Streptococcus pneumoniae stabilizes tumor necrosis factor α mRNA through a pathway dependent on p38 MAPK but independent of Toll-like receptors

Given the important role of the inflammatory response in the pathogenesis of streptococcal diseases [1], we were interested in studying mechanisms involved in expression of inflammatory cytokines during infection with S. pneumoniae. In a first set of experiments, we measured expression of TNF-α in murine macrophages after treatment with S. pneumoniae, N. meningitidis or pure ligands for TLR2 (Pam3Csk4), TLR4 (LPS), and TLR9 (ODN1826). As expected, cells responded to addition of bacteria or TLR ligands by secreting high amounts of TNF-α (Fig. 1A). LPS has been reported to stimulate TNF-α expression by both activating gene transcription, stabilizing the TNF-α mRNA, and supporting translation [11‐13]. To examine whether S. pneumoniae stabilized TNF-α mRNA, we treated cells with actinomycin D, which prevents de novo transcription of mRNA, and stimulated the macrophages with S. pneumoniae, N. meningitidis or LPS. At the indicated time-points post-stimulation, RNA was harvested and TNF-α mRNA levels were determined by real-time PCR. As shown in Fig. 1B, addition of actinomycin D led to a rapid decline in the levels of remaining TNF-α mRNA in cells receiving mock treatment. However, treatment with S. pneumoniae profoundly augmented the half-life of TNF-α mRNA (from about 30 min to about 6 hours). LPS and N. meningitidis also stabilized the TNF-α messenger but not to the same extent as S. pneumoniae. To examine whether the AREs of the TNF-α mRNA were required for the observed effect, we turned to a cell system derived from RAW264.7 cells, consisting of two cell lines stably expressing a reporter gene (chloramphenicol acetyl transferase, CAT), transcribed from a constitutive promoter. In one of the cell lines (RAW TNF-α 3'-UTR AU+) the 3'-UTR of the mRNA encoding the reporter gene is derived from the wild-type TNF-α mRNA, and in the other cell line (RAW TNF-α 3'-UTR AU÷) the AREs of the TNF-α 3'-UTR are mutated. When the RAW TNF-α 3'-UTR AU+ cell line was incubated in the presence of S. pneumoniae, a very strong increase in CAT levels was observed (Fig. 1C), whereas LPS and also N. meningitidis affected accumulation of CAT protein only to a more modest extent. Interestingly, the effect of S. pneumoniae on accumulation of CAT was largely abrogated in the RAW TNF-α 3'-UTR AU÷ cell line. Thus, S. pneumoniae stimulates production of TNF-α and stabilizes TNF-α mRNA through a mechanism involving the AREs in the 3'-UTR of the mRNA.

We next examined the pneumococcal requirement for mediating TNF-α mRNA stability. To this end, we compared the ability of live versus killed bacteria to induce CAT expression in the RAW TNF-α 3'-UTR AU+/AU- cell lines (Fig. 2A and 2B). As also shown in Fig. 1C, S. pneumoniae induced accumulation of CAT, and this was dependent on the 3'-UTR AREs (Fig. 2). Interestingly, S. pneumoniae killed either by heat-inactivation, UV light or penicillin displayed reduced ability to induce CAT expression, with the reduction ranging from almost 100% (heat-inactivated bacteria) to about 65% (penicillin-killed bacterial). Thus, stabilization of TNF-α mRNA by S. pneumoniae is dependent on the viability of the bacteria.

The S. pneumoniae SK1013 strain used in this study has previously been demonstrated to be recognized by TLR2 and 9 [16], with the former recognizing bacterial peptidoglycan and lipoteichoic acid and the latter recognizing bacterial DNA [16‐18, 26]. We have previously demonstrated that TLR9 recognizes live but not heat-killed S. pneumoniae [16] wherefore the data shown in Fig. 2 prompted us to examine the role of TLR9 in stabilization of TNF-α mRNA. We first examined how pure TLR agonists affected CAT expression in the RAW-TNF-α 3'-UTR AU+ cell line and compared this with the response evoked by live S. pneumoniae. The agonists for TLR2, 4, and 9 all enhanced CAT expression but to a rather moderate extent compared to what was observed in cells receiving live S. pneumoniae (Fig. 3A). To directly assess the role of TLR9, we treated cells with the TLR9 antagonist ODN2088 prior to addition of S. pneumoniae and measured CAT expression in lysates at later time points. Although ODN2088 inhibited induction of TNF-α by the TLR9 agonist ODN1826 by more than 95% (data not shown), the presence of the TLR9 antagonist had no effect on CAT expression induced by S. pneumoniae (Fig. 3B), thus suggesting that mRNA stabilization by S. pneumoniae is independent of TLR9.

To more broadly examine the role of TLRs in stabilization of TNF-α mRNA during S. pneumoniae infection, we pretreated cells with a cell-permeable MyD88 inhibitory peptide or an inactive control peptide prior to addition of bacteria. However, this treatment did not affect S. pneumoniae-induced CAT expression (Fig. 3C), although the inhibitory peptide did indeed inhibit TLR signaling both strongly and specifically (Fig. 3D). Thus, the ability of S. pneumoniae to stabilize TNF-α mRNA seems to be independent of bacterial signaling through TLRs.

The signaling pathways responsible for mediating mRNA stability have recently been studied extensively, and it appears that P38 MAPK plays a central role in this process [13]. To examine the role of p38 in stabilization of TNF-α mRNA in response to S. pneumoniae infection, we treated the RAW TNF-α 3'-UTR AU+ cell line with the p38 inhibitor SB202190 15 min prior to addition of S. pneumoniae. CAT was subsequently measured in cell lysates. As shown in Fig. 4A, the strong elevation of CAT protein levels observed after infection with live S. pneumoniae was significantly inhibited by the presence of SB202190. It has been reported that the p38 MAPK inhibitor SB203580, which is structurally very similar to SB202190, not only inhibits p38 but also the kinase receptor-interacting protein (RIP) 2 [27], which plays an important role in signaling downstream of the S. pneumoniae-activated PRR NOD2 [27]. Therefore, we also examined the effect of another RIP2 inhibitor, PP2 [27], on induction of CAT protein in our reporter system However, as illustrated in Fig. 4A, no effect of PP2 treatment towards bacteria-induced CAT production was observed.

In Fig. 2 we demonstrated that live but not heat-killed S. pneumoniae mediated stabilization of TNF-α mRNA. Therefore, we hypothesized that live versus heat-killed S. pneumoniae may differentially activate p38. RAW264.7 cells were stimulated with live or heat-killed S. pneumoniae for the indicated time intervals, whole-cell lysates were harvested and the levels of phosphorylated p38 were measured. Indeed, the results showed that live bacteria potently stimulated phosphorylation of p38, heat-killed bacteria were not able to elicit such a response (Fig. 4B).

To more thoroughly study the role of p38 in S. pneumoniae-mediated mRNA stabilization, we wanted to examine how inactivation of TLR signaling, which did not affect TNF-α mRNA stabilization (Fig. 3), might influence p38 activation. Therefore, we pretreated RAW264.7 cells and primary macrophages with the MyD88 inhibitory peptide or an inactive control peptide prior to addition of S. pneumoniae. The cells were lysed 30 min post-infection and levels of phospho-p38 were measured. In both cell types we observed strong activation of p38 in response to S. pneumoniae treatment and this was independent of the presence of the MyD88 inhibitory peptide (Fig. 4C and 4D). Thus, stabilization of TNF-α mRNA by S. pneumoniae is dependent on p38, and this kinase is activated by live bacteria through a mechanism independent of TLRs.

P38-mediated mRNA stabilization can occur through several mechanisms, including involvement of members of the MAPK-activated protein kinase family, in particular MK2 [28‐30]. To examine the role of this family of kinases in stabilization of TNF-α mRNA, we compared macrophages from C57BL/6 versus MK2-/- or MSK1/2-/- mice with respect to stabilization of endogenous TNF-α mRNA after challenge with S. pneumoniae. Furthermore, we also examined the effect of the Mnk inhibitor CGP57380 on production of CAT protein in the reporter cell line system. In all cases, no role for the MAPK-activated protein kinases was found (data not shown).

In order to gain more insight into the mechanism of p38 activation during infection with S. pneumoniae, we first examined the role of TAK1, which has been ascribed an important role in this process by activating a number of different pathways [31‐33]. RAW264.7 cells were seeded and treated with the TAK1 inhibitor (5Z)-7-oxozeaenol 15 min prior to stimulation with S. pneumoniae or anisomycin. Lysates were prepared 30 min post-stimulation, and levels of phospho-p38 were measured. As shown in Fig. 5A, both S. pneumoniae and anisomycin potently induced phosphorylation of p38. Whereas the TAK1 inhibitor did not affect activation of p38 in response to anisomycin, the ability of S. pneumoniae to induce p38 phosphorylation was abolished in the presence of the TAK1 inhibitor.

MKK3, as well as MKK4, and 6 are p38 upstream kinases and can be activated through TAK1 [34]. To evaluate the role of MKK3 in the S. pneumoniae-induced p38 activation pathway, we compared the ability of spleen cells from C57BL/6 and MKK3-/- mice to activate p38 in response to S. pneumoniae treatment. Anisomycin was included as a control, and was found to trigger p38 activation independently of MKK3 (Fig. 5B). By contrast, MKK3-/- cells responded to S. pneumoniae with significantly reduced, although not abolished, activation of p38. Taken together, these data demonstrate that S. pneumoniae activates p38 through a TLR-independent pathway involving the upstream kinases TAK1 and MKK3.

C57BL/6, MK2-/-, MSK1/2-/-, MKK3-/- mice were obtained from Taconic (Laven, Denmark), professor Mathias Gaestel (Hannover, Germany), professor Simon Arthur (Dundee, UK), and professor Richard A Flavell (Yale, USA), respectively, and bred in the animal facility of The Faculty of Health Sciences, University of Aarhus, under pathogen-free conditions. Peritoneal macrophages were harvested by lavage of the peritoneal cavity with PBS supplemented with 5% LPS-free heat-inactivated fetal calf serum (FCS) (Cambrex, Baltimore, MD, USA) and 400 μl heparin (Leo Pharma, Copenhagen, Denmark) per 100 ml. RAW264.7 macrophage-like cells and the derived cell lines RAW TNF-α 3' untranslated region (UTR) AU+ and RAW TNF-α 3'UTR AU÷ [38‐40] were maintained in DMEM supplemented with 5% FCS, antibiotics and for the latter 2 cell lines also with 500 μg/ml G418 (Roche, Basel, Switzerland). For experiments, primary macrophages were seeded in 96-well or 6-well tissue plates in RPMI 1640 medium containing 5% FCS at a density of 3.0 × 10⁵and 6.0 × 10⁶ cells per well, respectively. RAW264.7 and derived cell lines were seeded in 96-well or 6-well tissue plates at a density of 1.0 × 10⁵ and 6.0 × 10⁶ cells per well, respectively. After seeding, cells were left overnight to settle before further treatment.

The bacteria used were the S. pneumoniae strain SK1013 serotype 4 (TIGR4) and the N. meningitidis strain NGO93 serogroup B, serotype 15:P1.2, MLEE cluster I1, ET76. The bacteria were grown overnight in Brain Heart Infusion broth with 10% Levinthal broth (Statens Serum Institute, Copenhagen) reaching a concentration of 18.0 ± 2.2 × 10⁸ bacteria per ml as determined in a Thoma counting chamber. For stimulation, 300 μl and 10 μl from this stock was added to the cell cultures in 6- and 96-well plates, respectively, reaching final volumes of 3 ml and 100 μl, respectively. The bacteria were heat-inactivated by incubation of the cultures for 30 min at 65°C. To UV-inactivate the bacteria, the cultures were exposed to UV light for 10 min, and for killing of bacteria with penicillin, the cultures were incubated for 2 h with 1.0 μg/ml of the antibiotics.

For measurement of TNF-α. an ELISA duoset from R&D Systems was used, following recommendations of the manufacturer. The concentrations of coating Ab and detection Ab were 0.8 and 0.15 mg/ml, respectively. The ELISA was able to detect TNF-α from 15 to 2000 pg/ml.

Total RNA was extracted with TRIzol (Invitrogen, Carlsbad, CA, USA) according to the recommendations of the manufacturer. Briefly, cells were lysed in TRIzol, and chloroform was added, followed by phase separation by centrifugation. RNA was precipitated with isopropanol and pelleted by centrifugation. Pellets were washed with 80% ethanol and re-dissolved in RNase-free water. For cDNA generation, 1 μg of RNA was subjected to reverse transcription (RT) with oligo(dT) as primer and Expand Reverse Transcriptase (both from Roche). Prior to RT-PCR, RNA was treated with DNase I (Ambion, Austin, TX, USA) to remove any contaminating DNA, the absence of which was confirmed in control experiments where the reverse transcriptase enzyme was omitted (data not shown). To measure the relative amount of TNF-α mRNAs, amplification of sample cDNA was monitored with the fluorescent DNA binding dye SYBR (QIAGEN). The PCR primers used in this study were: TNF-forward, TGGGAGTAGACAAGGTACAACCC; TNF-reverse, AGAGGGAAATCGTGCGTGAC; β-actin-forward, GCTCCCCGGGCTGTATTCC; β-actin-reverse, CTCTCTTGCTCTGGGCCTCGT). The data on TNF-α were normalized to β-actin, and presented as normalized ratio.

The cells were seeded and treated as described for the appropriate amount of time. For cell lysis and measurement of CAT levels the CAT ELISA kit from Roche (Basel, Switzerland) was used.

To assay for levels of total and phosphorylated p38, cells were seeded in 6-well plates as described above and stimulated with bacteria as specified in the text. At different time points post-stimulation, cells were lysed using the Bio-Plex Cell Lysis Kit (Bio-Rad, Hercules, CA, USA) according to the recommendations of the manufacturer. Briefly, the cells were washed with 3 ml Cell Wash Buffer per well and treated with 1 ml Lysing Solution supplemented with PMSF followed by incubation for 20 min at 4°C. The suspension was centrifuged at 4500 × g for 20 min at 4°C and supernatants were harvested as whole-cell extracts.

The levels of total and phosphorylated p38 were measured using the Luminex Technology™ and kits purchased from Bio-Rad. Briefly, the filter plate was washed with assay buffer and freshly vortexed antibody-conjugated beads were added to each well. The plate was washed with assay buffer and samples (45 μg of protein in 50 μl lysis buffer) were added. After a brief shake (30 sec at 1.100 rpm), the plate was incubated with shaking (300 rpm) overnight at 4°C. After a wash-step, detection antibody was added to each well, and the plate was shaken and incubated shaking (300 rpm) at room-temperature in the dark for 45 min to 2 h. Subsequently, the plate was washed and incubated for 10 min with 50 μl of a streptavidine-PE solution with shaking (30 sec at 1.100 rpm, 10 min 300 rpm). Finally the plate was washed and 125 μl of assay buffer was added to each well and the plate was shaken for 10 sec at 1100 rpm and read immediately on the Bio-Plex reader.

Nitrite is generated by the rapid oxidation of NO. It is stable and its accumulation in the culture medium reflects the amount of NO produced. To assay nitrite we used the Griess reaction. Aliquots of 100 μl culture supernatants were mixed with equal volumes of Griess reagent (equal volumes of 0.1% N-(1-naphthyl)-ethylenediamine dihydrochloride and 1% p-aminobenzene sulphanilamide diluted in 2.5% phosphoric acid) in a 96-well microtitre plate (Maxisorb Immunoplate, Nunc). After 10 min incubation at room temperature the absorbance at a wavelength of 540 nm was measured in a microplate reader (model 450; Bio-Rad). A range of 2-fold dilutions of sodium nitrite (0.05 to100 μM) in RPMI medium was run in each assay to generate a standard curve. In most experiments, control cultures not infected or stimulated gave small nitrite values in the order of 0–1 μM. These figures were subtracted from the experimental values in the data presentation.

The work presented in this paper did not include research on humans. Animals, used for isolation of cells for in vitro experiments were sacrificed by cervical dislocation without receiving any prior experimental treatment. Therefore, to perform the experiments described in this article, no need for approval from ethics committees was required according to Danish law.