Comparative analysis of genes regulated by TNFa or IL6/sIL6R signaling in SF
To investigate the relative contribution of IL6 to the SF inflammatory response, we analyzed the expression pattern of a large group of cytokines, chemokines and matrix metalloproteases (MMPs) with important roles in RA pathophysiology. We first confirmed the lack of effect of IL6 alone compared to IL6/sIL6R or TNFα on the expression of known target genes such as CCL2 and IL6 (data not shown) in cultured SF. Furthermore, despite individual baseline differences on gene expression, both RA and non-RA cultured SF respond similarly to TNFa and/or IL6/sIL6R stimulation (data not shown) and therefore, SF from both healthy and RA donors were indistinctly used.
In dose-response experiments, RT-qPCR analyses showed that TNFα induced the expression of the cytokine IL6, chemokines genes il8/cxcl8, ccl2, CCL5 and ccl8, and MMPs genes mmp1, mmp3 and mmp10 (Fig. 1a). IL6/sIL6R induced the expression of IL6 itself, mononuclear-cells recruiting chemokines such as ccl2 orccl8 but no the neutrophil-recruiting chemokine gene il8/cxcl8 and in contrast to the robust activation of MMPs mediated by TNFa, only mmp1 was partially expressed upon stimulation by IL6/sIL6R trans-signaling at the higher dose tested (Fig. 1b). The magnitude of gene inductions by IL6/sIL6R was 2-100 times lower than by TNFa.
Overall, IL6 trans-signaling mediates effects partially overlapping to those of TNFα in SF, supporting the coordinated expression of cytokines, chemokines and matrix metalloproteases central to RA pathophysiology.
IL6/sIL6R trans-signal modulates the TNFa-induced inflammatory response of SF
To investigate IL6 trans-signaling regulation of the inflammatory response in SF, we stimulated SF cultures with different suboptimal doses of TNFα plus a fixed dose of IL6 and sIL6R (50ng/ml) according to dose-response assays of representative genes regulated by each factor (Fig. 1). Cooperative stimulation of SF with both TNFα and IL6/sIL6R enhanced the expression of common regulated genes such as il6,ccl8 and mmp1, although in the present dose-response analysis only a clear increase trend could be observed for CCL8 and MMP1. Further mRNA and protein expression analysis along the present study demonstrated the consistency of the increase shown for those genes. In contrast, mmp3, a gene specifically activated by TNFα but not IL6/sIL6R, was not affected by IL6 trans-signaling (Fig. 2a). IL6 has been shown to regulate the expression pattern of chemokines on stromal cells to drive the transition from the recruitment of neutrophils to mononuclear cells. Consistently, TNFα-induced mRNA and protein expression of the neutrophil-recruiting chemokine IL8/CXCL8 was partially inhibited by the trans-signal activation of IL6/sIL6R in SF, whereas TNFα and IL6/sIL6R cooperated to up-regulate the protein expression of mononuclear leukocytes chemoattractant chemokine CCL8 (Fig. 2a and 2b). Interestingly, IL6/sIL6R also inhibited the TNFα-induced expression of mmp10, an enzyme linked to the resolution of inflammation by macrophages (Fig. 2a).
To analyze the functional implications of the switch in the pattern of chemokines after the cooperative stimulation with TNFa and IL6/sIL6R, we performed a cell migration experiment using conditioned media from SF cultures as chemoattractant for leukocytes. Conditioned media from SF cultures treated with both TNFa and IL6/sIL6R reduced significantly the percentage of polymorphonuclear (PMN) cells in comparison with medium from TNFa-stimulated cultures (Fig. 2c). Likewise, TNFα plus IL6/sIL6R conditioned media induced a significant increase in the recruitment of mononuclear cells (MNC) (Fig. 2c).
IL6/sIL6R regulates the kinetics of the TNFa-inflammatory response
Expression of genes activated by continuous exposure to TNFα is determined by transcriptional and post-transcriptional mechanisms that regulate its level and kinetics[20-22]. To distinguish the potential regulation of these mechanisms upon induction with IL6/sIL6R, we first set the temporal pattern of induction for analyzed genes.
The kinetics of genes stimulated by TNFa in SF mostly fit into the three broad classes, as previously described in other cell types (Additional file 2: Figure S1a). Thus, the expression of an early gene (il6) was consistently detected at 0.5h upon induction, while intermediate genes (ccl2, il8/cxcl8) are observed before 2h and late expression genes (ccl8, mmp1) later than 2h. The expression of genes mediated by continuous exposure of IL6/sIL6R fit into a similar pattern of induction, although some of the genes that are common to both factors fall into a different category. The induction of il6 and mmp1 follows identical kinetics for both inflammatory factors. In contrast, the expression of ccl8 or ccl2 induced by IL6/sIL6R showed faster kinetics than that mediated by TNFa, showing ccl2 a less stable induction (Additional file 2: Figure S1b). These results more likely reflect differences in the underlying regulatory mechanisms induced by either inflammatory cytokine.
The expression kinetics of genes co-stimulated with TNFα and IL6/sIL6R may provide information about the molecular mechanisms operating in the cooperative induction of genes. For all TNFa-induced genes, kinetics was maintained after co-stimulation with IL6/sIL6R, but differences were observed in the time of the cooperative effect (Fig. 3). The increase of il6 and the decrease of il8/cxcl8 expression by IL6/sIL6R was detectable as soon as 30 minutes upon induction, more likely showing changes in transcription and/or mRNA stability mechanisms. However, increased expression of intermediate and late expression genes such as ccl2, ccl8 or mmp1 occurs at later time, suggesting that secondary factors may underlie the cooperative expression of these genes (Fig. 3a). A similar pattern of expression was obtained when we analyzed the protein released to the culture medium (Fig. 3b). Although not statistically significant, a moderate IL8/CXCL8 inhibition was detectable at 6h after induction, while enhanced expression of CCL8 was only detectable later at 24h (Fig. 3b).
IL6/sIL6R modulates the inflammatory expression profile through de novo transcriptional mechanisms
Accumulation of mRNA may be influenced by ongoing transcription or mRNA stability. To investigate the relative contribution of mRNA stability to the induction of genes mediated by TNFα or IL6/sIL6R in SF, we measured changes in mRNA expression over time after blocking transcription with actinomycin D (ActD). We determined mRNA expression after 24h of induction with either TNFα or IL6/sIL6R, relative to the baseline value before the addition of ActD. Our results showed that all tested genes stimulated by IL6/sIL6R responded similarly, with half-lives of mRNA transcripts decay varying from 0.8 to 2h (Fig. 4a). In contrast, mRNAs induced by TNFα were on average more stable, lasting more than 2h for most genes. The half-life of decay for il6 mRNA stimulated by TNFα was shorter than for the rest of the genes, but similar to that stimulated by IL6/sIL6R (0.8 to 2h) (Fig. 4b). We could not determine the half-life of decay for mmp3, since we found no consistent decrease in stability up to 4h after treatment with ActD. These results suggest that, while both transcriptional and post-transcriptional mechanisms are involved in the modulation of TNFa induced genes, the short half-life of IL6/sIL6R induced mRNAs may reflect a dominant role for de novo transcription. We also observed that co-stimulation of SF with both TNFα and IL6/sIL6R did not significantly modify the half-life of analyzed genes (Fig. 4c), suggesting that regulation of the mRNA stability of TNFa-induced genes is not affected by the modulation after trans-signal activation of IL6/sIL6R.
Recent reports have also demonstrated that a prolonged TNFα exposure for longer than 24h promotes the stability of mRNA expression in fibroblasts by epigenetic mechanisms, influencing the temporal order of induction of inflammatory genes[20, 22, 23]. To distinguish the potential role of these priming mechanisms, we first examined gene expression changes in response to TNFα withdrawal. We cultured SF in the presence of TNFα for 24h, removed the inflammatory input by washing the cells, and added new medium with adalimumab (ADA) to block residual TNFα and with IL6/sIL6R for additional 24h (Fig. 5). SF pre-exposed to TNFα and treated with ADA did not display enhanced induction of early (il6, ccl2 and IL8/CXCL8) nor late genes (mmp1 and MMP3) after IL6/sIL6R treatment, showing expression levels similar to the induced with only IL6/sIL6R (Fig. 5). We also observed that mRNA expression of MMP3 was more resistant to TNFα withdrawal probably due to its stability. These data suggest that the observed cooperative effect requires concomitant induction by both TNFα and IL6/sIL6R.
These results collectively support the view that, although regulation of the mRNA stability and priming mechanisms may determine the kinetics of TNFa-induced gene expression, cooperative induction by TNFa and IL6/sIL6R is more likely mediated through coordinated de novo transcriptional mechanisms.
Late crosstalk between TNFα and IL6/sIL6R is mediated by activation of the JAK-STAT pathway
The delayed effect on the cooperative expression of genes such as ccl8, ccl2 or mmp1 (Fig. 3a) suggested that either TNFα or IL6/sIL6R induces secondary mechanisms dependent of new protein synthesis. This hypothesis was tested by inhibiting the protein synthesis with cycloheximide (CHX). CHX inhibited the expression of late genes such as ccl8 and mmp1 induced by either TNFα (Additional file 3: Figure S2a) or IL6/sIL6R (Additional file 3: Figure S2b), suggesting that protein-synthesis-dependent pathways are partly involved in the expression of these genes in SF, in contrast to early or intermediate genes such as il6, il8/cxcl8 and ccl2.
Previous investigations have demonstrated that TNFα stimulation of SF induces the expression of several lymphocyte-attracting chemokines through a JAK signaling-mediated mechanism, dependent on autocrine release of type I Interferons (IFN). To confirm this possibility in our model, we analyzed the RSAD2 mRNA expression, a classical IFN-induced gene, after TNFa treatment. We observed a high induction of rsad2 that was completely inhibited in the presence of the JAK/STAT inhibitor ruxolitinib (RUXO) (Fig. 6a). As expected, induction of IL6/sIL6R dependent genes was abrogated by RUXO treatment (Fig. 6b). Further analyses demonstrated that RUXO significantly inhibited TNFa-induced expression of ccl2 and ccl8, implying that the secondary mediator acts through JAK/STAT. Interestingly, mmp1, other late gene regulated by CHX was not affected by RUXO, suggesting that a JAK/STAT-independent but protein-synthesis-dependent pathways is partly involved in mmp1 expression.
These experiments further revealed that a TNFα-induced autocrine mechanism is regulating part of the TNFα expression program in SF. This autocrine mechanism induced by TNFα shares a common JAK/STAT signaling pathway with IL6/sIL6R that may partially account for the cooperative expression of specific genes.