Background
Interleukin-17 (IL-17) is a pro-inflammatory cytokine mainly produced by Th17 cells and involved in several chronic inflammatory diseases, such as psoriasis or rheumatoid arthritis (RA) [
1,
2]. Commonly, an association is made between Th17 cells and the secretion of IL-17, using intracellular staining of IL-17 as an equivalent of actual demonstration of protein secretion. As shown in one of our previous studies, IL-17-positive cells acquire a plasma cell-like morphology under in vitro activation, which is correlated with enhanced concentrations of secreted IL-17. Moreover, in sections of inflamed tissue, IL-17+ cells have this plasma cell-like morphology [
3]. The connection between IL-17 intracellular staining and IL-17 secretion thus remained to be clarified.
During chronic inflammatory diseases, the recruitment of activated immune cells, including Th17 cells, to the site of disease, leads to interactions with local mesenchymal cells and this promotes cell proliferation and the chronicity of inflammation [
4‐
6]. These interactions between bone marrow-derived or synovium-derived mesenchymal cells and activated peripheral blood mononuclear cells (PBMC) promote Th17 cell expansion [
7].
Regarding the crucial role of cell–cell interactions in pro-inflammatory cytokine production, a mechanism remained to be identified. Among many options, podoplanin (pdpn) was considered as a possible candidate, as a molecule involved in cell–cell interactions. Indeed, pdpn modulates IL-8 secretion during interactions between platelets and synoviocytes through its receptor CLEC-2 on platelets [
8]. Moreover, the pdpn pathway is involved in different animal models of inflammatory diseases [
9,
10]; notably in a mouse model of RA, pdpn is upregulated in Th17 cells compared to other Th cell subsets [
9]. Pdpn expression is also upregulated in RA synovium and it might increase the migratory potential of activated synoviocytes in RA [
11]. These results lead us to consider pdpn as a potential new mechanism involved in the Th17 pathway during chronic inflammation.
The aim of this study was to clarify the difference between intracellular expression from IL-17 secretion and to identify a mechanism which leads to high IL-17 secretion during interactions between PBMC and synoviocytes, with a special interest in podoplanin.
Materials and methods
Samples
RA synoviocytes were obtained from synovial tissue of RA patients undergoing joint surgery and who fulfilled the American College of Rheumatology criteria for RA [
12]. Synovial tissue was minced into small pieces and then adhered in 6-well plates in Dulbecco’s modified Eagle’s medium (DMEM; Eurobio, Courtaboeuf, France) supplemented with 10 % fetal bovine serum (FBS; Life Technologies, Carlsbad, CA, USA), 2 mM L-glutamine and 100 U/ml penicillin/streptomycin. Cells were maintained at 37 °C in a humidified 5 % carbon dioxide incubator and used between passages 4 to 9. They are fibroblast-like synoviocytes which express pdpn and secrete IL-6 after stimulation by tumor necrosis factor (TNF) [
11,
13]. PBMC from healthy donors or RA patients were isolated by Ficoll-Hypaque (Eurobio, Courtaboeuf, France) density-gradient centrifugation. Adipose-derived stem cells (ASC) and human umbilical vein endothelial cells (HUVEC) were used as control cells.
T cell cloning procedure
To obtain CD4+ T cell clones, CD4+ T cells were isolated from peripheral blood mononuclear cells (PBMC) of healthy donors by immunomagnetic cell separation (Miltenyi Biotech, Bergisch Gladbach, Germany) and further divided into the CD161 + CCR6+ and CD161-CCR6- fractions by flow cytometry cell sorting (FACSAria, BD Bioscience, Franklin Lakes, NJ, USA). Both cell fractions obtained were seeded under limiting-dilution conditions (0.5 cell/well) in round-bottom microwell plates, containing 10
5 irradiated (60 Gy [9000 rad]) allogeneic peripheral blood mononuclear cells as feeder cells, 1 % phytohemagglutinin (PHA; vol/vol), and recombinant human IL-2 (50 U/mL). Blasts were expanded in the presence of feeder cells (10
5 cells/well) plus IL-2 (50 U/mL). T cell clones were obtained and evaluated for their cytokine production activity (interferon [IFN]-γ, and IL-17) by flow cytometry after stimulation with PMA plus ionomycin, as previously described [
14,
15]. T cell clones producing IL-17 alone were classified as Th17. Production of cytokines by each clone was arbitrarily considered as significant when the proportion of producing T cell blasts was >20 %. Th17 clones obtained from the CD161 + CCR6+ fraction were selected and further expanded in the presence of feeder cells plus IL-2 until they reached the number of approximately 10
6 blasts/clone, then they were transferred in 24-well plates and maintained in culture in the presence of IL-2 (50 U/mL).
Co-culture assays
Co-culture was initiated by seeding RA synoviocytes, ASC or HUVEC overnight in 96-well plates at a density of 2 × 104 cells/well in RPMI 1640 medium (Eurobio, Courtaboeuf, France) supplemented with 10 % human AB serum, 2 mM L-glutamine and 100 U/ml penicillin/streptomycin (complete RPMI). The next day, PBMC (1 × 105 cells/well) or Th17 clones (2 × 104 cells/well) were seeded in complete RPMI corresponding to 5:1 ratio or 1:1 ratio, respectively, in the presence or absence of phytohemagglutinin (PHA, 5 μg/ml). After 48 h, supernatants and cells were collected for analysis.
Transwell assay
RA synoviocytes were seeded in 24-well plates at a density of 1 × 105 cells/well in complete RPMI. After overnight incubation, PBMC were added directly on top of synoviocytes or in a cell culture insert (0.4 μm pore size) at a concentration of 5 × 105 cells/well, in the presence or absence of PHA (5 μg/ml). After 48 h, supernatants and PBMC were recovered for analysis.
Cell fixation
Synoviocytes or PBMC were fixed for 1 h, at 4 °C, in phosphate-buffered saline (PBS) 0.01 % paraformaldehyde before co-culture.
Enzyme-linked immunosorbent assays (ELISA)
IL-17A, IL-6, and IL-1β productions were evaluated from culture supernatants with commercially available DuoSet ELISA kits, according to the manufacturer’s instructions (R&D Systems, Minneapolis, MN, USA).
Flow cytometry
Allophycocyanin (APC), phycoerythrin (PE), PE-cyanine-7 or eFluor 450-conjugated antibodies (eBiosciences, San Diego, CA, USA) were used to stain cells. EFluor 450-CD3 (48–0038), PE-Cy7-CD4 (25–0049) and PE-pdpn (12–9381) were used for cell surface staining, according to the manufacturer’s instructions. PBMC were fixed and permeabilized for APC-IL-17A (17–7179) intracellular staining. Cells were incubated in cold PBS and cold 2 % paraformaldehyde in the dark during 1 h for fixation step. For permeabilization, cells were incubated in PBS with 0.2 % Tween at 37 °C for 15 min, before intracellular staining. Flow cytometry staining buffer from eBiosciences (San Diego, CA, USA) was used for staining protocol. Analysis was done with the Kaluza software (Beckman Coulter, Brea, CA, USA).
Monocyte contribution
To remove monocytes, PBMC were pre-incubated during 2 h at 37 °C. The percentage of removal monocytes by adherence was verified with flow cytometry. This method allowed removing at least 50 % of monocytes.
Inhibition of podoplanin
A dose–response curve was performed to determine the concentration of purified anti-human podoplanin antibody (14–9381, eBiosciences, San Diego, CA, USA) required for maximal inhibition. PBMC were pre-incubated for 4 h with different concentrations of anti-podoplanin (0, 1, 5, 10 and 20 μg/ml) before co-culture assay. According to the results of dose–response curve, all experiments studying podoplanin were realized with anti-podoplanin at 5 μg/ml. A control antibody against the BetV1 allergen with no effect in the assays was used at the same concentration (anti-BetV1 Ab, Dendritics, Lyon, France).
Small interfering RNA (siRNA)
A mixture of four small interfering RNA (siRNA) provided as a single reagent (siGENOME human SMARTPool siRNA) specific for podoplanin (M-017560-01-0005) was purchased from Dharmacon, Lafayette, CO, USA. RA synoviocytes were seeded at a density of 5 × 105 cells/12-well plate before transfection and used at 80–90 % confluence. Cells were transfected with control siRNAs (siCONTROL siRNA as a negative control (mock) and siGLO peptidylpropyl isomerase B (PPIB) (cyclophilin B) as a positive control) or target siRNAs (siGENOME SMARTPool pdpn siRNA) by nucleofection (Amaxa Pharma, London, UK) according to the manufacturer’s instructions (program U23; Human Dermal Fibroblast Nucleofector kit). Dose– and time–response experiments were performed to determine the best time point and the best suitable concentration of siRNA duplexes needed for efficacious RNA silencing. Cells were nucleofected with 37.6, 100 or 376 nM of pdpn siRNA (siPdpn), per 5 × 105 cells. Twenty-four hours or 48 h post-transfection, RNA was extracted and RT-PCR performed to detect the RNA silencing. The best condition was at 100 nM, at 48 h, thus this condition was used for the presented experiments.
RNA extraction and real-time PCR
Total RNA of synoviocytes was extracted using the RNeasy Mini Kit (Qiagen®, Hilden, Germany) and quantified with the Quant-it kit assay (Invitrogen™ by Thermo Fisher Scientific, Grand Island, NY, USA) following the manufacturer’s instructions. cDNA was synthesized using the QuantiTect reverse transcription kit (Qiagen®) according to the manufacturer’s instructions. SYBR green-based real-time qRT-PCRs were performed on the CFX96 Real-Time PCR Detection System (BioRad, Hercules, CA, USA) using the QuantiFast SYBR green kit and QuantiTect primers (Qiagen®). Cycle threshold values were normalized with respect to the endogenous control gene glyceraldehyde 3-phosphate dehydrogenase (GAPDH). The relative expression of pdpn and PPIB genes in the different conditions was determined using the comparative threshold cycle method as described by the manufacturer.
Statistical analysis
Statistical analyses for co-culture assays were performed using two-way ANOVA. For transwell, monocytes and podoplanin experiments a paired nonparametric Wilcoxon test was used. All analyses were performed with GraphPad Prism 6 software (GraphPad Software, San Diego, CA, USA). p values less than or equal to 0.05 were considered as significant.
Discussion
Cell interactions between mesenchymal and immune cells are known to induce the production of pro-inflammatory cytokines and also to affect the survival of both cell types [
7,
20‐
22]. In the context of RA, a co-culture system between RA synoviocytes and PBMC was used to mimic the in vivo situation. Cell interactions were sufficient to provide a necessary activation state for the secretion of IL-6 and IL-1β and this is in agreement with a previous study showing that MSC-PBMC interactions increased IL-6 and IL-1β mRNA [
7]. These observations reflect how high levels of pro-inflammatory cytokines, including IL-6 and IL-1β, can be produced locally by the RA synovium [
23]. Furthermore, co-culture with autologous cells, PBMC and synoviocytes from the same patient, validated our co-culture in vitro model mimicking the in vivo inflammation.
IL-6 and IL-1β are known to be involved in Th17 cell differentiation [
16‐
18]. Th17 cells in turn secrete IL-17 which acts on synoviocytes, often in synergy with other cytokines such as TNF-α, IL-1β or granulocyte-macrophage colony-stimulating factor (GM-CSF) [
2,
24‐
26]. Considering the results on IL-6 and IL-1β production, the effect of cell interactions on the Th17 pathway was studied to differentiate the intracellular expression from the secretion of IL-17. The percentage of CD3 + CD4 + IL-17+ cells was evaluated in PBMC alone or in co-culture. Cell contact alone had no major effect on Th17 differentiation measured by intracellular staining. Only TCR activation had a modest effect. This indicated that Th17 differentiation requires cell activation more than cell–cell contact.
When looking at the production of IL-17 and in contrast to that of IL-6 and IL-1β, cell interactions were not sufficient to induce a high IL-17 secretion. Its production required two signals, TCR activation and cell–cell contact. Moreover, activation of synoviocytes by TNF alone in co-culture without TCR stimulation had no effect on IL-17 production. In fact, IL-17 secretion during cell interactions was dependent on T cell but not synoviocyte activation. Transwell experiments confirmed that cell interactions were crucial to have an elevated IL-17 secretion even in the presence of TCR activation.
These results reveal a major discrepancy between intracellular and secreted IL-17. The intracellular presence of IL-17 inside Th17 cells does not mean by itself a high secretion of IL-17. The presence of Th17 cells even with TCR activation was not enough to have the release of high levels of IL-17. It was necessary that activated cells could physically interact with mesenchymal cells, derived from either synovium, bone marrow, or adipose tissue. Thus, TCR activation and cell–cell contact are two needed mechanisms leading to a high IL-17 production and this differs from intracellular expression of IL-17. Both mechanisms are present during RA pathogenesis [
13,
27‐
30], and this could explain the presence of IL-17 in the joints of RA patients [
31,
32].
Furthermore, the secretion of IL-17 was variable depending on experiments which used different donors for PBMC and RA synoviocytes. This variability reflected the heterogeneity which characterizes the RA population suggesting the variable contribution of IL-17 in the inflammatory process. The variability in IL-17 secretion observed in our experiments could explain in part the heterogeneity of the response to an anti-IL-17 treatment in RA patients [
33,
34]. Such heterogeneity remains to be explained. One explanation could be the contribution of gene polymorphisms in the regulation of the Th17 pathway.
IL4R gene polymorphisms have been associated with RA severity by increasing the Th17 cell frequency and by modulating the inhibitory effect of IL-4 on Th17 development [
35] and the modulation of IL-17 production [
36].
IL-23R polymorphisms have been implicated in IL-17A expression in RA [
37].
The critical contribution of interactions between immune cells and mesenchymal cells indicated the need to identify a molecular mechanism. The limited contribution of monocytes suggested a molecule present on lymphocytes or on mesenchymal cells or on both. Pdpn, which is a type I transmembrane protein, appeared as a good candidate. Pdpn-mediated interaction of RA synoviocytes and platelets modulates IL-8 production [
8]. Furthermore, in the SKG spontaneously occurring arthritis model, pdpn is upregulated in Th17 cells compared to other Th cell subsets [
9] and in the synovium of RA patients [
11]. In a mouse model of multiple sclerosis, mice treated with anti-pdpn present a delayed onset of symptoms [
10]. Based on these observations, an antibody against pdpn was used in the co-culture system and siRNA specific for pdpn was used on synoviocytes. Both means induced an inhibition of IL-17 production and confirmed the role of pdpn in the IL-17 secretion.
These results focused on the podoplanin expressed by RA synoviocytes but it was known that Th17 cells could express pdpn, notably in an experimental arthritis model and in clinical RA [
38,
39]. In accordance with this, the three different tested protocols, the pre-incubation of PBMC, the pre-incubation of synoviocytes or the pre-incubation of both cells, gave similar results (data not shown). Acting first on synoviocytes or PBMC did not affect the inhibitory effect of the anti-pdpn. This is in line with the expression of pdpn by Th17 cells and the fact that the effect of anti-pdpn could be both direct on Th17 cells and indirect by acting on synoviocytes to inhibit the IL-17 production during cell interactions. In addition, the lower percentage of inhibition obtained with siPdpn compared with the anti-pdpn Ab could also indicate the respective involvement of pdpn expressed by synoviocytes and by Th17 cells. The regulation of pdpn in PBMC and specifically in Th17 cells remains to be clarified.
The interaction between pdpn and its receptor could occur in the two directions, from synoviocytes to PBMC, or from PBMC to synoviocytes. The receptor of pdpn CLEC-2 is known to be mainly expressed by platelets [
40] and also by mature dendritic cells or peripheral blood B lymphocytes [
41‐
44]. Its expression in our co-culture system could be studied to provide a new insight on the pathway by which pdpn could influence the IL-17 secretion. Currently, there is no evidence for its expression on Th17 cells and this could also suggest a possible new receptor for pdpn. A recent study has shown that pdpn can negatively regulate CD4+ effector T cell functions through pdpn-CLEC-2 interaction [
45]. A high pdpn expression was found on “nonpathogenic” Th17 lymphocytes while “pathogenic” Th17 cells expressed less pdpn. Thus, pdpn displays two divergent functions which may depend on different ligands. One ligand, such as CLEC-2 could mediate T cell inhibition while another ligand could promote inflammation by stimulating pro-inflammatory cytokine production. Furthermore, our results demonstrated that the inhibition of the interaction mediated by pdpn decreased at least by 50 % the secretion of IL-17 and of IL-1β, but not that of IL-6. Furthermore, in both PBMC and synoviocytes, pdpn expression was increased in co-culture with TCR activation which correlates with the high IL-17 production. These results suggested that cell interactions of synoviocytes with activated immune cells increased pdpn expression that contributed to the high IL-17 secretion.
If podoplanin seems to be a good potential therapeutic target, investigating the effect of cell interactions on other signaling molecules involved in Th17 differentiation and function could be interesting. Indeed, the interaction with mesenchymal stem cells (MSC) could on the one hand enhance Th17 activity [
7,
46] but on the other hand it could repress Th17 molecular program through PD-1 [
47]. Furthermore, IL-17A can induce soluble PD-1 (sPD-1) which level is increased in RA serum. This overexpression of sPD-1 might block the inhibitory PD-1 pathway [
48]. Investigating the PD-1 pathway in co-culture system could allow identifying another mechanism. Signaling lymphocytic activation molecule (SLAM) is another candidate as it promotes the differentiation of IL-17-secreting effectors [
49]. Inducible T cell costimulator (ICOS) signaling, which belongs to the CD28 costimulatory molecule superfamily, has been also shown to play a critical role in the generation and maintenance of human Th17 cells [
50] and it could be another candidate.
Abbreviations
ASC, adipose-derived stem cells; ELISA, enzyme-linked immunosorbent assay; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; HUVEC, human umbilical vein endothelial cells; IFN, interferon; IL, interleukin; MSC, mesenchymal stem cells; PBMC, peripheral blood mononuclear cells; PHA, phytohemagglutinin; pdpn, podoplanin; PPIB, peptidylpropyl isomerase B; RA, rheumatoid arthritis; siRNA, small interfering RNA; TCR, T cell receptor; TNF, tumor necrosis factor
Acknowledgements
The authors acknowledge Dr. Odile Damour, Laboratoire des substituts cutanés, Hôpital Edouard Herriot, for providing ASC and Pr. Francesco Annunziato, Department of Experimental and Clinical Medicine, University of Florence, for providing Th17 clones.