The activity of p38 SAPK is important for controlling many aspects of cell physiology and its role in gene expression regulation has been well established. For instance, the Saccharomyces cerevisae p38 SAPK orthologue Hog1 plays a key role regulating osmostress-induced genes required for yeast adaptation to osmotic shock . In order to shed light on the role of p38 SAPK in genome-wide transcriptional regulation in response to stress, we have performed whole genome microarray analyses and compared three different treatments (TNFα, anisomycin and NaCl) that are able to activate p38 SAPK, albeit with different kinetics. The three p38 SAPK activators induced the up-regulation of a specific set of genes, with a small core set of genes being induced by all of them.
In general, genes induced by the three stimuli were found to be involved in the regulation of different biological processes. As expected, TNFα elicited immune and stress responses. On the other hand, anisomycin specifically induced genes involved in transcription repression and chemokine activity suggesting that this antibiotic partially mimics TNFα, which strongly induced several chemokines and cytokines. Both NaCl and anisomycin had in common the regulation of RNA biogenesis whereas NaCl specifically enhanced many components of the ribonucleoprotein complex, suggesting that protein translation is a key process for cell adaptation to osmostress. On top of that, osmostress also up-regulated DNA associated proteins, such as histones and proteins associated to the nucleolus and the nuclear pore complex. Notably, the three stimuli regulated to some extent genes involved in cell migration, proliferation and cell death. However, the major GO molecular function induced by p38 SAPK activation was the up-regulation of transcription factors.
A significant amount of up-regulated genes was dependent on p38 SAPK, based on their diminished expression in the presence of SB203580 and in p38α-/- cells. This raises the interesting question of how SAPK activity is able to specifically up-regulate different gene programs depending on the stimuli. A plausible explanation would be that the final output results from the integration at the gene promoters of various converging signals. Thus, the presence of additional activating signals would be necessary to achieve the highest transcriptional up-regulation in a given context, while inhibition of p38 SAPK would cut one of the inputs resulting in a reduced expression. Alternatively, but less likely, the different kinetics observed in the activation of p38 SAPK might lead to the activation of a different set of genes.
It is well established as well, that p38 SAPK controls the stability of several mRNAs through a not fully understood mechanism that requires the activity of MK2 and the presence of AU-rich elements (AREs) located at the 3' UTR of specific mRNAs. These sequences are involved in the recruitment of ARE-binding proteins that contribute to mRNA stabilisation or degradation . There are about 950 human mRNAs containing ARE sequences. A high-throughput study from Frevel and co-workers showed that the stability of up to 42 mRNAs bearing ARE sequences were enhanced upon the treatment of the human monocytic cell line THP-1 with lipolysaccharide in a p38 SAPK dependent manner . We have crossed Frevel's results with ours and found the overlapping of up to 12 different genes (e.g. Tnfaip3, Gro1, Gro2, Gch1, PTGS2/COX2, Ccl2, Cxcl10, Junb, Csf1, Irf1, Pmaip1 and Cited2). Notably, most of these genes where up-regulated by TNFα which suggests that different p38 SAPK activators may contribute differently to mRNA stabilization. Moreover, we have seen that osmostress strongly up-regulated the ARE binding protein TPP which has been shown to suppress inflammation by accelerating the decay of cytokine mRNAs upon p38 SAPK activation . Therefore further experimental work is necessary to asses the contribution of transcription and mRNA stabilization in response to stress conditions. In any case, p38 SAPK appears to be instrumental for the proper response of a majority of genes in response to all three stimuli tested.
In yeast, there is a large conserved environmental stress response (ESR), which consists in a core of genes that respond to a diverse type of stress; such as heat shock, osmostress, oxidative stress and others [21, 22]. This core of genes includes a number of stress defence genes and seem to be important for adaptation to stress. In mammals, such an ESR has not been reported. Interestingly, we found a small number of genes that is common to all three stress treatments and might be relevant. Notably, the main enriched molecular function induced by the three treatments was the up-regulation of transcriptional modulators. Thus, transcriptional regulation could be a key shared process to build up the appropriate gene response for adaptation to cell stress. Moreover, the three treatments up-regulated several members of the DUSP family of genes, which have MAPK phosphatase activity and are known to play a key role in the down-regulation of MAPK signalling [23, 24].
It is worth noting that when p38 SAPK dependence was analysed in a time course experiment, more than 90% of the genes that responded to osmostress at 45 minutes were p38 SAPK-dependent, however, after 8 h only 60% of the up-regulated genes required p38 SAPK activity. Thus, the SAPK pathway appears to be critical at the initial response, whereas other signaling pathways are probably more relevant at later times.
Regarding how cells adapt to stress over time, our GO analyses revealed that the primary response to NaCl addition was the up-regulation of solute and ion trans-membrane transporters as well as sensorial G-protein coupled receptors. Significantly, 65% of the early genes (71 out of 114 up-regulated at 45 minutes) were shut down at later times. At the intermediate time point, 130 genes were specifically up-regulated and were mainly involved in the control of transcription and ribosome biogenesis. However after 8 h of osmostress, 279 genes were specifically up-regulated. These late-induced genes were mainly included in the GO terms transcription factor activity, chemokine and cytokine activity, protein kinase activity, stress response, growth control and protein dephosphorylation. Altogether, we conclude that on one hand, osmostress up-regulates many ion trans-membrane transporters to quickly overcome the changes in osmolarity between the extra-cellular environment and the intracellular milieu, whereas on the other hand, long exposure to osmostress promotes cell survival, growth and mediates a TNFα-like response through the up-regulation of chemokine/cytokine activities. On top of that, protein dephosphorylation is probably necessary to down-regulate p38 SAPK activity and to shut down other signalling pathways activated at early times. Notably, gene transcription is still necessary for cell adaptation at this late time point.
To date, few whole genome studies on p38 SAPK regulated genes have been reported. An early work used primary endothelial cells treated with TNFα in the presence of SB203580 for 5 h to identify TNFα-induced and p38 SAPK dependent genes. These authors reported 58 TNFα-induced genes, 12 of which depended on p38 SAPK to some extent . In contrast, we have found that just after 45 minutes of incubation with TNFα there is a strong dependency on p38 SAPK for gene up-regulation. This difference would be consistent with our results showing that SAPK dependency decreases over time upon osmostress treatment. Therefore, the genes up-regulated 5 h after cell stimulation observed by Vienmman and co-workers may represent a late response to the cytokine, which would be driven by secondary signalling events rather than by TNFα itself. In agreement with this idea, our data shows that TNFα strongly induced the up-regulation of chemokines and cytokines that once secreted to the media may trigger a paracrine cell response leading to the activation of other signalling pathways and gene transcription programs. Similarly a more recent study reported a genome wide screening using rat fibroblast-like synoviocytes also treated with TNFα. These authors reported the up-regulation of 141 genes by this cytokine, 30% of which were dependent on the activation of p38 SAPK . Again the low p38 SAPK dependency reported in this work is not surprising if we take into account that cells were harvested 24 h after the addition of TNFα. We have crossed-checked their TNFα up-regulated list of genes with ours and found that only six genes overlapped: Cxcl2, Gm190, Csf1, Pde4b, Phlda1 and Ier3. As discussed above, the most plausible explanation for the discrepancy with our study is probably due to the fact that we have evaluated the primary TNFα response on immediate early genes whereas Zer et al reported a late TNFα-mediated gene response. On another hand, a wide genome microarray analysis performed on exponentially growing human keratinocytes treated with SB203580 but without challenging the cells with extra-cellular stimuli has been reported. This study concluded that p38 SAPK positively controlled genes involved in cell proliferation, keratinocyte differentiation and MAPK regulation . In agreement with our results, the authors also found the up-regulation of some DUSP genes, indicating that this is likely to be an important event to down-regulate MAPK activity. Moreover Gazel's work and ours concludes that the transcription factors c-fos and n-Myc are key players of stress-mediated responses. However, the overall gene overlapping between the set of genes that we have found and that of Gazel and co-workers is negligible which might be due to the fact that in such study the authors only evaluated gene regulation mediated by the p38 SAPK basal activity. Our conclusions based on transient activation of the p38 SAPK also differ from a report using proliferating cardiomyocytes which showed that p38 SAPK basal activity played an important role in regulating extra-cellular matrix genes through the transcription factors C/EBPβ and TEF-1 .