Abstract
Macrophages are responsible for the control of inflammation and healing, and their malfunction results in cardiometabolic disorders. TGF-β is a pleiotropic growth factor with dual (protective and detrimental) roles in atherogenesis. We have previously shown that in human macrophages, TGF-β1 activates Smad2/3 signaling and induces a complex gene expression program. However, activated genes were not limited to known Smad2/3-dependent ones, which prompted us to study TGF-β1–induced signaling in macrophages in detail. Analysis of Id3 regulatory sequences revealed a novel enhancer, located between +4517 and 4662 bp, but the luciferase reporter assay demonstrated that this enhancer is not Smad2/3 dependent. Because Id3 expression is regulated by Smad1/5 in endothelial cells, we analyzed activation of Smad1/5 in macrophages. We demonstrate here for the first time, to our knowledge, that TGF-β1, but not BMPs, activates Smad1/5 in macrophages. We show that an ALK5/ALK1 heterodimer is responsible for the induction of Smad1/5 signaling by TGF-β1 in mature human macrophages. Activation of Smad1/5 by TGF-β1 induces not only Id3, but also HAMP and PLAUR, which contribute to atherosclerotic plaque vulnerability. We suggest that the balance between Smad1/5- and Smad2/3-dependent signaling defines the outcome of the effect of TGF-β on atherosclerosis where Smad1/5 is responsible for proatherogenic effects, whereas Smad2/3 regulate atheroprotective effects of TGF-β.
Introduction
Macrophages are key cells of the innate immune system that orchestrate various physiological and pathological processes of an infectious and noninfectious nature. The malfunction of macrophages is a cause of chronic inflammation leading to most serious human pathologies including solid tumor and cardiometabolic disorders (1–4). Macrophages play a highly important role in carcinogenesis, starting from transformed cell control, contributing to tumor initiation by chronic inflammation, and supporting tumor growth and metastasis (1). This high versatility of macrophages and their important role in various diseases prompted researchers worldwide to investigate the possibilities of therapeutic targeting of macrophages or even their use in cell-based therapies (5–7). Strong arguments in favor of therapeutic targeting of macrophages is provided by recent findings made by us and others demonstrating high levels of macrophage functional plasticity (5, 8, 9). It was demonstrated that every combination of cytokines, hormone growth factors, or even nutritional conditions leads to the development of a specific macrophage phenotype (10–12). Highly specific effects on macrophages are demonstrated for cytokines like IFN-γ, various ILs, CC-chemokines, glucocorticoids, retinoids, and increased glucose concentration in culture medium (13–17). These experimental studies are well supported by clinical findings showing involvement of macrophages in the pathogenesis of diseases, such as atherosclerosis (18), obesity-induced diabetes (19), and diabetic complications (20). Although clinical data also indicate the complexity of the macrophage activation system, it was suggested to classify the type of macrophage activation in parallel with Th1-Th2 dichotomy as type 1 (M1) and type 2 (M2), where M1 represents the activation of macrophages capable of stimulating acute inflammation and showing bactericidal activity, whereas M2, having impaired bactericidal activity and promoting resolution of inflammation, shows healing and tolerogenic reactions (21). This classification led to a disputed grouping of stimuli capable of macrophage activation. For instance, the group of type 2 activators comprises hormones (glucocorticoids); vitamins A and D3; IL-4, IL-13, and IL-10; and growth factor TGF-β.
Although the effects of ILs on macrophages are fairly well studied (11, 22), TGF-β remains to be the most puzzling cytokine in regard to its effects on macrophages. Nevertheless, TGF-β plays an important role in the pathogenesis of many diseases where macrophages play a key role as well. Multifunctional growth factor TGF-β is essential for development and normal cell functioning, and is also involved in the pathogenesis of various diseases, including cancer (23, 24). Several attempts to target or use TGF-β therapeutically were proved to be unsuccessful not only because of incomplete knowledge of signaling pathways activated, but also because of unpredictable cell-specific effects. Being a potent activator of various cellular processes, TGF-β remains a highly attractive tool for therapy development provided better knowledge about its effects on various cell populations is available.
The effects of TGF-β on macrophages play important roles in various diseases. The best studied is the role of TGF-β in the development of tumor-associated macrophages. TGF-β promotes the recruitment of monocytes into tumors (25) and their differentiation into macrophages (26). TGF-β predominantly inhibits macrophage functions necessary for the acute phase of inflammation. For example, TGF-β inhibits LPS-induced and proinflammatory cytokine–induced expression of a number of macrophage activation markers (e.g., matrix metalloproteinase-12 -12, inducible NO synthase, TNF) in a Smad-dependent manner (27–30). The knowledge of the signaling that is induced by TGF-β in macrophages is, however, limited to the activation of the canonical Smad2/3-mediated pathway (12, 31).
We have previously demonstrated that macrophages with IL-4 and dexamethasone-induced surface expression of TGF-βRII respond to TGF-β1 through the activation of a complex, multistep gene expression program. We found that indeed, TGF-β1 activated in IL-4/dexamethasone stimulated macrophage phosphorylation of Smad2/3, a classical signaling event for TGF-β1. However, expression profiling of TGF-β–activated genes led to somewhat unexpected results showing activation of not only Smad2/3-dependent genes, like Smad7, but also of some genes whose expression was not shown to be Smad2/3 dependent so far, like Id3 (12). We hypothesized that in macrophages, TGF-β1 activates more complex signaling cascades than just the Smad2/3-mediated one.
In this study, we used Id3 as a marker of TGF-β–activated signaling in macrophages and studied the mechanism of its expression regulation. The search for novel, potentially Smad2/3-dependent regulatory elements in the Id3 gene locus allowed for the identification of a novel enhancer located between +4517 and 4662 bp. Although being capable of amplifying the promoter activity by a factor of 10 and working synergistically with a known Id3 enhancer, its sequence did not contain any Smad2/3 binding elements that were confirmed experimentally. Furthermore, we checked whether TGF-β–dependent activation of Id3 in macrophages is maintained by an epigenetic mechanism and demonstrated that inhibition of histone deacetylases has an inhibitory effect on Id3 expression. Because the activation of Id3 in endothelial cells was shown to be dependent on Smad1/5 (typical for BMP stimulation), we next checked whether TGF-β1 activates Smad1/5 in macrophages. Indeed, we were able to demonstrate that on stimulation with TGF-β1 (but not with BMPs), Smad1/5 is phosphorylated in macrophages. Using pharmacologic inhibitors of ALK5, we showed that this receptor is responsible for Smad1/5 activation. Inhibition of Smad1/5 phosphorylation required lower concentrations of the inhibitor and allowed for the testing of the groups of genes that are Smad1/5 dependent in macrophages. In conclusion, we demonstrated that TGF-β1 induces in macrophages not only phosphorylation of Smad2/3, but also Smad1/5, which is typical for BMPs. In addition to ID3, these signaling pathways lead to the activation of the proatherogenic gene HAMP, coding for hepcidin, which alters iron metabolism in macrophages, and the receptor for plasminogen urokinase activator PLAUR, which is associated with plaque instability.
Materials and Methods
Plasmids
To clone the 5′ flanking region of the Id3 gene, we used a clone RZPDB737B108D6 containing the fragment of chromosome 1 comprising the Id3 gene locus. A 1.1-kb fragment of the 5′ region flanking the Id3 gene was amplified using Expand High Fidelity polymerase mix (Roche). The primers used for amplification were F994 5′-ATA GAC GCG TCA ATT TAA GCG GGC TGT GA-3′ and R1072 5′-CAA AAC TCG AGG AAG TCC CGC TAC AGT G-3′. The amplified fragment contained a region from -1090 to +41 related to the transcription start of the Id3 gene identified according to GenBank mRNA sequence with accession number NM_002167 (http://www.ncbi.nlm.nih.gov/nuccore/NM_002167). The obtained PCR fragment was cloned into the pGL3basic reporter vector to obtain pGL3-ID3prom plasmid.
Identified evolutionary conserved regions (ECRs) were amplified using Expand High Fidelity polymerase mix (Roche). The primers for amplification of the ECR1 sequence were F1260 5′-AAA AGG TAC CGA CTT TAA AGG GCC CCA AT-3′ and R1260 5′- GGA CAC GCG TAG GCT AGC TCA GGC CTT CTG C-3′. The primers for amplification of the ECR2 sequence were F1261 5′-CGG AGG ATC CGC CGA GCC GGC CCC GCC TC-3′ and F1261 5′-CGC AGT CGA CGG CTG CGG AGG GAG CCG GAG-3′. The fragment of ECR1 comprising a region from -3177 to -2660 was cloned upstream of the promoter sequence in the pGL3-ID3prom plasmid. The fragment of ECR2 comprising a region from +4517 to +4662 was cloned into the pGL3-ID3prom plasmid downstream of the luciferase coding sequence. The plasmid phRL-TK was used as an internal transfection control.
To generate plasmids for the overexpression of recombinant Smad1, Smad2, Smad3, Smad4, and Smad5, we amplified the coding sequences using primers presented in Table I and cloned into the pcDNA3.1 vector. The sequences of the cloned fragments were verified by sequencing.
Cells and mediators
The isolation and cultivation of human monocytes/macrophages were done as described previously (8, 11). In brief, the cells were purified from individual buffy coats using density gradients followed by CD14+ magnetic cell sorting (Miltenyi Biotech, Bergisch Gladbach, Germany). Macrophages were cultured at 1 × 106 cell/ml in X-vivo 10 serum-free medium (Cambrex, Verviers, Belgium), supplemented with cytokines and/or dexamethasone as indicated. A detailed protocol is available online at: http://www.methods.info/index.html.
Human IL-4, TGF-β1, bone morphogenetic protein (BMP)-2, BMP-4, and BMP-7 were from TEBU Peprotech (Frankfurt am Main, Germany). Cytokines were used at a final concentration of 10 ng/ml or as indicated. Dexamethasone (Sigma-Aldrich, Munich, Germany) was used at a concentration of 1 × 10−7 M.
Inhibitors of histone deacetylases (HDACs), trichostatin A (TSA), MS-275, and Apicidin (all from Sigma-Aldrich), were used at concentrations indicated in the figures. Inhibitors of signaling pathways, SB431542 (inhibitor of ALK4, ALK5, and ALK7), SB203580 (inhibitor of p38), SP600125 (inhibitor of JNK JNK), PD98059 (inhibitor of MAPK kinase, ERK kinase, and MEK kinase), and BML-275 (inhibitor of ALK2, ALK3, and ALK6) were from Sigma-Aldrich. Concentrations used are indicated in the figures.
HepG2 cells were obtained from American Type Culture Collection and cultivated in VLE-RPMI-1640 (Biocrom, Berlin, Germany) supplemented with 10% FBS.
Protein phosphorylation analysis
For the analysis of protein phosphorylation, 1–3 × 106 cells were lysed in 50 μl lysis buffer containing 50 mM Tris HCl, pH 7.4, 1 mM EDTA, 150 mM NaCl, 1% NP-40, 5 mM NaF, 0.25% Na deoxycholate, 2 mM NaVO3, and 1× Complete protein inhibitors (Roche, Mannheim, Germany). The samples were then separated in a 12% PAGE and transferred to a nitrocellulose membrane. For detection, rabbit anti-human Smad2 mAb (catalog 3102; Cell Signaling Technology, Danvers, MA) was used at a 1:1000 dilution and rabbit anti-human phospho-Smad2 mAb (catalog 3108; Cell Signaling Technology) at a 1:500 dilution. As a secondary Ab, anti-rabbit IgG HRP-linked whole Ab (GE Healthcare) was used at a dilution of 1:5000. Chemiluminescence detection was performed using SuperSignal Pico peroxidase substrate (Pierce, Rockford, IL).
Transfection and luciferase assay
Primary monocyte-derived macrophages were transfected using Amaxa Nucleofector. For this, macrophages were cultivated in X-vivo 10 medium in the presence of IL-4 and dexamethasone for 6 d. Before detachment, the plates were placed on ice for 15 min. Detached cells were harvested by centrifugation, and culture medium was saved for cell recovery and further culture. A total of 1.5 × 106 to 2 × 106 cells was transfected using Amaxa Cell Line Nucleofector Kit V (Lonza, Basel, Switzerland) and program Y-001. For each transfection, 0.75 μg reporter plasmid and 0.25 μg internal control plasmid were used. The macrophages were resuspended in saved culture medium for recovery. Cells were allowed to recover for 24 h before further stimulation.
HepG2 cells were transfected using Lipofectamine reagent (Invitrogen) according to the recommendations of the manufacturer. Transfection was carried out in six-well plates. For each transfection, 3 μg reporter plasmid and 1 μg internal control plasmid was taken. After 3 d of transfection, the cells were harvested and lysed in the passive lysis buffer provided with the luciferase assay kit. Luciferase and Renilla luciferase were measured using the Dual Luciferase assay kit (Promega). Luminescence measurements were performed with Lumoscan (Thermo Labsystems, Frankfurt, Germany).
Small interfering RNA transfection
The following small interfering RNAs (siRNAs) were used to knock down Smad1 and Smad5: SMAD1HSS106247 (Smad1.1; Invitrogen) for Smad1, SMAD5HSS106258 (Smad5.3) (Invitrogen) for Smad5, and Stealth RNAi siRNA Negative Control Lo GC (LO GC; Invitrogen) as a nonspecific negative control. HepG2 cells (2 × 105 cells) were seeded in six-well plates and transfected on the same day. Transfection complexes were prepared as follows: 300 ng siRNA was mixed with 100 μl Opti-MEM I Reduced-Serum Medium (Life Technologies) in Eppendorf tubes by pipetting; 12 μl HiPerFect Transfection Reagent (Qiagen) was added to the tubes, without touching the walls, and mixed by vortexing. siRNA/transfection reagent complexes were allowed to form for 10–15 min at room temperature. The complexes were added dropwise to the six-well plates and the plates were gently swirled to mix. The cells were incubated for 48–72 h at normal culturing conditions and were directly used for Western blotting (72 h) or used for transfection with reporter plasmids for the luciferase assay (48 h). The efficiency of knockdown was confirmed using Western blotting with the following Abs: Smad5 Ab (catalog 9517; Cell Signaling Technology) used at a 1:1000 dilution and Smad1 Ab (catalog no. 9743; Cell Signaling Technology) at a 1:1000 dilution. GAPDH rabbit mAb (HRP Conjugate; Cell Signaling Technology) was used at a dilution of 1:1000 as a loading control. As a secondary Ab, anti-rabbit IgG, HRP-linked Ab (catalog no. 7074S; Cell Signaling Technology) was used at a dilution of 1:2000. Western blotting was performed according to the manufacturer’s protocol (Cell Signaling). Chemiluminescence detection was performed using SuperSignal Pico peroxidase substrate (Pierce, Rockford, IL). ImageJ 1.48 software (available at: http://imagej.nih.gov/ij/) was used to assess the relative OD of Smad1 and Smad5 protein bands. Normalization was performed against densitometry of GAPDH bands to adjust for variations in loading.
Knockdown of Smad1 and Smad5 in macrophages
Macrophages stimulated with IL-4 and IL-4/Dex were transfected after 3 d using the siRNAs Smad1.1, Smad5.3, and LO GC. The best ratio of siRNA to transfection reagent was 600 ng siRNA with 24 μl HiPerFect Transfection Reagent for a six-well plate. For each donor, a well with 4 × 106 cells in 4 ml medium was transfected using specific siRNA for Smad1, specific siRNA for Smad5, and an unspecific negative control. The macrophages were stimulated with TGF-β on day 6 for 3 h and collected for analysis. Samples for quantitative RT-PCR (qRT-PCR) were taken for RNA isolation using the E.Z.N.A. Total RNA Kit I (Omega Bio-Tek), and cDNA was synthesized using the RevertAid First Strand cDNA Synthesis Kit (Fermentas) as described in this article. The final stimulations for each donor were as follows: IL-4/Dex and IL-4/Dex/TGF-β. The stimulations and knockdowns were performed in duplicate for each donor (one experiment for Western blotting and one experiment for qRT-PCR). Western blotting was used to confirm successful siRNA knockdown, whereas qRT-PCR was used to determine changes in ID3 expression. Analysis was done as described earlier in this work.
Real-time RT-PCR analysis
RNA isolation was performed using the RNeasy Mini kit (Qiagen, Hilden, Germany). For first-strand cDNA synthesis, 500 ng total RNA was treated with 2 U RNase-free DNase (Fermentas, St. Leon-Rot, Germany). DNase-treated RNA was used for reverse transcription with the RevertAid First Strand cDNA Synthesis Kit (Fermentas) using oligo dT primers. Real-time qRT-PCR analysis was performed using TaqMan Universal PCR master mix (Applied Biosystems, Darmstadt, Germany). qRT-PCR analysis was performed using the primers listed in Table II (all from Eurofins MWG GmbH, Ebersberg, Germany). For normalization, the expression of GAPD mRNA was measured. The experiments were performed using a Stratagene Mx3005 instrument. The following program was used: 95°C 10 min; 50 cycles: 95°C 10 s, 60°C 60 s.
Results
We have previously demonstrated that in mature macrophage, the expression of Id3 mRNA is activated by TGF-β1 stimulation (12). Bioinformatics analysis of the Id3 promoter showed, however, that it does not contain the Smad binding elements (SBEs) necessary for the activation of gene expression by Smad2/3. Because enhancers were previously described for Id3 (32), we searched for evolutionally conserved regions that may function as enhancers and enable Smad2/3-dependent upregulation of Id3 in macrophages. Using ECR browser (http://ecrbrowser.dcode.org/), we found two conserved regions, ECR1 located between -3177 and -2660 bp upstream of transcription start and ECR2 located between +4517 and 4662 bp downstream of the gene (sequences provided in Supplemental Table 1). ECR1 overlaps with the enhancer described by Shepherd et al. (32), whereas ECR2 has not been described to date.
To test the impact of identified ECRs on the regulation of Id3 expression, we cloned them into a pGL3basic vector together with about a 1-kb fragment of the Id3 promoter (Fig. 1A). Obtained plasmids were transfected using nucleofection into primary monocyte-derived macrophages cultivated in the presence of IL-4 and dexamethasone for 6 d. Cells after transfection were further cultivated for 24 h and then stimulated with TGF-β1 for the next 24 h. Cells were harvested and luciferase activity was measured. Without TGF-β1 stimulation, all reporter constructs showed low promoter activity. Stimulation of macrophages with TGF-β1 led to a 2-fold upregulation of the promoter activity. Addition of ECR1 or ECR2 led to a 5-fold increase of promoter activity. Reporter plasmid containing both of them showed nearly 12-fold upregulation of promoter activity, indicating that ECRs have an additive effect (Fig. 1B).
Mapping of regulatory regions of Id3 gene. (A) Structure of Id3 genomic loci indicating enhancers located within ECR1 and ECR2. (B) Luciferase reporter gene assay of TGBβ1-induced Id3 promoter in primary monocyte-derived macrophages; the results are the average of three independent experiments. (C) Luciferase reporter gene assay of the Id3 promoter and enhancers activities in HepG2 cells, n = 3. (D) Luciferase reporter gene assay of TGF-β1–induced Id3 promoter and enhancer activities in HepG2 cells, n = 3. *p < 0.05, **p < 0.01, ***p < 0.001.
For more detailed analysis of promoter activity, HepG2 cells were used. Reporter plasmids were transfected into HepG2 cells, and luciferase activity was measured 3 d after transfection. It was found that the addition of ECR1 increased the activity of the promoter by a factor of 3.5, which corresponds to published results (32). The addition of ECR2 alone led to a 7-fold upregulation of promoter activity. When used in combination, ECR1 and ECR2 showed an additive effect leading to almost a 10-fold upregulation of promoter activity (Fig. 1C). Obtained results indicate that in HepG2 cells, the effects of ECRs on promoter activity is similar to that in primary macrophages. However, the effect of TGF-β1 stimulation is less pronounced.
We next tested whether the activity of the promoter alone or in combination with ECRs may be increased by TGF-β stimulation. For this, HepG2 cells were transfected with reporter constructs and, after 5 h of transfection, stimulated with TGF-β1. To control the efficiency of TGF-β signaling activation, we transfected HepG2 cells with the pGL-CAGAluc reporter plasmid. It was found that stimulation of transfected cells with TGF-β1 leads to a slight increase of promoter activity. The addition of ECRs did not lead to an amplification of the effect of TGF-β (Fig. 1D). We concluded that neither ECR contains SBEs.
To identify which Smads are involved in the regulation of Id3 expression, we performed cotransfection experiments. The reporter assay preformed in HepG2 cells using pGL3-ID3promECR1ECR2 revealed that cotransfection of expression constructs for Smad2, Smad3, or their combination leads to a decrease in activity of the Id3 promoter (Fig. 2A), whereas cotransfection of Smad1- and Smad5-expressing constructs increases the promoter activity (Fig. 2B). It is important to note that cotransfection of Smad1 and Smad5 together showed the strongest effect on Id3 promoter activity. To further confirm the role of Smad1 and Smad5 in the regulation of Id3 expression, we used knockdown of Smad1 and Smad5 with siRNA. HepG2 cells were transfected with Smad1 and Smad5 siRNA or their combination. This induced significant knockdown of expression of both proteins (Fig. 2E). Cells with knocked down Smad1 and Smad5 were transfected with pGL3-ID3prom and pGL3-ID3promECR1ECR2. Luciferase analysis has shown that knockdown of either Smad1 or Smad5 or their combination leads to ∼40% reduction of Id3 promoter activity. In primary human monocyte-derived macrophages, knockdown of Smad1 and Smad5 was less efficient than in HepG2 cells. However, the expression of Id3 mRNA in macrophages with Smad1 or Smad5 knockdown was reduced to ∼30%. These data indicate that the Id3 promoter is regulated by the Smad1/5 pathway.
Activity of regulatory sequences of Id3 gene is upregulated by Smad1 and Smad5. Regulatory sequences of Id3 gene were analyzed using luciferase reporter assay in cotransfection experiments. (A) Luciferase activity of pGL3-ID3prom ECR1 ECR2 plasmid cotransfected with plasmids containing TGF-β–specific Smad2 and Smad3. (B) Luciferase activity of pGL3-ID3prom ECR1 ECR2 plasmid cotransfected with plasmids containing BMP-specific Smad1 and Smad5. All experiments were performed three times in duplicates. (C) Luciferase activity of pGL3-ID3prom ECR1 ECR2 in HepG2 cells with knockdown of Smad1 and Smad5, n = 4. (D) Luciferase activity of pGL3-ID3prom in HepG2 cells with knockdown of Smad1 and Smad5, n = 4. (E) Western blot analysis of Smad1 and Smad5 knockdown by siRNA in HepG2 cells, diagram represents relative OD of the bands. (F and G) Real-time RT-PCR analysis of Id3 mRNA expression in macrophages with knockdown of Smad1 (F) or Smad5 (G), n = 3. (H) Western blot analysis of Smad1 and Smad5 knockdown by siRNA in primary human macrophages; diagram represents relative OD of the bands. *p < 0.05, **p < 0.01, ***p < 0.001.
Epigenetic mechanisms involved in Id3 expression regulation
Changes in chromatin structure are involved in the regulation of gene expression in macrophages in response to various stimuli (33–35). To assess the effect of chromatin structure on Id3 gene expression, we used one pan–-HDAC inhibitor (TSA) and two selective HDAC inhibitors: MS-275, an inhibitor of HDAC1, and Apicidin, an inhibitor of HDAC2 and HDAC3.
Macrophages stimulated with IL-4 and Dex were pretreated for 1 h with HDAC inhibitors used at different concentrations. After 1-h pretreatment, the cells were stimulated with TGF-β1 for 3 h. Cells were harvested and used for RNA isolation. The level of Id3 mRNA expression was measured using real-time PCR (Fig. 3). It was found that TSA used at a concentration of ≥1 μM reduced the expression of Id3 mRNA by a factor of 4.5 (Fig. 3A). Similar inhibition was observed when Apicidin was used (Fig. 3B). The inhibitor of HDAC1 did not show any significant effect on Id3 expression (Fig. 3C).
Id3 mRNA expression is suppressed by histone deacetylases. Real-time RT-PCR analysis of the effects of histone deacetylase inhibitors TSA (A), Apicidin (B), and MS275 (C) on Id3 mRNA expression induced by TGF-β in primary human monocyte-derived macrophages. Macrophages were differentiated in the presence of IL-4 and dexamethasone for 6 d, pretreated with inhibitors, and then stimulated with TGF-β1 for 3 h. Data in (A) and (B) are representative of three independent experiments; data in (C) are average of nine independent experiments. The p value represents paired comparison of Id3 gene activity with and without HDAC inhibitor induction in TGF-β1–stimulated MφIL-4/Dex (paired Student t test). *p < 0.05.
Obtained data confirm the importance of HDAC for the activation of Id3 expression; however, the observed effect does not explain the activation of Id3 in macrophages, stimulated by TGF-β.
TGF-β activates BMP-specific signaling in macrophages
Next, we analyzed which signaling cascades are activated in macrophages upon stimulation with TGF-β1 and/or BMP-2, BMP-4, and BMP-7. Macrophages cultivated in the presence of IL-4 and dexamethasone for 6 d were treated with TGF-β1 and BMPs or their combinations for 3 h. After incubation, the cells were harvested and lysed in a lysis buffer supplemented with phosphatase and protease inhibitors. Amounts of phosphorylated Smad 2/3 (pSmad2/3), Smad1/5 (pSmad1/5), as well as total Smad2 and total Smad1 were analyzed using Western blot (Fig. 4A).
TGF-β activates Smad1 and Smad5 in primary human monocyte-derived macrophages via ALK5. (A) Western blot analysis of Smad phosphorylation upon stimulation of primary human monocyte-derived macrophages with TGF-β1, BMPs, and their combinations. Macrophages were differentiated in the presence of IL-4 and dexamethasone for 6 d and stimulated as indicated for 3 h. Experiment was repeated at least three times with similar results. (B) Western blot analysis of the effects of ALK5 inhibitor SB431542 on activation of Smad1/5 and Smad2/3. Macrophages were differentiated in the presence of IL-4 and dexamethasone for 6 d, pretreated with the inhibitor for 1 h, and then stimulated with TGF-β for 3 h. Experiment was repeated three times with similar result. (C) Western blot analysis of the influence of PD98059 used in range of concentration from 25 to 100 μM on Smad1/5 phosphorylation. MφIL-4/Dex were treated with PD 98059 for 1 h alone and then stimulated for 3 h with TGF-β1. Experiment was repeated three times with similar result. (D) Luciferase activity of pGL3-ID3prom ECR1 ECR2 plasmid with BRE1 and BRE2 deletions in HepG2 cells, n = 3. (E) Structure of pGL3-ID3prom ECR1 ECR2 plasmid with BRE1 and BRE2 deletions.
Phosphorylation of both Smad2/3 and Smad1/5 was found under the stimulation of macrophages with TGF-β1. Among BMPs tested, only BMP-4 induced moderate phosphorylation of Smad2/3, but not Smad1/5. Stimulation of macrophages with BMP-2 or BMP-7 failed to induce either Smad1/5 or Smad2/3 phosphorylation (Fig. 4A).
To test whether newly identified enhancer in ECR2 contains BMP response elements (BRE), we analyzed its sequence using Mat inspector web application (http://www.genomatix.de). Two BREs were identified: BRE1, CTGGCGCCCGG between bp 95 and 105; and BRE2, TCTGGCTCCGG between bp 127 and 137 (1-st bp of ECR2 was taken as 1). For the analysis of the impact of identified sites on the activity of the enhancer reporter constructs with deleted BRE1, BRE2 and both sites together were created (Fig. 4E). Reporter assay in HepG2 cells showed that deletion of BRE1 leads to a 50% reduction of the promoter activity, whereas deletion of BRE2 has no effect (Fig. 4D).
To identify which receptors are responsible for the activation of Smad1/5 upon stimulation of macrophages with TGF-β1, we inhibited type I TGF-β receptors ALK4/5/7 using pharmacologic inhibitor SB431542. Macrophages cultivated in the presence of IL-4 and dexamethasone for 6 d were treated with different concentrations of SB431542 for 1 h and then stimulated with TGF-β1 for 3 h. Phosphorylation of Smad1/5 and Smad2/3 was analyzed using Western blot. It was found that SB431542 used at a concentration of 1 μM significantly inhibited phosphorylation of Smad1/5 and totally abolished it when used at a concentration of 5 μM (Fig. 4B). In contrast, phosphorylation of Smad2/3 was reduced when SB431542 was used at concentrations from 5 to 25 μM (Fig. 4B). Complete inhibition of Smad2/3 signaling was not reached within the concentration range tested, which allows for the dissection of Smad1/5 and Smad2/3 signaling pathways using different concentrations of the inhibitor.
Because noncanonical activation of Smad1/5 by TGF-β1 via Erk kinase was described previously (36), we further tested whether inhibition of Erk will influence Smad1/5 phosphorylation. Macrophages were pretreated with Erk inhibitor PD98059 for 1 h and stimulated with TGF-β1 for 3 h. Western blotting showed no effect of this inhibitor on phosphorylation of Smad1/5 (Fig. 4C).
It was concluded that in mature macrophages, TGF-β1 activates Smad1/5-mediated signaling via TGF-β type I receptors and does not activate Smad-independent alternative activation signaling pathways.
Dissecting Smad1/5- and Smad2/3-dependent TGF-β–inducible genes in macrophages
SB431542, an inhibitor of ALK4/5/7, provides a unique possibility to dissect Smad1/5- and Smad2/3-dependent TGF-β–induced signaling in macrophages. The fact that SB431542 used at a concentration of 5 μM entirely inhibits Smad1/5 phosphorylation, whereas leaving sufficient Smad2/3 phosphorylation was used for the design of the experiment. Macrophages were cultivated for 6 d in the presence of IL-4 and dexamethasone and then treated for 1 h with SB431542 at a concentration of 5 μM. The cells were then stimulated for 3 h with TGF-β1. The cells were harvested and used for RNA isolation. The following genes that showed significant upregulation after 3 h of stimulation with TGF-β (12) were selected for the analysis: Id3, HAMP, SLC7A5, Smad7, PLAUR, and TMCC3.
Using real-time PCR, we demonstrated that the expression of ID3, HAMP, and PLAUR mRNA was reduced nearly to background level upon inhibition of Smad1/5 signaling. In contrast, only a partial reduction in mRNA expression was observed in the case of TMCC3, Smad7, and SLC7A5 (Fig. 5), which corresponds to partial inhibition of Smad2/3-mediated signaling (Fig. 4B). Thus, we found that ID3, HAMP, and PLAUR are indicative genes of the TGF-β–activated Smad1/5 signaling in mature macrophages.
ALK5 inhibitor discriminates between TGF-β–induced, Smad1/5-, and Smad2/3-dependent genes in macrophages. Real-time RT-PCR analysis of Smad1/5- and Smad2/3-dependent genes in macrophages treated with low concentration of ALK5 inhibitor SB431542. Primary human monocyte-derived macrophages were differentiated in the presence of IL-4 and dexamethasone for 6 d, pretreated with the inhibitor, and stimulated with TGF-β1 for 3 h, n = 3, *p < 0.05, **p < 0.01.
Discussion
In this study, we demonstrated for the first time, to our knowledge, that in macrophages TGF-β activates not only Smad2/3-mediated signaling, but also Smad1/5 signaling rather typical for BMPs. TGF-β–induced signaling in macrophages was studied previously in various systems involving cell lines and primary cells. However, these studies were limited to the investigation of canonical Smad2/3-mediated signaling. It was shown that Smad3 is essential for downregulation of MHC class II in astrocytes in a mouse model (37) and of CD163 in primary human macrophages and macrophage cell lines (31). Our previous study demonstrated activation of Smad2 and Smad3 in mature primary human monocyte-derived macrophages (12).
As observed by us, activation of Id3 expression by TGF-β in macrophages (12) provided a unique possibility to study TGF-β signaling in macrophages. First, the study of regulatory elements of the Id3 gene was performed. It is known for Smad-dependent genes that SBEs may be located far away from the basic promoter region. Thus, an SBE-containing enhancer of the Foxp3 gene is located 2.1 kb upstream from the transcription start site (38). We used analysis of the ECR to identify novel regulatory elements in the Id3 locus. Out of identified candidate ECRs, the upstream one was reported before and was shown to be Smad1/5 dependent (32). The second one, located 4.5 kb downstream of the promoter region, was, to our knowledge, identified for the first time by us. A bioinformatics search demonstrated absence of SBEs in this enhancer that was confirmed experimentally using the luciferase assay. However, its impact on general Id3 promoter activity in HepG2 was ∼2-fold stronger than that of the upstream enhancer. These results suggest either the involvement of epigenetic mechanisms or the Smad2/3-independent, TGF-β–activated signaling pathway in Id3 expression regulation.
Analysis of the effect of HDACs revealed that they do not affect direct Id3 expression regulation by TGF-β. The observed selective inhibition suggests an indirect involvement of HDAC2 or HDAC3, but not HDAC1. The effect of HDAC inhibitors may be explained if their effect on regulatory proteins other than histones is considered. In the context of TGF-β–induced signaling, particularly important are two HDAC substrates: Smad7 and TGF-β type II receptor (39, 40). It was demonstrated that Smad7 interacts with specific HDACs (40) that deacetylate the protein in response to TGF-β signaling (41). The acetylation of Smad7 was shown to be drastically reduced in the presence of HDAC1, HDAC3, and HDAC6 (40).
Tajima and colleagues (42) published that in non-stimulated cells Smad7 is located in the cell nucleus and translocates to the plasma membrane upon TGF-β stimulation. This translocation depends on the Smurf E3-ubiquitin ligases. At the cell membrane, the Smad7–Smurf complex interacts with the activated receptors and inhibits TGF-β signaling by blocking the interaction of receptor-activated Smads and the receptors (43). In addition, Smurf ubiquitinates Smad7 preparing it for degradation (40). Acetylated Smad7 is protected from Smurf-mediated ubiquitination and degradation. Therefore, inhibition of HDACs leads to an accumulation of Smad7 in the cell, which, in turn, potentiates the negative feedback mechanism of TGF-β–induced signaling. Strength of interaction between Smad7 and HDACs taken together with the results of HDAC-mediated inhibition of Id3 expression suggests that the key role in regulation of Smad7 acetylation in macrophages is played by HDAC3.
In addition to canonical Smad2/3-dependent signaling, noncanonical TGF-β–mediated signaling was reported for various cell types (44). It was shown that TGF-β activates Smad1 and Erk1/2 pathways in fibroblasts (45). Activation of Smad1/5/8 by TGF-β was also reported for C2C12 mouse mesenchymal stem cells and HepG2 human hepatocellular carcinoma cells (46), endothelial cells (47), and smooth muscle cells (48). Interestingly, in the latter study, inhibition of ALK5 alone was sufficient to block both phosphorylation of Smad2/3 and of Smad1/5/8. In our work, we used SB431542 as a specific inhibitor of ALK4/5/7 (49). Observed by us, inhibition of Smad1/5 phosphorylation by SB431542 indicates the necessity of ALK5 for Smad1/5 activation in macrophages, suggesting the model proposed by Goumans et al. (50) for endothelial cells is also valid for macrophages (Fig. 6). In this model, activation of the TGF-βRII/ALK5 complex leads to recruitment of ALK1 that, in turn, activates Smad1/5 phosphorylation. The role of TGF-βRIII–endoglin in this process was established recently (51, 52). This can be also true for macrophages, because we have previously demonstrated that macrophages express endoglin, and that this expression does not depend on the type of activation (12).
Summary of TGF-β–induced signaling in primary human monocyte-derived macrophages.
Activation of the BMP-specific pathway by TGF-β in macrophages can be very important for understanding atherogenesis. Accumulating evidence indicate that TGF-β has both atherogenic and atheroprotective functions. In contrast, genes, activated by Smad1/5-dependent BMP signaling, contribute to plaque instability. Hepcidin (53) alters iron metabolism in macrophages and reduces lipid efflux (54), whereas increased expression of the receptor for plasminogen urokinase activator PLAUR is associated with unstable plaques and plaque rupture (55). We clearly demonstrated in this article that TGF-β alone is sufficient for activation of both Smad2/3- and Smad1/5-dependent signaling in macrophages, whereas macrophages do not respond to BMP. We suggest that the dual role of TGF-β in atherosclerosis can be explained by the activation of these two pathways and that the activation of Smad2/3 is responsible for atheroprotective effects, whereas activation of Smad1/5 leads to expression of proatherogenic genes.
In conclusion, this study of TGF-β–activated signaling in macrophages using regulation of Id3 expression as a model system revealed a novel enhancer sequence of the Id3 gene. It was established that in macrophages, TGF-β activates both TGF-β–specific and BMP-specific signaling cascades that can explain a dual role of TGF-β in atherosclerosis.
Disclosures
The authors have no financial conflicts of interest.
Footnotes
↵2 J.K. and A.G. share senior authorship.
This work was supported by Deutsche Forschungsgemeinschaft Grants GRK880/3 (to J.K. and S.G.) and GRK1874 (to J.K. and A.G.), a Bundesministerium für Bildung und Forschung, project BIO-IN (to J.K.), Deutscher Akademischer Austauschdienst (to K.A.), and a European Commission Seventh Framework Programme IMMODGEL project (to J.K.). Revision of this paper was supported by Russian Science Foundation Projects 14-15-00396 (to A.G.) and 14-15-00350 (to J.K.), and the Tomsk State University Competitiveness Improvement Program.
The online version of this article contains supplemental material.
Abbreviations used in this article:
- BMP
- bone morphogenetic protein
- BRE
- BMP response element
- ECR
- evolutionary conserved region
- HDAC
- histone deacetylase
- qRT-PCR
- quantitative RT-PCR
- SBE
- Smad binding element
- siRNA
- small interfering RNA
- TSA
- trichostatin A.
- Received January 28, 2013.
- Accepted November 13, 2014.
- Copyright © 2015 by The American Association of Immunologists, Inc.
References
In this issue
