutic goal in RA is to stop the chronic synovial inflammation and thereby prevent the subsequent cartilage and bone damage in the affected joint. Synovial fibroblasts play an important role in this process since they can secrete proinflammatory cytokines that attract and activate immune cells, produce tissue-destructive enzymes, and invade cartilage [
2,
3]. How these RA synovial fibroblasts (RASF) are activated is currently unknown, although one of the hypotheses is that they are activated by infiltrating immune cells. Previously we have shown that CCR6
+, but not CCR6
–, memory T helper (memTh) cells can activate RASF [
4].
CCR6
+ memTh cells are characterized by interleukin (IL)-17A production and RAR-related orphan C receptor (RORC) expression. This subset contains the classic Th17 cells, but also contains, for example, Th17.1 cells that produce high levels of interferon (IFN)γ [
5]. Other evidence that suggests a role for these IL-17A-producing memTh cells is that Th17 cells and IL-17-mediated signaling are required for the development of murine autoimmune arthritis [
6,
7]. Furthermore, CCR6
+ memTh cells are more activated and more prevalent in the blood of treatment-naive early RA patients and are also found in the synovial fluid of RA patients [
4,
8].
Upon interaction with CCR6
+ memTh cells, RASF secrete proinflammatory mediators such as IL-6, matrix metalloproteases (MMPs), and prostaglandin E2 (PGE2) via IL-17A and tumor necrosis factor (TNF)α. In turn, these molecules, especially PGE2, further activate the CCR6
+ memTh cells to produce more IL-17A, thereby creating a proinflammatory feedback loop that could drive the chronic synovial inflammation [
4,
9]. Therefore, inhibiting this proinflammatory loop may be beneficial in the treatment of RA.
We have previously shown that combining the active vitamin D metabolite 1,25(OH)
2D
3 with TNFα blockade, a commonly used therapy in RA, additively suppresses the proinflammatory loop between CCR6
+ memTh and RASF [
10]. However, clinical translation of these findings is challenging since high concentrations of 1,25(OH)
2D
3 were used. Interestingly, the effects of 1,25(OH)
2D
3 on Th17-related cytokines, such as IL-17A, TNFα, and IL-22, can be augmented through combination with dexamethasone (DEX) [
11]. DEX is a synthetic glucocorticoid (GC) which is clinically used for fast resolution of inflammation and is often combined with vitamin D supplements to prevent the osteoporotic side effects of the drug [
12].
Given the immunomodulatory capacities of both DEX and 1,25(OH)2D3, we here investigated whether this combination could suppress CCR6+ memTh cells, RASF, and their interaction. Furthermore, the potential use of these findings for improving current anti-TNFα therapy is explored.
Methods
Subjects
Healthy control peripheral blood mononuclear cells (PBMCs) were obtained from buffycoats (Sanquin, Amsterdam, the Netherlands). For validation of findings in healthy PBMCs, RA PBMCs were isolated from treatment-naive RA patients who were embedded in the tREACH study, which was ethically approved by the METC Rotterdam. Relevant clinical information is summarized in Additional file
1 (Table S1). RASF were grown from synovial explants after joint replacement surgery. All patients signed informed consent.
Cell sorting
PBMCs were isolated from peripheral blood using ficoll-based cell separation and frozen in liquid nitrogen until use. For sorting of CCR6+ memTh cells (CD4+CD45RO+CCR6+CD25low/int), PBMCs were stained using antibodies against CD4, CCR6, CD25 (BioLegend, San Diego, CA, USA), and CD45RO (BD Biosciences, San Diego, CA, USA). Before sorting CCR6+ memTh cells, the cells were prepurified using CD4 microbeads via automated magnetic-activated cell sorting (Miltenyi Biotec, Leiden, The Netherlands). Dead cells were excluded using 4′6-diamidino-2-phenylindole dilactate (DAPI) and CCR6+ memTh cells were sorted on a FACSAriaIII sorter (BD Biosciences).
Cell culture
RASF were obtained by culturing small synovial biopsies in a culture flask with Dulbecco’s modified Eagle’s medium (DMEM; Gibco, Waltham, MA, USA), supplemented with 10% fetal calf serum (FCS; Gibco) and 100 IU/ml penicillin/streptomycin (pen/strep; Lonza, Verviers, Belgium). After RASF were grown out of the synovial biopsies, cells were passaged and used for experiments between passage 3 and 8. For coculture with sorted T cells or stimulation experiments, RASF were plated at a density of 1 × 104 cells/well in a 96-well plate 24 h before the T cells or stimulation medium were added. Where indicated, RASF were stimulated with 2 ng/ml recombinant TNFα and 5 ng/ml recombinant IL-17A (R&D Systems, Minneapolis, MN, USA).
Sorted CCR6+ memTh cells were stimulated with 300 ng/ml soluble anti-CD3 and 400 ng/ml soluble anti-CD28 (Sanquin, Amsterdam, the Netherlands) at a density of 2.5 × 104 cells/ml in Iscove’s modified Dulbecco’s medium (IMDM; Gibco) supplemented with 10% FCS, 100 IU/ml pen/strep, 2 mM l-glutamine, and 50 μM β-mercaptoethanol (Sigma-Aldrich, St. Louis, MO, USA). Cells were treated with or without 1,25(OH)2D3 (Leo Pharmaceutical Products, Ballerup, Denmark), dexamethasone (Sigma-Aldrich), and etanercept (anti-TNFα, Pfizer, New York, NY, USA) dissolved in 100% ethanol at the indicated concentrations and added for the full duration of the culture simultaneously with the stimulatory compounds. Control conditions contained an equal volume of 100% ethanol which never exceeded 0.1%.
Flow cytometry
Cultured cells were restimulated with 50 ng/ml phorbol 12-myristate 13 acetate (PMA), 500 ng/ml ionomycin (Sigma-Aldrich), and GolgiStop (BD Biosciences) for 4 h. Cells were then stained with Fixable Viability Dye eFluor506 (eBioscience, San Diego, CA, USA), fixated with 2% paraformaldehyde and permeabilized using 0.5% saponine. Intracellular cytokines were stained with monoclonal antibodies against IL-17A, IL-22 (eBioscience), and IFNγ (BioLegend).
Apoptosis was assessed using 7AAD/Annexin V staining, performed according to the manufacturer’s instructions (eBioscience). All samples were measured on the FACSCantoII Flow Cytometer (BD Biosciences).
Enzyme-linked immunosorbent assay (ELISA)
In culture supernatant after 3 days of culture, the concentration of IL-17A, IL-22, IFNγ, TNFα, IL-10, IL-6, IL-8 (Ready-Set-Go, eBioscience), MMP1, MMP3 (DuoSet ELISA, R&D Systems), and PGE2 (Prostaglandin E2 Parameter Assay Kit, R&D Systems) were measured using ELISA. The manufacturers’ protocols were followed.
Statistical analysis
Differences between experimental treatment groups were tested using analysis of variance (ANOVA) with Bonferroni post-hoc tests. p values below 0.05 were considered statistically significant. Analyses were performed using Prism software version 6.01 (GraphPad Software, La Jolla, CA, USA).
Discussion
This study showed that 1,25(OH)2D3 and DEX can additively inhibit synovial inflammation modeled by CCR6+ memTh-RASF cocultures. Furthermore, combining low doses of DEX and 1,25(OH)2D3 with TNFα blockade demonstrated added value over TNFα blockade alone.
Similar to our previous results in PBMCs and memTh cells [
11], 1,25(OH)
2D
3 was a stronger modulator of IL-17A, IL-22, IFNγ, and IL-10 in CCR6
+ memTh cells than DEX. This could be due to the reported resistance of CCR6
+ memTh cells to cytokine inhibition and apoptosis induction by GCs [
13]. Interestingly, in asthma, GC resistance is increased with decreasing vitamin D serum levels [
14]. Since the vitamin D receptor can enhance the activity of the GC receptor at promoter sites, 1,25(OH)
2D
3 may be able to overcome GC resistance through this mechanism [
15]. Notably, in Crohn’s disease Th17.1 cells, one of the subpopulations within CCR6
+ memTh cells, are the most GC-resistant cells [
16]. Since these cells can also be found at the site of inflammation in juvenile idiopathic arthritis [
17] and RA (unpublished observations) further research into cell type-specific modulation by GCs may further elucidate its immunosuppressive actions in RA.
Independent of the differences in modulation of cytokines derived from CCR6
+ memTh cells, either cultured alone or together with RASF, both 1,25(OH)
2D
3 and DEX inhibited the activation of synovial fibroblasts as demonstrated by decreased IL-6, IL-8, MMP1, MMP3, and PGE2. These data suggest that DEX has strong direct effects on RASF, which was confirmed by RASF-only cultures. Although others have shown that 1,25(OH)
2D
3 can inhibit MMP1 and MMP3 from RASF under IL-1β stimulation [
18], this was not observed in our cultures with stimulation of TNFα and IL-17A. Interestingly, whereas TNFα stimulation of RASF was generally more potent than stimulation with IL-17A, the combination of TNFα and IL-17A induced a higher level of IL-6 and MMP1 and a striking stronger increase in IL-8 and MMP3. It has been previously observed in keratinocytes that IL-8 was more strongly induced by TNFα and IL-17A than IL-6 [
19]. Also, another study in RA fibroblast-like synoviocytes suggested a slightly stronger additive effect on IL-8 than IL-6, although they only used 1 ng/ml for both TNFα and IL-17A [
20]. Since DEX and 1,25(OH)
2D
3 inhibit IL-8 and MMP3 both after TNFα single stimulation or TNFα-IL-17A combination, this may be an effective way to suppress even the strong stimulation that is potentially present in the RA joint.
Based on other studies showing that the vitamin D receptor and GC receptor can cooperate to enhance one another’s functions [
15,
21], we also expected synergistic effects of 1,25(OH)
2D
3 and DEX in CCR6
+ memTh cells or RASF. However, this synergy was not observed, except that TNFα could only be inhibited when 1,25(OH)
2D
3 and DEX were combined. Instead, the data indicate that 1,25(OH)
2D
3 and DEX additively suppress inflammation through targeting different inflammatory pathways. 1,25(OH)
2D
3 indirectly reduces RASF activation through modulation of IL-17A and could reduce activation of other immune cells, such as macrophages, by inhibiting IFNγ [
22]. DEX, on the other hand, directly affects the RASF and reduces immune cell activation, immune cell attraction, and tissue destruction through regulation of IL-6, IL-8, and MMPs, respectively. Finally, 1,25(OH)
2D
3 and DEX are both capable of inducing the anti-inflammatory cytokine IL-10 in CCR6
+ memTh cell monocultures or in coculture with RASF, which could further contribute to inhibiting synovial inflammation.
Because of the strong immunomodulatory effects of 1,25(OH)
2D
3 and DEX, we postulated that they could be beneficial in the treatment of autoimmune diseases. This study demonstrated that DEX especially could augment the effect of TNFα blockade on RASF activation, even up to the point that no differences could be seen between the lowest and highest dose of anti-TNFα. Adding 0.1 nM 1,25(OH)
2D
3 did not contribute to this effect, but 10 nM enhanced the effects of DEX and TNFα blockade. The concentration of 0.1 nM (approximately 40 pg/ml) corresponds to the 20–80 pg/ml of 1,25(OH)
2D
3 that has been found in the synovial fluid of RA patients [
23]. However, the synovial fluid is not always a perfect representation of the situation in the synovium. Furthermore, it has been reported that immune cells are capable of converting 25(OH)D
3 into 1,25(OH)
2D
3 [
24,
25], suggesting that the local concentration of active vitamin D in the inflamed synovium may be higher than 0.1 nM and therefore could contribute to the anti-inflammatory effects of DEX and anti-TNFα.
Although these data indicate that a combination of DEX and TNFα blockade could be beneficial in the treatment of RA, some limitations of this study should be considered. Firstly, the current study has focused on an in-vitro culture model with only T cells and RASF, whereas an inflamed joint also contains other cell types [
26]. Although 1,25(OH)
2D
3 and DEX are both known for their wide range of immunomodulatory properties [
12,
27], the exact effects of combining these with TNFα blockade have not been investigated. Furthermore, the data suggest that dose reduction of TNFα blockade could be possible on combination with DEX since there is no dose-dependent effect anymore. It should be noted that the physiological concentration of etanercept in the synovial fluid has not been elaborately studied. One study found the concentration of etanercept in the synovial fluid was around 20 ng/ml in two patients receiving 50 mg etanercept every 2 weeks after 5 weeks, thus in between two dosages [
28]. However, one patient who received 50 mg every week had a concentration of 100 ng/ml after 5 weeks, suggesting that the concentration shortly after etanercept injection is in the range of the 0.1 μg/ml that was used in this study [
28]. The inhibitory effects of RASF activation by this dose of etanercept have been drastically enhanced by DEX and 1,25(OH)
2D
3.
A final point to consider is that TNFα blockade has more effects than only suppressing RASF activation [
29]. Therefore, it is possible that our in-vitro model overlooks interactions that may arise between the compounds and a more complex environment. To study this, murine models for RA, such as collagen-induced arthritis, could be used as a first validation of our study. After that, a clinical trial should be designed in which a low dose of DEX is combined with TNFα blockade. Due to the perceived immunomodulatory effects of 1,25(OH)
2D
3, adding vitamin D supplements to this treatment will not only prevent the osteoporotic side effects of DEX, but it may also further enhance the therapeutic effects.