Results

M-CSF Induces NF-κB Transcriptional Activity in Human Monocyte–Derived Macrophages (MDMs) and Mouse Macrophage Cell Line, RAW 264.7
To determine if M-CSF induced NF-κB DNA binding in human macrophages, we performed EMSA analysis on nuclear lysates from M-CSF-treated MDMs. Similar to previous reports [11], nuclear NF-κB constitutively bound DNA in non-stimulated monocytes (Figure 1A). Interestingly, adding M-CSF did not alter NF-κB DNA binding by EMSA. In contrast, after transiently transfecting human MDMs with pNF-κB-SEAP constructs containing four NF-κB consensus binding sequences, M-CSF treatment of the transfected cells resulted in a 2.3-fold increase in SEAP release in the culture media compared to PBS (vehicle)-treated transfected MDMs (Figure 1B). As a control, the pTAL-SEAP construct lacking NF-κB binding sites was used. Cells transfected with the pTAL-SEAP construct did not produce SEAP in the absence or presence of M-CSF (Figure 1B).
10.1371/journal.pone.0028081.g001 Figure 1  M-CSF induces NF-κB transcriptional activity in macrophages.
(A) Nuclei were extracted from human MDMs treated without or with M-CSF for 15 or 30 minutes. NF-κB DNA binding activity was analyzed by EMSA Shown is a representative blot from three independent experiments. NS: non-stimulated. (B) Human MDMs transiently transfected with pTAL-SEAP or pNF-κB-SEAP constructs were treated with M-CSF in X-vivo medium and incubated for 6 hours before the collection of medium. NF-κB activity was analyzed by measuring the amount of SEAP secreted into the medium and data are expressed as fold increase of SEAP activity over that in pTAL-SEAP transfected resting cells. (C) RAW 264.7 cells were transiently transfected with pTAL-SEAP or pNF-κB-SEAP construct. Cells were serum starved for 4 hours prior to 2 hours stimulation with mouse recombinant M-CSF (100 ng/ml). Culture media was collected to measure SEAP production. Data are expressed mean ± S.E.M, for three independent experiments.   We next investigated whether M-CSF induced NF-κB activity in the mouse macrophage cell line, RAW 264.7. RAW 264.7 cells were transfected with either the NF-κB-SEAP reporter or control pTAL-SEAP construct. As shown in Figure 1C, M-CSF treatment of RAW 264.7 cells increased NF-κB reporter activity by 2.5-fold over that of non-treated cells. Together, our data demonstrate that M-CSF induced NF-κB transcriptional activity in macrophages.

PKC Inhibition Reduces NF-κB Activity in Human MDMs and RAW 264.7 Cells
Since NF-κB is activated by PKC in several cell types [36], we next determined if M-CSF-induced NF-κB transcriptional activity was dependent on PKC activation and if calcium, a co-factor for conventional PKC isoform activation, was important. In addition, because of the number of conventional PKCs existing within mononuclear cells, pharmacological inhibitors of PKC family activation was used to determine the relationship of PKC activation to M-CSF-induced cellular survival. MDMs were transfected with the pNF-κB-SEAP reporter and then treated with the general PKC inhibitor, Ro-31-8220; the conventional PKCα/β inhibitor Gö-6976; or the intracellular calcium blocker BAPTA/AM. Ro-31-8220 significantly suppressed M-CSF-induced NF-κB activity compared to cells treated with M-CSF alone (Figure 2A). In addition, Gö-6976 and BAPTA/AM also blocked NF-κB activity in M-CSF-treated MDMs. Trypan blue exclusion analysis did not indicate cell death suggesting that this suppression was not due to non-specific toxicity. These data indicate that M-CSF mediated NF-κB activation through calcium-dependent conventional PKC activation.
10.1371/journal.pone.0028081.g002 Figure 2  Inhibition of PKC reduces NF-κB activity in M-CSF stimulated macrophages.
(A) Human MDMs transfected with pNF-κB-SEAP were pre-incubated with inhibitors; Ro-31-8220, Gö-6976, or BAPTA/AM for 30 minutes prior to 6 hours M-CSF stimulation. (B) RAW 264.7 cells transfected with the pNF-κB-SEAP construct were pre-incubated with inhibitors for 30 minutes prior to 2 hours of stimulation with mouse recombinant M-CSF. The NF-κB activity was analyzed by measuring SEAP production in the medium and data are expressed as fold increase of SEAP activity over that in pTAL-SEAP transfected resting cells. The graph represents mean ± S.E.M for three independent experiments. *The p-values of cells treated with inhibitors/M-CSF compared to vehicle/M-CSF were ≤0.05.   We next investigated whether PKC inhibition affected NF-κB activity in the mouse macrophage cell line, RAW 264.7. Similar to MDMs, PKC inhibitors Ro-31-8220, Gö-6976 and BAPTA/AM reduced M-CSF-induced NF-κB activity in a dose-dependent manner in RAW 264.7 macrophages (Figure 2B). To ensure that PKC specifically regulated NF-κB p65 and not the closely related family member, c-Rel, we co-transfected c-Rel and NF-κB-SEAP constructs into the Raw 264.7 cell line and measured NF-κB activity in response to M-CSF. There was no increased NF-κB activity in cells expressing c-Rel (Figure S1). These observations indicate that PKC functioned upstream of NF-κB p65 in MDMs and RAW 264.7 cells.

NF-κB and PKC(s) Mediate Human MDM Survival in Response to M-CSF Stimulation
Since PKC and NF-κB are critical in cell survival [37], [38], we hypothesized that M-CSF promoted cell survival through PKC in human MDMs. To detect apoptosis, the expression of cleaved caspase-3, a marker of apoptosis, was analyzed in the presence or absence of M-CSF and PKC inhibitors. In MDMs pretreated with PKC inhibitors (Ro-31-8220, Gö-6976, or BAPTA/AM) and stimulated with M-CSF, cleaved caspase-3 was elevated to levels of cells treated with vehicle alone (Figure 3A). In contrast, cells incubated with M-CSF and vehicle had less cleaved caspase-3 than cells incubated with vehicle alone or M-CSF with PKC inhibitors. These data supported the hypothesis that M-CSF-induced NF-κB activity was regulated by conventional PKCs, not novel PKCs.
10.1371/journal.pone.0028081.g003 Figure 3  Inhibition of PKC or NF-κB induces apoptosis in MDMs.
(A) MDMs were pre-incubated in RPMI medium containing inhibitors (Ro-31-8220: 5 µM, Gö-6976: 5 µM, BAPTA/AM: 2.5 µM) for 30 minutes prior to the addition of M-CSF. As a control, untreated cells were incubated with dimethyl sulfoxide (DMSO). Cell lysates were resolved by SDS-PAGE and immunoblotted with antibody recognizing the active cleaved form of caspase-3. The blots were reblotted with β-actin that served as a loading control. The ratio of active caspase-3 bands (17 kD and 19 kD) to β-actin control was determined by densitometry analysis (bottom panel). Data represents the mean ± S.E.M from two independent donors. (B) Apoptosis of the treated MDMs was also measured by Annexin V-FITC and propidium iodine (PI) staining and analyzed by flow cytometry. The percentage of surviving cells (Annexin V/PI negative) for the cells treated with vehicle/M-CSF was arbitrarily set as 100. Data shown represent the mean ± S.E.M from three independent experiments. *The p-values of inhibitors/M-CSF compared to vehicle/M-CSF were ≤0.05.   To confirm that PKC inhibition decreased cell survival, cells were treated with M-CSF in the absence or presence of PKC inhibitors and then also examined for apoptosis by Annexin V/PI staining. PKC inhibitors in the presence of M-CSF reduced the number of Annexin V/PI negative MDMs compared to cells treated with M-CSF alone (Figure 3B). These observations further suggested that MDM survival is promoted by M-CSF and critically involves conventional PKCs and NF-κB activity.

M-CSF-Induces PKCα Kinase Activity
Since our data suggested that conventional PKCs were involved in activating NF-κB in response to M-CSF in primary human macrophages, we next investigated whether the conventional PKC, PKCα was a downstream target of M-CSF in human MDMs. PKCα was immunoprecipitated from human MDMs at the indicated time points (Figure 4), and then a PKC kinase assay was performed using a fluorescein-tagged peptide as a substrate. As shown in Figure 4 (upper panel), the peptide substrate was maximally phosphorylated within 10 minutes of stimulation (1.8-fold over resting cells, lower panel), and then returned to basal levels by 15 minutes. This effect was not seen in M-CSF-stimulated monocytes when isogenic antibodies (IgG) were used for immunoprecipitation. Western blots of identical samples using an antibody recognizing PKCα demonstrated that equal amounts of PKCα were assayed (Figure 4, middle panel).
10.1371/journal.pone.0028081.g004 Figure 4  M-CSF activates PKCα in human MDMs.
MDMs treated with M-CSF (100 ng/ml) for varying amounts of time were lysed and immunoprecipitated using anti-PKC antibody or control IgG antibody. One half of the samples was used to analyze PKC kinase activity using a fluorescein tagged peptide and visualized by agarose gel electrophoresis (top panel), while the other half was subjected to Western blot analysis to confirm equal amounts of PKCα were immunoprecipitated from each sample (middle panel). The kinase assay was quantitated using Quantity One software (Bio-Rad) (bottom panel). Data represents the average fold increase of PKCα activity in non-stimulated samples compared to M-CSF-treated MDM ± S.E.M for three independent experiments. NS: non-stimulated. * The p-values of M-CSF stimulated compared to non-stimulated were ≤0.05.

M-CSF-Dependent Activation of PKC Does Not Regulate the Classical NF-κB p65 Activation Pathway
Canonical activation of NF-κB occurs via phosphorylation and degradation of IκBα leading to the release and nuclear translocation of the NF-κB p50/p65 heterodimer to transactivate target genes [37], [39]. Since conventional PKC activity was important in regulating M-CSF-induced NF-κB activation, we next investigated whether IκBα degradation was regulated by PKC. Cells were treated with cyclohexamide (CHX) to inhibit protein synthesis of IκBα, and its degradation was followed. As shown in Figure 5A, PKC inhibition with Ro-31-8220 did not alter M-CSF-induced IκBα degradation, suggesting that M-CSF-induced PKC activity augmented NF-κB transcriptional activity by an alternative pathway, like post-translational modification of NF-κB p65.
10.1371/journal.pone.0028081.g005 Figure 5  M-CSF-induced PKC activity does not regulate IκBα degradation but regulates the phosphorylation of NF-κB p65 at Ser276.
(A) MDMs were pretreated with cycloheximide (CHX) in the absence or presence of Ro-31-8220 for 30 minutes prior to M-CSF stimulation for the indicated times. Whole cell lysates were subjected to Western blotting with anti-IκBα antibody. Data are representative of three independent experiments. (B) MDMs were pretreated with either vehicle or Ro-31-8220 for 30 minutes before the addition of M-CSF for 10 minutes. Whole cell lysates were resolved by SDS-PAGE and phospho-Ser276 or phospho-Ser536 of NF-κB p65 was detected using phospho-specific antibodies to either residue of NF-κB p65. (C) Whole cell lysates of RAW 264.7 cells treated with vehicle control or Ro-31-8220 in the absence or presence of M-CSF were subjected to Western blot analysis with phospho-Ser276 or phospho-Ser536 NF-κB p65 antibodies. (D) Cytosolic and nuclear fractions of RAW 264.7 were obtained from the treated cells and immunoblotted for phospho-NF-κB. The purity of the cytosolic and nuclear fractions was confirmed by immunoblotting with GAPDH and Lamin B, respectively. Shown are representative blots from three independent experiments.

M-CSF Induces Phosphorylation of NF-κB Ser276 in a PKC-dependent Fashion
Since M-CSF did not regulate NF-κB activation by influencing IκBα, we next sought to determine if M-CSF affected NF-κB p65 by post-translational mechanisms. Thus, we examined the phosphorylation of NF-κB p65 with specific phospho-NFκB p65 (Ser276 and Ser536) antibodies. M-CSF induced the phosphorylation of Ser276 but not Ser536 of NF-κB p65 in MDMs. Compared to vehicle, the general PKC inhibitor Ro-31-8220 reduced Ser276 phosphorylation, but not Ser536, phosphorylation in M-CSF-stimulated cells (Figure 5B). Furthermore, M-CSF-stimulated NF-κB p65 phosphorylation at residue Ser276 in RAW 264.7 cells was also PKC dependent (Figure 5C). These studies suggested that PKC(s) regulated Ser276 phosphorylation but not Ser536 in both human MDMs and mouse macrophages after M-CSF stimulation.
We next performed cellular fractionation to identify the cellular location of phosphorylated NF-κB p65 in Raw 264.7 cells. Non-phosphorylated NF-κB p65 was located in both cytosolic and nuclear fractions, but phosphorylated Ser276 and Ser536 NF-κB p65 was primarily located in nuclear fraction after M-CSF stimulation (Figure 5D). Notably, constitutive phosphorylation of Ser536 NF-κB p65 was found in these cells. Importantly, Ro-31-8220 reduced M-CSF-induced Ser276 phosphorylation of NF-κB p65 in both the cytosolic and nuclear fractions, while M-CSF-induced NF-κB p65 Ser536 phosphorylation was present in the nucleus regardless of PKC inhibition. These observations indicate that M-CSF-induced Ser276 and Ser536 are regulated differently by conventional PKC activation in mononuclear phagocytes. The purity of the cytosol and nuclear cell fractions was confirmed by immunoblotting with GAPDH and Lamin B, respectively (Figure 5D).

M-CSF-dependent PKC Regulates NF-κB-targeted Genes
NF-κB induces a number of downstream genes, including the IκB family. Among the IκB molecules, IκBα is highly induced by NF-κB activation [40]. Having shown that PKC regulated NF-κB activity in M-CSF-stimulated MDMs, we next determined whether inhibition of PKC activity decreased expression of NF-κB-regulated genes. We treated both MDMs and RAW 264.7 cells with the PKC inhibitor Ro-31-8220 for 30 minutes and then stimulated with M-CSF. IκBα gene was measured by qRT-PCR. As shown in Figures 6A and 6B, M-CSF enhanced IκBα gene expression and PKC inhibition by Ro-31-8220 decreased IκBα gene expression in both MDMs and RAW 264.7 cells (p<0.01), demonstrating that PKC affected NF-κB-regulated gene expression in macrophages.
10.1371/journal.pone.0028081.g006 Figure 6  Inhibition of M-CSF-induced PKC reduces NF-κB-regulated genes in both MDMs and RAW 264.7 cells.
MDMs (A and C) and RAW 264.7 (B) cells were pretreated with Ro-31-8220 or solvent control DMSO for 30 minutes prior to M-CSF stimulation for the indicated times. Total RNA was isolated and converted to cDNA. Real-time RT PCR was performed using primers for IκBα, BCL-xl or GAPDH. Data are expressed as relative fold increase of IκBα or BCL-xl gene expression upon treatment over non-stimulated cells. Data represent the mean ± S.E.M for three independent experiments.   To further define the role of PKC in mediating human MDM survival in response to M-CSF, we examined the expression of the anti-apoptotic gene BCL-xL, which is also regulated by NF-κB. As shown in Figure 6C, Ro-31-8220 reduced M-CSF-stimulated BCL-xL expression compared to cells treated with M-CSF and the vehicle control DMSO (p<0.05).

Identification of PKCα as the Upstream Activator of NF-κB in Myeloid Cells
Even though the PKC family consists of 10 members, finding that PKCα/β inhibitors and intracellular calcium inhibitors reduced M-CSF-induced NF-κB activity, suggested PKCα was involved in NF-κB activation after M-CSF treatment. To confirm the role of PKCα in NF-κB activation in macrophages, constructs for either wildtype (WT)-PKCα or kinase-deficient (KD)-PKCα was co-transfected with the pNF-κB-SEAP reporter gene and SEAP secretion was measured. As shown in Figure 7A, MDMs co-transfected with pNF-κB-SEAP and WT-PKCα had a 1.8-fold increase in NF-κB transcriptional activity after M-CSF activation compared with NS cells (p = 0.05), similar to M-CSF-treated cells expressing only pNF-κB-SEAP. Transfecting human macrophages with the KD-PKCα construct significantly reduced M-CSF-induced NF-κB activity compared to WT-PKCα transfected cells (p = 0.016). Similarly, RAW 264.7 cells transfected with WT-PKCα had 2.5-fold more NF-κB transcriptional activity after M-CSF activation compared to unstimulated RAW 264.7 cells (NS) transfected with WT-PKCα (Figure 7B). Expression of the KD-PKCα construct into RAW 264.7 cells reduced M-CSF-induced NF-κB activity to 1.5-fold (p = 0.045) compared to cells transfected with WT PKCα.
10.1371/journal.pone.0028081.g007 Figure 7  PKCα regulates phosphorylation of NF-κB p65 at Ser276.
(A) MDM or (B) RAW 264.7 cells were transiently transfected with pNF-κB-SEAP along with either WT-PKCα or the kinase-deficient (KD)-PKCα construct at a 1∶5 ratio. Cells were serum starved and stimulated with M-CSF and then SEAP secretion in the medium was measured. Data is from of three independent experiments. The p-value of cells transfected with KD compared to those transfected with WT was 0.05. (C) MDMs were removed from the plate using accutase and apoptosis of MDMs was measured by flow cytometry using Annexin V-FITC and propidium iodine (PI). (D) Whole cell lysates from the transfected RAW 264.7 cells were subjected to Western blot analysis with phospho-Ser276 or phospho-Ser536 NF-κB p65 antibodies. Blots were immunoblotted with PKCα to determine equal protein expression for the PKCα constructs. β-actin served as a loading control. Shown is a representative blot from three independent experiments. (E) MDM or (F) RAW 264.7 cells were transiently transfected with a pNF-κB-SEAP along with either 100 nM PKCα siRNA or control siRNA for 20-24 hours. Cells were serum starved for 2-4 hours and stimulated with 100 ng/ml M-CSF for 6 hours for MDM or RAW 264.7 for 2 hours and then SEAP secretion in the medium was measured. Shown is data of three independent experiments. (G) MDMs were removed from the plate using accutase and apoptosis of MDMs was measured by Annexin V-FITC and propidium iodine (PI) staining and analyzed by flow cytometry. (H) Whole cell lysates from the transfected MDM and RAW 264.7 cells were subjected to Western blot analysis with PKCα antibody. β-actin served as a loading control. Shown is a representative blot from at least three independent experiments.   Cell survival was also examined in MDMs expressing either WT-PKCα or KD-PKCα constructs by Annexin V-FITC and PI staining. As expected, M-CSF increased MDM survival as measured by the percent of Annexin V/PI negative cells. Similarly, expression of WT-PKCα protected cells from apoptosis. In contrast, expression of KD-PKCα decreased M-CSF-induced cell survival (p<0.01) (Figure 7C).
Next, we examined the effect of expressing the PKCα constructs on NF-κB phosphorylation. As shown in Figure 7D, expression of KD-PKCα in RAW 264.7 cells did not affect the constitutive phosphorylation at Ser536 of NF-κB p65, but attenuated the phosphorylation at Ser276. Expression of WT-PKCα did not effect the phosphorylation of either residue with or without M-CSF stimulation. These observations demonstrate that PKCα is important in M-CSF-regulated cell survival and NF-κB activation and likely regulated through phosphorylation of Ser276 of NF-κB p65.
To further validate the impact that PKCα played in M-CSF-induced NF-κB transcriptional activity, we next employed PKCα siRNA treatment of MDM or RAW cells. A pool of specific PKCα siRNA were transfected into MDM or Raw 264.7 cells in the presence or absence or M-CSF. Reducing native PKCα expression decreased M-CSF-induced NF-κB transcriptional activity in both MDM (Figure 7E) (p = 0.012) and Raw 264.7 cells (Figure 7F) (p = 0.01). We also examined cell survival of the MDMs by Annexin V-FITC and PI staining after PKCα siRNA transfection. As shown in Figure 7G, M-CSF-induced MDM survival was reduced in the cells transfected with PKCα siRNA compared with cells transfected with control siRNA (p = 0.047). In Figure 7H, we confirmed that PKCα siRNA transfection decreased PKCα protein expression in both MDM and Raw 264.7 cells. Our results indicated that PKCα regulated NF-κB activation and M-CSF-regulated cell survival.

PKC is Essential in Regulating NF-κB Activity
M-CSF-dependent PKC activity facilitated NF-κB p65 phosphorylation at Ser276 but not Ser536. To confirm the role of PKC and p65 Ser276 phosphorylation on NF-κB activity, we co-transfected the NF-κB-SEAP construct with either WT NF-κB p65 or NF-κB p65 276S/A constructs in Raw 264.7 cells, then treated cells with the PKC inhibitor Ro-31-8220 in the absence or presences of M-CSF. As expected in cells transfected with vector control, inhibiting PKC reduced M-CSF-stimulated NF-κB activity compared to cells treated with M-CSF and the vehicle DMSO (p = 0.001) (Figure 8A). In contrast, expressing WT NF-κB p65 (p65 WT) increased NF-κB activity, while the general PKC inhibitor Ro-31-8220 decreased this NF-κB activity (p = 0.001). Notably, the introduction of NF-κB p65 276 S/A construct significantly reduced NF-κB activity compared with the WT NF-κB p65 construct (p = 0.005) and M-CSF treatment was unable to overcome this inhibition.
10.1371/journal.pone.0028081.g008 Figure 8  NF-κB p65 Ser276 is essential in regulating NF-κB activity.
(A) Raw 264.7 cell line was transiently transfected with pNF-κB-SEAP along with empty vector or plasmid encoding either NF-κB p65 WT or NF-κB p65 276S/A. The cells were transfected for 18-24 hours, serum starved for 4 hours, and then incubated with 10 µM of Ro-31-8220 for 30 minutes prior to treatment with 100 ng/ml of M-CSF for 2 hours and SEAP secretion in the medium was measured. (B) NF-κB p65−/− cell line was transiently transfected with pNF-κB-SEAP along with empty vector or plasmid encoding either NF-κB p65 WT, NF-κB p65 276S/A, or NF-κB p65 536S/A. The cells were cultured for 24 hours and then serum starved for 4 hours. Cells were then incubated in fresh DMEM medium for 2 hours and SEAP secretion in the medium was measured. The results shown are fold change over empty vector + pTAL-SEAP. (C) NF-κB p65−/− cell line was transiently transfected with pNF-κB-SEAP with either empty vector or plasmid encoding either NF-κB p65 WT or NF-κB p65 276S/A. The cells were transfected for 24 hours and serum starved for 4 hours, and then incubated with 10 µM of Ro-31-8220 for 30 minutes prior to treatment with 10 ng/ml of TNFα. The supernatant were collected after 2 hours of treatment and SEAP secretion in the medium was measured. The results shown are the fold change over empty vector + pNF-κB-SEAP. Data shown are mean ± S.E.M for at least three independent experiments performed in duplicate.   Furthermore, we co-transfected the NF-κB-SEAP construct along with either the WT NF-κB p65, NF-κB p65 276S/A or NF-κB p65 536S/A constructs into a NF-κB p65−/− murine fibroblast cell line and measured NF-κB activity. As predicted, expression of WT NF-κB p65 (p65 WT) in NF-κB p65−/− cell line constitutively activated NF-κB compared to cells transfected with vector control without any stimulation (Figure 8B). In comparison, transfecting the NF-κB p65 276S/A construct reduced NF-κB activity by 5-fold (p = 0.006, WT NF-κB p65 vs. NF-κB p65 276S/A), while expressing the NF-κB p65 536S/A construct increased NF-κB activity in the p65−/− cells to levels similar to WT NF-κB p65 transfected cells.
Since M-CSF-induced PKC activation regulated NF-κB activity via Ser276 residue of NF-κB p65 in primary human MDMs and RAW 264.7 cells, we next examined if this occurred in NF-κB p65−/− fibroblasts in response to a native stimulus for these cells, TNFα. NF-κB activity was measured in the WT NF-κB p65 or NF-κB p65 276S/A transfected cells treated with TNFα in the absence or presence of Ro-31-8220 (Figure 8C). As expected, TNFα increased NF-κB activity in NF-κB p65−/− cells expressing WT p65 and (p = 0.001) PKC inhibitors decreased NF-κB activity to the non-stimulated (NS) level (p = 0.007). Introducing NF-κB p65 276 S/A constructs significantly reduced NF-κB activity compared with the WT NF-κB p65 construct (p = 0.001). Notably, TNFα failed to increase NF-κB activity in the NF-κB p65−/− cells expressing NF-κB p65 276S/A constructs. These observations are similar to macrophages overexpressing the NF-κB p65 276S/A (Figure 8A). Our data demonstrate that Ser276 of NF-κB p65 is essential in regulating NF-κB activity and suggests that PKC regulates NF-κB activity by modulating the phosphorylation of NF-κB p65 at Ser276 residue.