MSK1 (mitogen- and stress-activated kinase 1) and MSK2 are nuclear protein kinases that regulate transcription downstream of the ERK1/2 (extracellular-signal-regulated kinase 1/2) and p38α MAPKs (mitogen-activated protein kinases) via the phosphorylation of CREB (cAMP-response-element-binding protein) and histone H3. Previous studies on the function of MSKs have used two inhibitors, H89 and Ro 31-8220, both of which have multiple off-target effects. In the present study, we report the characterization of the in vitro and cellular properties of an improved MSK1 inhibitor, SB-747651A. In vitro, SB-747651A inhibits MSK1 with an IC50 value of 11 nM. Screening of an in vitro panel of 117 protein kinases revealed that, at 1 μM, SB-747651A inhibited four other kinases, PRK2 (double-stranded-RNA-dependent protein kinase 2), RSK1 (ribosomal S6 kinase 1), p70S6K (S6K is S6 kinase) (p70RSK) and ROCK-II (Rho-associated protein kinase 2), with a similar potency to MSK1. In cells, SB-747651A fully inhibited MSK activity at 5–10 μM. SB-747651A was found to inhibit the production of the anti-inflammatory cytokine IL-10 (interleukin-10) in wild-type, but not MSK1/2-knockout, macrophages following LPS (lipopolysaccharide) stimulation. Both SB-747651A and MSK1/2 knockout resulted in elevated pro-inflammatory cytokine production by macrophages in response to LPS. Comparison of the effects of SB-747651A, both in vitro and in cells, demonstrated that SB-747651A exhibited improved selectivity over H89 and Ro 31-8220 and therefore represents a useful tool to study MSK function in cells.
- cAMP-response-element-binding protein (CREB)
- mitogen-activated protein kinase (MAPK)
- mitogen- and stress-activated kinase (MSK)
- ribosomal S6 kinase (RSK)
MAPK (mitogen-activated protein kinase) signalling cascades are important mediators of the cellular response to a wide range of stimuli, including mitogens, growth factors, cytokines and cellular stress [1–3]. How these cascades produce the correct response in cells has been the subject of extensive study. Small-molecule inhibitors of the ERK1/2 (extracellular-signal-regulated kinase 1/2) cascade or p38α/β have been used extensively as a means to probe the functions of these pathways, and have played a crucial role in the advancement of this area. The impetus for the continued development of these compounds has come from the demonstration that ERK1/2 and p38 play key roles in cancer and autoimmunity, and MAPK pathway inhibitors have also shown efficacy in models of cancer and inflammation [3–6].
Both ERK1/2 and p38α are known to have multiple substrates, including several downstream kinases . The ERK1/2 cascade has been shown to be essential for the activation of the RSK (ribosomal S6 kinase) family of kinases, whereas p38α MAPK activates the kinases MAPKAPK2 [MAPKAP (MAPK-activated protein) kinase 2] and MAPKAPK3. In contrast with RSK and MAPKAPK2, which are activated downstream of a single MAPK cascade, two further groups of kinases, MNKs (MAPK-interacting kinase) and MSKs (mitogen- and stress-activated kinases) can be activated by either ERK1/2 or p38α pathways [8,9]. These downstream kinases probably play key roles in mediating the physiological functions of ERK1/2 and p38 and also represent potential drug targets.
Two isoforms of MSK, termed MSK1 and MSK2, have been identified in mammalian cells. MSKs are most closely related to the RSK family of kinases and, similar to RSK, they contain two kinase domains in a single polypeptide. The use of cell-permeant inhibitors that block MAPK signalling has demonstrated a role for both ERK1/2 and p38α in the activation of MSKs [10–12]. For stimuli such as PMA or EGF (epidermal growth factor), which predominantly activate ERK1/2 but not p38α, MSK activation is blocked by pre-treatment of the cells with inhibitors of MKK1 (MAPK kinase 1) and MKK2 that prevent the activation of ERK1/2. MSK activation by stimuli such as anisomycin that predominantly activate p38, but not ERK1/2, is blocked by pre-incubation with the p38α/β inhibitor SB-203580, whereas, for stimuli that activate both ERK1/2 and p38, such as TNF (tumour necrosis factor) or LPS (lipopolysaccharide), a combination of MEK1/2 (MAPK/ERK kinase 1/2) and p38 inhibitors are required to completely block MSK activation [12,13]. The p38 inhibitor SB-203580 targets both p38α and p38β; however, it has been shown using mouse knockouts that fibroblasts lacking p38β activate MSK1 normally, whereas, in cells lacking p38α, p38-dependent MSK1 activation was abolished, indicating that p38α and not p38β is the major p38 isoform involved in MSK activation [14,15].
Mice lacking MSK1 or MSK2, and also a double knockout of both MSK1 and MSK2, are viable and fertile, but show enhanced inflammation in immune models as well as impairments in some models of memory [11,16,17]. Using cells from MSK-knockout mice, it has been shown that MSK1/2 is critical for the mitogen- and stress-activated phosphorylation of the transcription factors CREB (cAMP-response-element-binding protein) and ATF1 (activating transcription factor 1) as well as the chromatin proteins histone H3 and HMG14 (high-mobility group protein 14) in a variety of cell types, including fibroblasts, macrophages and cortical neurons [11,18–20]. In addition, MSKs have been suggested to phosphorylate RARα (retinoic acid receptor α)  and p65/RelA . This suggests a role for MSKs in the regulation of immediate-early gene transcription. In line with this, the induction of several immediate-early genes, including c-fos, JunB, nur77, IL-1ra [IL-1 (interleukin-1) receptor antagonist] and miR-132 (microRNA 132), in response to mitogens or cellular stress is reduced in cells from the MSK1/2-knockout mice [11,15,23–25].
Although the use of mouse germline manipulation has been a useful way of investigating MSK function, a selective cell-permeant inhibitor of MSKs would be of great help in elucidating MSK signalling mechanisms and function. Until recently, however, selective small-molecule inhibitors of the downstream kinases RSK, MSK and MAPKAPK2 have not been available. Ro 31-8220 and H89 inhibit RSK and MSK, but these compounds are non-selective and inhibit many other protein kinases [26,27], whereas H89 has also been shown to target proteins other than kinases . In addition, they can also affect the activation of the ERK1/2 MAPK pathway by some stimuli, therefore limiting their utility to study MSK or RSK function [27,29]. Despite this, owing to the lack of alternatives, H89 and Ro 31-8220 have been used to study MSK function in cells [12,17,22,27,30,31]. There have been several reports on the identification of more selective inhibitors for these kinases. Three reports have described the identification of novel inhibitors of RSK that block its activity both in vitro and in cells [32–34], and two recent papers have described the first MAPKAPK2 inhibitors [35,36]. With the exception of the MAPKAPK2 compounds reported by Revesz et al.  for which data are not yet available, profiling of these inhibitors has demonstrated that they are selective for RSK or MAPKAPK2 over MSKs [26,34,36]. (1H-Imidazo[4,5-c]pyridin-2-yl)-1,2,5-oxadiazol-3-ylamine derivatives have also been reported to act as potent inhibitors of MSK1 activity in vitro, with up to 300-fold selectivity for MSK1 over RSK [37,38]. However, the properties of such compounds in cells have not been reported. In the present study, we characterize the in vitro and cellular properties of SB-747651A, a selective and cell-active inhibitor of MSKs with properties superior to H89 and Ro 31-8220 for analysis of MSK signalling pathways.
MATERIALS AND METHODS
SB-747651A was synthesized by GlaxoSmithKline as described previously [37,38]. Requests for SB-747651A should be addressed to Alastair.firstname.lastname@example.org. H89 and Ro 31-8220 were obtained from Calbiochem.
Kinase inhibitor specificity profiling
Kinase selectivity profiling for MSK1 inhibitors was carried out as described previously [26,27,39] (http://www.kinase-screen.mrc.ac.uk/). Briefly, protein kinase assays were carried out at room temperature (21°C) and were linear with respect to time and enzyme concentrations under the conditions used. Assays were performed for 40 min using a Biomek 2000 Laboratory Automation Workstation in a 96-well format (Beckman Instruments). The concentration of magnesium acetate in the assays was 10 mM, and the concentration of [γ-33P]ATP (800 c.p.m./pmol) used was selected to be close to the kinase's Km value for ATP (see Table 2). Assays were initiated with MgATP and stopped by the addition of 5 μl of 0.5 M orthophosphoric acid. Aliquots (30 μl) were then spotted on to P30 filtermats, washed four times in 75 mM phosphoric acid to remove ATP and once in methanol, then dried and counted for radioactivity. To determine IC50 values, kinase activities were determined at ten inhibitor concentrations ranging from 3 nM to 100 μM as indicated.
Primary MEFs (murine embryonic fibroblasts) were prepared as described previously . MEF, HEK (human embryonic kidney)-293 and HeLa cells were maintained in DMEM (Dulbecco's modified Eagle's medium) supplemented with 10% FBS (fetal bovine serum), 2 mM L-glutamine, 50 units/ml penicillin G and 50 mg/ml streptomycin (Invitrogen). HEK-293 cells were transfected using a modified calcium phosphate protocol as described previously . Before stimulation, MEF, HEK-293 and HeLa cells were serum-starved for 16 h. BMDMs (bone-marrow-derived macrophages) were isolated as described previously  and cultured in DMEM supplemented with 10% FBS, 5 ng/ml CSF (colony-stimulating factor), 2 mM L-glutamine, 50 units/ml penicillin G and 50 μg/ml streptomycin (Invitrogen). Animals were maintained in accordance with EU and U.K. regulations, and work was carried out under a U.K. Home Office Project Licence and subject to local ethical review.
Samples were run on 4–12% polyacrylamide gels (Novex, Invitrogen) and transferred on to nitrocellulose membranes. Antibodies that recognize total ERK1/2, phospho-ERK1/2 (which recognizes phospho-Thr202/Tyr204 of ERK1 or phospho-Thr185/Tyr187 of ERK2), phospho-Thr308 PKB (protein kinase B), phospho-Ser473 PKB, phospho-GSK3 (glycogen synthase kinase 3) (phospho-Ser21 GSK3α, phospho-Ser9 GSK3β), phospho-Thr359 RSK, total p38α, phospho-Thr180/Tyr182 p38α, phospho-Thr334 MAPKAPK2, phospho-Thr180/Tyr182 JNK (c-Jun N-terminal kinase), phospho-Thr389 p70S6K (S6K is S6 kinase, also called RSK), phospho-Ser235/Ser236 S6K, phospho-Ser133 CREB (also recognizes phospho-Ser63 in ATF1) and phospho-Ser360, phopsho-Ser376 or phospho-Thr581 MSK1 were from Cell Signaling Technology. Antibodies against phospho-Ser212 MSK1, Ser750/Ser752 MSK1, Nur77 and phospho-Ser354 Nur77 have been described previously [10,29]. Residue numbers are for human proteins. HRP (horseradish peroxidase)-conjugated secondary antibodies were from Pierce, and detection was performed using the enhanced chemiluminescence reagent from Amersham Biosciences.
Quantitative RT (reverse transcription)–PCR
Cells were treated as indicated, then lysed and total RNA isolated using the NucleoSpin RNA purification method (Qiagen). RNA was reverse-transcribed (iScript; BioRad Laboratories) and real-time PCR was carried out using SYBR Green-based detection. 18S rRNA levels were used as normalization controls and relative mRNA levels were calculated using the equation: where E is the efficiency of the PCR, ct is the threshold cycle, u is the mRNA of interest, r is the reference gene (18S RNA), s is the sample and c is the unstimulated control sample. The PCR efficiency was determined experimentally, and the identity of the PCR products was confirmed by sequencing. The primers used are listed in Table 1.
The levels of TNF, IL-12p40, IL-12p70 and IL-10 in culture medium were determined using a multiplex assay from BioRad Laboratories using Luminex-based technology. Assays were performed according to the manufacturer's protocols.
SB-747651A inhibits activity of the N-terminal kinase domain of MSK1
The (1H-imidazo[4,5-c]pyridin-2-yl)-1,2,5-oxadiazol-3-ylamine derivative SB-747651A (Figure 1A) has been identified as a potent ATP-competitive inhibitor of MSK1 in vitro [37,38]. We found that SB-747651A inhibited MSK1 with an IC50 value of 11 nM in vitro (Figure 1B). To establish which of the two kinase domains of MSK1 is the target of SB-747651A, we determined the effect of SB-747651A on key MSK1 phosphorylation sites. FLAG–MSK1 was expressed in HEK-293 cells by transient transfection and activated by stimulation in the presence or absence of 5 μM SB-747651A with either UV-C, which activates MSK1 via p38, or PMA, which activates MSK1 via ERK1/2. The activation of p38 in response to UV-C, or the activation of ERK1/2 in response to PMA, was not significantly affected by SB-747651A, as judged by immunoblotting with phospho-specific antibodies against their TXY activation motifs (Figure 1C).
MSKs contain two kinase domains, and their activation mechanism is complex and involves multiple phosphorylation steps [10,40]. Initially, MSK1 is phosphorylated on three sites (Ser360, Thr581 and Thr700) by the upstream kinases ERK1/2 and p38α. This results in the activation of the C-terminal kinase domain, which then autophosphorylates three residues, including the T-loop (Ser212) and hydrophobic motif (Ser376) of the N-terminal domain. This in turn activates the N-terminal domain of MSK1, which is responsible for autophosphorylating three residues at the C-terminus of MSK1 (Ser750, Ser752 and Ser758) as well as phosphorylating MSK substrates [10,40].
To determine which kinase domain SB-747651A targets, the phosphorylation of the key activation residues was examined in the presence of SB-747651A. ERK1/2 and p38 phosphorylate Ser360 and Thr581 in MSK1  and, as expected, the phosphorylation of these sites was induced by UV-C or PMA (Figure 1C). The UV-C-induced phosphorylation of these MAPK sites was not affected by pre-incubation with SB-747651A. However, unexpectedly, the PMA-induced phosphorylation of Thr581, and to a lesser extent that of Ser360, was reduced by pre-incubation with SB-747651A. One explanation of this could be that MSK inhibition promotes dephosphorylation of Thr581, although it is not clear why this effect was much more pronounced with PMA than with UV-C stimulation. Consistent with this, we have found previously that point mutants that inactivate MSK1 decrease the amount of Thr581 phosphorylation observed in cells following PMA stimulation [10,40]. Phosphorylation of Thr581 activates the C-terminal domain of MSK1, which then autophosphorylates Ser376 and Ser212 . The phosphorylation of these sites induced by UV-C or PMA was not greatly affected by pre-incubation with SB-747651A, suggesting that this inhibitor did not target the C-terminal domain of MSK1. The N-terminal kinase domain of MSK1 has been shown to autophosphorylate Ser750 and Ser752 in the C-terminus of MSK1 . The UV-C- and PMA-induced phosphorylation of these sites was inhibited by SB-747651A, consistent with this compound targeting the N-terminal kinase domain of MSK1.
Selectivity profile of SB-747651A
Interpretation of data generated with ATP-competitive kinase inhibitors requires a detailed understanding of the kinase target profile of the compound in question. To this end, the selectivity of SB-747651A was determined against a panel of 117 kinases in vitro. For comparison, the selectivity profiles of H89 and Ro 31-8220, two compounds previously described as MSK inhibitors, were also determined (Figure 2A and Table 2). At 1 μM, H89 showed reasonable selectivity, but limited potency, for MSK1. Additionally, six kinases were more strongly inhibited than MSK1 by H89. Ro 31-8220 showed poor selectivity for MSK1 and targeted many kinases in the screen. These results confirm and extend previously published selectivity data demonstrating the limitations of these compounds as MSK1 inhibitors [26,27]. In addition to MSK1, SB-747651A was also found to inhibit four other AGC kinases PRK2 (double-stranded-RNA-dependent protein kinase 2), ROCK (Rho-associated protein kinase), S6K1 and RSK to a similar degree as MSK1 (Table 2). SB-747651A also showed some activity against PKA (protein kinase A) and PKB as well as inhibiting three non-AGC kinases Pim-1 (provirus integration site for Moloney murine leukaemia virus-1), Pim-3 and MELK (maternal embryonic leucine zipper kinase). IC50 values were determined for SB-747651A against MSK, PKA, RSK, PKB, p70S6K, ROCK and PRK2 at ATP concentrations close to the Km of the kinase for ATP. IC50 values against MSK, RSK, p70S6K, PRK2 and ROCK-II were all found to be in the range 10–100 nM, whereas the IC50 values for PKB and PKA were 190 and 300 nM respectively (Figure 2B). This compares with an IC50 value of 0.05 μM for MSK1. Thus, although SB-747651A is a potent MSK1 inhibitor, it can also inhibit several other AGC kinases to a similar degree in vitro.
SB-747651A inhibits the action of MSKs in cells
The results for inhibitors from in vitro kinase assays do not always translate directly into their properties in cells. This can be due to differences in the ATP concentrations between the two systems and the ability of the inhibitor to cross the cell membrane.
MSK1 and MSK2 are known to be required for the phosphorylation of CREB and ATF1 downstream of ERK1/2 or p38 signalling. To determine whether SB-747651A was able to inhibit MSK1 and MSK2 activities in cells, serum-starved HeLa cells were incubated with concentrations of SB-747651A ranging from 0.1 to 10 μM and then stimulated with PMA, UV-C or anisomycin (Figure 3). PMA was found to activate ERK1/2, but not p38, in HeLa cells, and the activation of ERK1/2 was not affected by pre-incubation of the cells with SB-747651A. PMA treatment also resulted in the phosphorylation of the MSK substrates CREB and ATF1. These phosphorylation events were significantly reduced by pre-incubation of the cells with 10, 5 or 1 μM SB-747651A (Figure 3A). Anisomycin and UV-C were found to strongly activate the p38 and JNK MAPKs in HeLa cells, and the activation of these kinases, or the phosphorylation of the p38 substrate MAPKAPK2, were not affected by SB-747651A. Anisomycin and UV-C also stimulated CREB and ATF1 phosphorylation, and this was reduced to baseline levels by pre-incubation in 10 or 5 μM SB-747651A. Phosphorylation of CREB and ATF1 was also significantly reduced by 1 μM SB-747651A, whereas 0.5 and 0.1 μM were less effective (Figures 3B and 3C). Taken together, these data suggest that SB-747651A is effective at inhibiting the activity of MSKs in cells without affecting the upstream MAPK signalling pathways. Consistent with this, SB-747651A was also able to inhibit PMA-induced CREB phosphorylation in primary MEFs (Figure 4A). To confirm that SB-747651A could also inhibit MSK2 in cells, MEFs were isolated from MSK1-knockout mice, as in these cells the remaining PMA-induced CREB phosphorylation is catalysed predominantly by MSK2 . SB-747651A was also able to inhibit CREB phosphorylation in these cells with a similar IC50 value to that seen in wild-type cells (Figure 4B).
SB-747651A can also inhibit PKB, RSK and p70S6K in cells
As SB-747651A was found to have off-target activity in vitro, including inhibition of PKA, PKB, RSK and p70S6K, the ability to inhibit these kinases in cells was also examined. PKA is activated by forskolin, which is able to elevate cAMP levels in cells. Analysis of the PKA substrate CREB after forskolin stimulation showed that concentrations of SB-747651A above 5 μM resulted in some inhibition of CREB phosphorylation (Figure 5A). This was not due to an effect of SB-747651A on MSK, which also phosphorylates CREB in response to other signals, as forskolin-induced CREB phosphorylation has been previously shown to be unaffected by knockout of MSK1 and MSK2 .
In addition to MSKs, PMA also activates RSK downstream of ERK1/2 in HeLa cells. The activation of RSK was not blocked by SB-747651A, as judged by immunoblotting to detect the phosphorylation of RSK on Thr369. RSK is able to phosphorylate Nur77 on Ser354 . As phosphorylation of endogenous Nur77 is hard to detect with the antibodies currently available, Nur77 was overexpressed in HeLa cells by transient transfection. The phosphorylation of Nur77 in response to PMA was also blocked by concentrations of SB-747651A over 5 μM (Figure 5B). GSK3 is a substrate for both RSK and PKB; however, it has been shown previously that in response to PMA, which is a strong activator of RSK but not of PKB, GSK3 phosphorylation requires RSK but not PKB . PMA stimulation of GSK3 phosphorylation in HeLa cells was blocked by SB-747651A at concentrations of 1 μM or above (Figure 5C). In response to IGF (insulin-like growth factor), GSK3 is phosphorylated by PKB and not RSK, so the effect of SB-747651A on the IGF-stimulated GSK3 phosphorylation in HeLa cells was also examined . Surprisingly, treatment with high concentrations of SB-747651A increased the phosphorylation of PKB on both Thr308 and Ser473, sites that correlate with the activation of PKB. Despite this, SB-747651A was able to inhibit the IGF-stimulated phosphorylation of the PKB substrate GSK3 at concentrations of 5 and 10 μM (Figure 5D).
Both IGF and PMA can activate S6K1 in HeLa cells. SB-747651A was able to greatly reduce the phosphorylation of the S6K substrate ribosomal S6 protein at concentrations down to 0.5 μM, lower than that required to inhibit CREB phosphorylation. At higher concentrations, SB-747651A also inhibited the phosphorylation of S6K itself in response to PMA, and to a lesser extent IGF (Figures 5E and 5F). This is probably due to the ability of SB-747651A to inhibit PKB and RSK, as both of these kinases have been reported to be involved in the activation of S6K [7,42,43].
SB-747651A inhibits MSK-dependent transcription
A major role of MSKs in cells is the transcriptional regulation of specific genes. The effect of SB-747651A on the induction of MSK1/2-dependent immediate-early genes was therefore examined in cells from both wild-type and MSK1/2-knockout mice. Stimulation of primary MEFs with anisomycin has been shown previously to activate MSK via the p38α pathway , and this is required for phosphorylation of CREB and the maximal induction of several immediate-early genes including c-fos, nur77 and Nor-1 [11,15]. Consistent with these previous reports, anisomycin was able to induce the transcription of these genes in wild-type MEFs, and this induction was significantly reduced in MSK1/2-knockout cells. Pre-incubation of the wild-type MEFs with SB-747651A was found to partially inhibit the induction of these genes in wild-type cells, and the remaining levels of mRNA for nur77, Nor-1 and c-fos were similar to those observed in anisomycin-stimulated knockout MEFs (Figure 6). SB-747651A did not have any significant effect on the induction of c-fos and nur77 in anisomycin-stimulated MSK1/2-knockout MEF cells, demonstrating that its effect was specific to MSKs and not the inhibition of an off-target kinase. SB-747651A did not have a general inhibitory effect on all immediate-early genes, as the induction of Egr1 (early growth response 1), a gene which is induced independently of MSK in response to anisomycin, was unaffected by pre-incubation of the cells with SB-747651A.
SB-747651A mimics the effect of MSK1/2 knockout in macrophages
Previous work on MSK1/2-knockout macrophages has shown that MSKs regulate the transcription of several anti-inflammatory cytokines, including IL-10 and IL-1ra [13,23]. As a result, MSK1/2-knockout macrophages overproduce pro-inflammatory cytokines relative to wild-type cells, and MSK1/2-knockout mice are more sensitive to endotoxic shock. Although overall macrophage development appears normal in the MSK1/2-knockout mice, it is possible that the knockout phenotype could be affected by subtle differences in macrophage development rather than acute regulation of cytokine transcription by LPS. Therefore it was of interest to see whether pharmacological inhibition of MSK1/2 would have similar effects. Hence the effects of SB-747651A on LPS-induced transcription were examined. As a further comparison, H89 and Ro 31-8220 were tested in parallel. In wild-type macrophages, LPS induced the transcription of nur77, and this was greatly reduced by pre-incubation of the cells with SB-747651A, Ro 31-8220 or H89. As expected, MSK1/2 knockout resulted in a lower induction of nur77 by LPS than wild-type cells and the remaining induction was not significantly (P>0.05) affected by SB-747651A in the knockout cells. In contrast, H89 resulted in a significant increase in nur77 transcription in MSK1/2-knockout cells, whereas Ro 31-8220 blocked the residual induction of nur77 in MSK1/2-knockout cells (Figure 7A). Similar results were also obtained for c-fos (results not shown). SB-747651A also blocked IL-10 induction in response to LPS in wild-type but not MSK1/2-knockout macrophages, whereas Ro 31-8220 or H89 at 25 μM inhibited IL-10 transcription in both wild-type and MSK1/2-knockout cells (Figure 7B).
SB-747651A also reduced the initial (Figure 7C), but not prolonged (results not shown), transcription of IL-6, which is consistent with the effect of MSK1/2 knockout on IL-6 transcription . In contrast, both H89 and Ro 31-8220 strongly inhibited IL-6 transcription. This was due to inhibition of an enzyme other than MSK1 or MSK2 as it also occurred in MSK1/2-knockout cells (Figure 7C). TNF transcription was not significantly affected by SB-747651A or MSK1/2 knockout; however, similar to IL-6, both H89 and Ro 31-8220 inhibited TNF transcription in both wild-type and MSK1/2-knockout macrophages (Figure 7D).
MSK1/2 knockout has previously been shown to result in decreased IL-10 secretion as well as transcription in macrophages . Consistent with this, pharmacological inhibition of MSKs with SB-747651A also inhibited IL-10 secretion (Figure 8A). MSK1/2 knockout has also been shown to result in increased secretion of TNF, due to a negative regulation of TNF translation by MSKs. SB-747651A increased TNF secretion in wild-type but not MSK1/2 knockouts in response to LPS stimulation (Figure 8B). In contrast, neither H89 nor Ro 31-8220 increased TNF secretion, possibly due to their MSK-independent effects on TNF transcription (Figure 7D)
IL-10 is a potent inhibitor of IL-12 production by macrophages. Previously, it has been shown that knockout of MSK1 and MSK2 results in elevated secretion of IL-12 in response to LPS, and that this is a result of the decreased IL-10 secretion by the MSK1/2 knockouts . SB-747651A increased the secretion of IL-12p40 in wild-type macrophages stimulated by LPS to levels similar to that seen in MSK1/2-knockout cells. SB-747651A also increased IL-12p70 secretion in wild-type cells, but not to the levels seen in MSK1/2-knockout cells. This was probably due to an off-target effect of SB-747651A, as this compound was found to reduce IL-12 secretion in MSK1/2-knockout cells following LPS stimulation. Despite this, SB-747651A more closely mimicked the effect of MSK1/2 knockout on IL-12 production than either H89 or Ro 31-8220 (Figures 8C and 8D).
In the present paper, we describe the characterization of SB-747651A, an inhibitor of MSK1 and MSK2 that represents an improvement in terms of selectivity over previous compounds. The use of small-molecule inhibitors has proven to be a powerful method of dissecting the roles of signalling cascades in cells. However, care should be taken in the interpretation of these experiments since most inhibitors have additional off-target effects in cells. Additionally, the complexity of signalling networks can mean that inhibition of a particular kinase can lead to changes in either upstream or parallel pathways due to the modulation of feedback loops. Extensive characterization of an inhibitor is therefore required before its cellular effects can be properly understood. In the present study, we show that SB-747651A is an effective inhibitor of MSKs in cells and, despite some off-target effects, represents an improvement on the existing alternatives H-89 and Ro 31-8220.
SB-747651A was originally reported as a potent and selective inhibitor of MSK1 , with a reported 300-fold selectivity for MSK compared with RSK. In the present study, we have found that, although SB-747651A is a potent MSK inhibitor, it also inhibits several other kinases, including RSK and MSKs. In contrast with the first report , we did not observe significant selectivity for SB-747651A between MSK1 and RSK. The reason for this difference is unclear, but may relate to differences in the conditions used for the MSK and RSK assays in the different studies. SB-747651A targets the N-terminal kinase domain of MSK1, which is a member of the AGC kinase family. Interestingly, the other kinases (S6K, RSK, PKR2, ROCK, PKB and PKA) inhibited at a similar IC50 value to MSK are also members of the AGC family. SB-747651A also showed some weaker activity against two other AGC kinases in the panel [PKC (protein kinase C) and SGK (serum- and glucocorticoid-induced kinase)]; however, less activity was seen against the non-AGC kinases, suggesting that this inhibitor may target a motif common to the AGC family. This off-target activity was not restricted to in vitro assays, as we also found that SB-747651A could block the actions of PKA, RSK, PKB and S6K in cells at similar concentrations to those required to inhibit MSK1.
SB-747651A was found to affect the activation of several kinases downstream of PI3K (phosphoinositide 3-kinase) signalling, notably PKB and S6K. SB-747651A was able to inhibit the phosphorylation of p70S6K in response to PMA and, to a lesser extent, IGF. This effect is probably due to the ability of SB-747651A to inhibit RSK and PKB. p70S6K is activated by phosphorylation of its hydrophobic motif by mTOR (mammalian target of rapamycin), which then allows the recruitment of PDK1 (phosphoinositide-dependent kinase 1), which in turn phosphorylates the T-loop of p70S6K. The activation of mTOR requires Rheb GTPase, which in turn is dependent on the phosphorylation of TSC2 (tuberous sclerosis complex 2). Downstream of IGF, TSC2 is phosphorylated by PKB, whereas RSK can also phosphorylate TSC2 downstream of mitogens [7,42–46]. Thus the ability of SB-747651A to block PKB and RSK in cells could inhibit p70S6K phosphorylation by preventing the phosphorylation of TSC2. SB-747651A also increased PKB phosphorylation at both Thr308 and Ser473 in response to IGF, which could be due to a negative-feedback loop from p70S6K to IGF signalling. p70S6K can phosphorylate and inhibit the activity of IRS-1 (insulin receptor substrate-1), and so the ability of SB-747651A to inhibit p70S6K (preventing IRS-1 phosphorylation) may promote IGF-mediated activation of PI3K, which as a result increases PKB phosphorylation .
Despite the caveats regarding off-target activities, SB-747651A provides a significant improvement on the previous alternatives. Both Ro 31-8220 and H89 have been used as MSK inhibitors. SB-747651A shows a better selectivity profile for MSK1 compared with Ro 31-8220 in vitro (Table 2). In addition, analysis of Ro 31-8220 action in macrophages shows multiple effects that could not be attributed to inhibition of MSKs (Figures 7 and 8, and results not shown). In vitro, 1 μM H89 showed better selectivity than Ro 31-8220; however, at this concentration MSK1 inhibition was only 75%. A further drawback is that H89 is a more potent inhibitor of PKA than MSKs, and PKA shares in vivo targets such as CREB, ATF1, RARα and p65/RelA with MSKs [21,48,49]. In addition to its actions as a kinase inhibitor, H89 has also been reported to act as an antagonist of β-adrenergic receptors and to modulate the activity of certain Ca2+ and K+ channels [28,50]. In the present study, we show that H89 also affects transcription and cytokine production that are independent of MSKs (Figures 7 and 8).
Analysis of LPS-induced transcription and cytokine production in macrophages (Figures 7 and 8) demonstrated that SB-747651A more closely replicated the effect of MSK1/2 knockout than H89 and Ro 31-8220. Despite this, as shown in Figure 4, it is not without off-target activities. To confirm that an effect of SB-747651A in cells was due to inhibition of MSK, additional experiments could be used. Several commercially available inhibitors could help to distinguish MSK from PKB and S6K. Rapamycin inhibits mTOR and blocks the activation of S6K, whereas Akti (Akt inhibitor) can block PKB activation . It has been shown previously that MSK is not directly inhibited by any of these inhibitors [26,27,39]. Alternatively, a combination of PD184352 and SB-203580 (which inhibit the ERK1/2 and p38 MAPK pathways respectively) will block MSK activation , but they do not normally affect PI3K signalling pathways. Therefore, if a process is insensitive to rapamycin and Akti but sensitive to SB-747651A as well as to a combination of PD184352 and SB203580, PKB and S6K can be excluded. Distinguishing between MSK and RSK is still problematic as, like MSKs, it is activated by ERK1/2. One possibility would be the use of recently described RSK inhibitors, which do not affect MSK activity [32,33]. A second complementary approach would be the use of RNAi (RNA interference) knockdown or gene-targeted mice.
In summary, in the present study we show that SB-747651A is an effective inhibitor of MSKs in cells and, although it has some off-target effects, it does replicate many of the effects of MSK1/2 knockout in cells.
Shaista Naqvi carried out the work in macrophages and MSK2 knockouts, Joanne Darragh analysed the gene induction in MEFs, and Andrew Macdonald and Claire McCoy carried out the work to characterize the mechanism and off-target effects of SB-747651A. Alastair Reith was responsible for screening campaigns to identify MSK inhibitors and selection of SB-747651A. Simon Arthur co-ordinated and designed the study. All authors contributed to writing the paper.
This work was supported by the MRC.
We thank the National Centre for Protein Kinase Profiling (http://www.kinasescreen.mrc.ac.uk/) for help with the kinase selectivity screens.
Abbreviations: Akti, Akt inhibitor; ATF1, activating transcription factor 1; BMDM, bone-marrow-derived macrophage; CREB, cAMP-response-element-binding protein; Egr1, early growth response 1; ERK, extracellular-signal-regulated kinase; FBS, fetal bovine serum; GSK, glycogen synthase kinase; HEK, human embryonic kidney; IGF, insulin-like growth factor; IL, interleukin; IL-1ra, IL-1 receptor antagonist; IRS, insulin receptor substrate; JNK, c-Jun N-terminal kinase; LPS, lipopolysaccharide; MAPK, mitogen-activated protein kinase; MAPKAP, MAPK-activated protein; MAPKAPK, MAPKAP kinase; MEF, murine embryonic fibroblast; MKK, MAPK kinase; MSK, mitogen- and stress-activated kinase; mTOR, mammalian target of rapamycin; PI3K, phosphoinositide 3-kinase; Pim, provirus integration site for Moloney murine leukaemia virus; PKA, protein kinase A; PKB, protein kinase B; PKC, protein kinase C; PRK2, double-stranded-RNA-dependent protein kinase 2; RAR, retinoic acid receptor; ROCK, Rho-associated protein kinase; RSK, ribosomal S6 kinase; RT, reverse transcription; S6K, S6 kinase; TNF, tumour necrosis factor; TSC, tuberous sclerosis complex
- © The Authors Journal compilation © 2012 Biochemical Society