PGE2 is involved in numerous inflammatory associated diseases including the chronic inflammatory disease periodontitis. The aim of this study was to investigate the global gene expression profile of TNFα-stimulated human gingival fibroblasts with a focus on signal transduction pathways related to the expression of mPGES-1 and COX-2. We used a time-course factorial design for the oligonucleotide microarray hybridizations, which enabled us to statistically discern the interaction effect between the TNFα treatment and time aspect in the experiments studied. To our knowledge, this is the first study investigating the global gene expression profile of TNFα-stimulated gingival cells as an in vitro model of gingival inflammation. In the current study, we demonstrate a TNFα-stimulated up-regulation of mPGES-1 and COX-2 expression, both at mRNA and protein levels, accompanied by enhanced PGE2 production. In addition, our microarray results showed a regulation of immune response, apoptosis and signal transduction gene categories in response to TNFα treatment of gingival fibroblasts, including positive regulation of the signal pathways JNK and NF-κB.
The TNFα up-regulated mPGES-1 and COX-2 expression, accompanied by increased PGE2 production, is in line with our previously published results on gingival fibroblasts . In contrast to the mPGES-1 and COX-2 enzymes, the expression of the PGE synthases mPGES-2 and cPGES was not affected by TNFα treatment, highlighting the importance of mPGES-1 and COX-2 in the regulation of inflammatory-induced PGE2 production. However, when considering the similar kinetics of PGE2 production and COX-2 expression, as well as the magnitude of COX-2 induction compared to mPGES-1, COX-2 seems to be the more important enzyme driving the TNFα induced PGE2 production in gingival fibroblasts. Increased PGE2 production, via induction of the PGE2-synthesizing enzymes mPGES-1 or COX-2, has also been reported in other cell types stimulated with TNFα, including synovial cells, chondrocytes and colonocytes [38, 50, 51].
Various intracellular signaling pathways have been reported to be involved in inflammatory-induced PGE2 synthesis, mainly through regulation of COX-2, which is the most widely studied enzyme of the PGE2 synthesis chain. In contrast, limited reports exist regarding the regulation of the PGE synthases downstream of COX-2. To further explore the regulation of mPGES-1 and related PGE2-synthesizing enzymes in gingival fibroblasts, we here used a global gene expression profiling approach to achieve a broader view of the genes and signal pathways related to the regulation of mPGES-1, in parallel with COX-2, using microarray analysis of TNFα-treated gingival fibroblasts. The effect of TNFα on global gene expression profiles has previously been investigated in synovial fibroblasts, preosteoblasts and HeLa cells, but not in gingival cells [52–56]. With regard to periodontal disease, microarray analysis of gingival tissue has been used in an attempt to define subclasses of periodontitis and to evaluate the effect of periodontal therapy [57, 58]. In addition, blood cell gene expression profiling has been performed in subjects with aggressive periodontitis . Concerning gingival fibroblasts, microarray studies have been performed on unstimulated cells from healthy and inflamed gingival tissue, and on IL-1β-stimulated immortalized cells [60, 61]. However, a majority of the abovementioned microarray studies on cell cultures use only one time point of RNA analysis after stimulation. In this work, using primary gingival fibroblasts, we employed a time-course factorial design to extract as much relevant information as possible from our data set. Commonly, when using a simpler design for a time series experiment, it becomes difficult to relate the gene expression differences at the different time points to each other. This is greatly facilitated by the use of a time-course factorial design, which allows for the identification of the genes that are differentially regulated between the time points due to TNFα treatment .
Our results from the microarray analysis indicated the JNK and NF-κB pathways as possible targets for interrupting the TNFα-induced signal transduction leading to increased expression of the PGE2-synthesizing enzymes mPGES-1 and COX-2. Thus, we next further investigated the involvement of JNK and NF-κB in the TNFα-stimulated mPGES-1 and COX-2 expression. By using inhibitors specific for these signal pathways, we could demonstrate that JNK and NF-κB are partly involved in the complex network of intracellular signal transduction pathways leading to increased expression of mPGES-1 and COX-2, as well as PGE2 production, in response to the inflammatory cytokine TNFα. To our knowledge, this is the first study pointing out the involvement of JNK in up-regulation of mPGES-1 expression in TNFα-stimulated cells. However, previous studies have shown that mPGES-1 is stimulated by IL-1β through the JNK pathway in cardiac myocytes, cardiac fibroblasts and A549 human lung epithelial cells [36, 62, 63]. In this study we also show that JNK is involved in TNFα-induced COX-2 expression in gingival fibroblasts, suggesting that JNK-dependent decrease of mPGES-1 may not be the major event whereby the inhibition of this signal pathway exerts its effect on PGE2 production. The finding that COX-2 expression was somewhat more influenced by inhibition of the JNK pathway, suggests that the JNK pathway may be more significant for COX-2 induction by TNFα in gingival fibroblasts. The involvement of JNK in TNFα-induced COX-2 expression is in line with results obtained from human alveolar epithelial cells and murine osteoblasts [64, 65]. Similarly, IL-1β- and lipopolysaccharide-induced COX-2 has been shown to be mediated through JNK in other cell types [66, 67]. In contrast, it has also been reported that COX-2 expression induced by epidermal growth factor is unaffected by JNK inhibition in astrocytes, emphasizing the differences between cell types as well as the inflammatory stimuli used for investigation of signal transduction pathways .
The JNK pathway has been implicated in chronic inflammatory disorders such as rheumatoid arthritis and inflammatory bowel disease [69, 70]. Moreover, JNK-1 deficiency as well as chemical JNK inhibition has been demonstrated to prevent joint destruction in rodent models of rheumatoid arthritis [69, 71, 72]. Thus, our novel finding that JNK is partly involved in the regulation of TNFα-induced mPGES-1 expression, which in concert with COX-2 regulates PGE2 production in gingival fibroblasts, indicates that this signal pathway may also be of importance in the pathogenesis of periodontitis.
In addition to the JNK pathway, we also found that the NF-κB pathway was involved in the regulation of mPGES-1 and COX-2 in gingival fibroblasts stimulated with TNFα. Our results pointing out the NF-κB pathway in the regulation of mPGES-1 are in line with our previous observations . Induction of mPGES-1 expression using another inflammatory mediator, IL-1β, has been reported to be mediated by NF-κB in A549 cells . In regard to COX-2 regulation, the involvement of the NF-κB pathway in the signal transduction of TNFα-induced COX-2 expression has been observed in other cell types, including endothelial cells and astrocytes [74, 75]. Concerning PGE2 production, no decrease was observed in TNFα-stimulated PGE2 production in 24 h cultures, in contrast to the inhibition observed in 6 h cultures. One reason for the lack of inhibition of PGE2 in 24 h cultures by the NF-κB inhibitor Bay, in contrast to mPGES-1 and COX-2 expression, may be a toxic effect of the inhibitor, affecting the release of PGE2 although no visual signs of cellular toxicity were observed. However, when using the NF-κB inhibitor Ro, reported to inhibit NF-κB via selective inhibition of TNFα-induced IκBα , it decreased the TNFα-stimulated expression of mPGES-1 and COX-2 as well as the production of PGE2 in 24 h cultures. Furthermore, both the NF-κB inhibitors Bay and Ro as well as the JNK inhibitor SP decreased the basal expression of mPGES-1 in cells not treated with TNFα, which might be due to slightly raised basal levels of mPGES-1 expression resulting from a lingering effect of the serum present in growth medium before the start of cell culture experiments.
A time-course factorial microarray analysis, like the one performed in this work, yields massive amounts of data. In this study we have focused on the signal transduction aspect, especially the JNK and NF-κB pathways, in order to investigate the regulation of mPGES-1 and COX-2 expression in relation to PGE2 production. Inhibition of JNK and NF-κB signal pathways by SP and Ro abolished the production of PGE2, although the induction of mPGES-1 and COX-2 by TNFα was not completely abrogated. One explanation for this might be that other enzymes may contribute to the production of PGE2 stimulated by TNFα. For instance, it is known that the signal pathways JNK and NF-κB, in addition to the COX-2 and mPGES-1 enzymes, are also involved in the regulation of cPLA2, an upstream key enzyme of the PGE2 synthesis reported to be induced by TNFα in gingival fibroblasts [76–78]. Another explanation for the strong inhibition of PGE2 production, in contrast to the partial reduction of mPGES-1 and COX-2 expression, may be a synergistic effect of the concerted inhibition of these two enzymes, since they are functionally coupled and responsible for the coordinated PGE2 synthesis . In addition, one has to be aware that the JNK and NF-κB pathways activated by TNFα may not be entirely responsible for the increased expression of mPGES-1 and COX-2. There are many additional TNFα-regulated genes and pathways involved in the regulation of inflammatory conditions, including PGE2 regulatory enzymes, that merit further study, and investigations are ongoing to continue charting the genome-wide effect of TNFα on gingival fibroblasts.