Obesity is associated with macrophage accumulation and inflammation in adipose tissue. Macrophage-secreted factors have been reported to inhibit the differentiation of preadipocytes into adipocytes and to modulate adipogenic extracellular matrix gene expression. To enlarge our understanding of macrophages and the scope of their interactions with preadipocytes, we investigated their effect on preadipocyte survival. Acute exposure of 3T3-L1 preadipocytes to J774A.1 macrophage-conditioned medium (MacCM) stimulated platelet-derived growth factor receptor (PDGFR) tyrosine phosphorylation by 4.1-fold. There were significant increases in the phosphocontent of downstream PDGFR targets Akt and ERK1/2 (5.3-fold and 2.4-fold, respectively) that were inhibited by PDGF immunoneutralization or by the selective PDGFR inhibitor imatinib. Serum-free J774A.1-MacCM or RAW264.7-MacCM completely prevented 3T3-L1 preadipocyte apoptosis normally induced by serum deprivation. Addition of PDGF alone to serum-free control medium was sufficient to prevent 3T3-L1 preadipocyte apoptosis. Inhibition of PDGFR activation by MacCM, either by addition of imatinib or by PDGF immunodepletion of MacCM, effectively disrupted the prosurvival effect. In summary, our data indicate that MacCM promotes preadipocyte survival in a PDGF-dependent manner.
- platelet-derived growth factor
obesity, defined as an increase in adiposity, results from a chronic positive energy balance. Expansion of adipose tissue occurs through the formation of new adipocytes, via the differentiation of stromal precursor preadipocytes, and the enlargement of existing adipocytes (9). Deficiency in the number of preadipocytes, or their capacity to differentiate, may lead to excessively hypertrophied and dysfunctional adipocytes that overproduce immune cell chemoattractant proteins (7, 13, 14). Concomitantly, the population of adipose tissue macrophages rises due to monocyte infiltration. This abnormal cellular remodeling of adipose tissue is associated with low-grade inflammation and insulin resistance (38, 39).
Paracrine communication among macrophages, adipocytes, and preadipocytes is believed to contribute to obesity-associated adipose tissue dysfunction (12). We and others have reported that macrophages produce anti-adipogenic factors that restrain the differentiation of preadipocytes into adipocytes (4, 5, 15, 17, 33, 40). Macrophage-secreted factors have also been observed to have variable effects on preadipocyte proliferation (17, 23). Apoptosis, the other key preadipocyte fate pertinent to adipose tissue remodeling (11, 21, 28), has not been evaluated with respect to potential macrophage paracrine interactions.
The precise identity of the macrophage products influencing preadipocyte responses is not known. Here, we demonstrate that macrophage-conditioned medium (MacCM) activates platelet-derived growth factor (PDGF) signaling in 3T3-L1 preadipocytes. Our studies indicate that serum-free MacCM prevents apoptosis of preadipocytes undergoing serum deprivation and that this prosurvival effect depends on the presence of PDGF.
MATERIALS AND METHODS
Preparation of MacCM.
J774A.1 (from ATTC, Manassas, VA) and RAW264.7 (from Dr. X. Zha, Ottawa, Canada) macrophages were grown in DMEM supplemented with 10% fetal bovine serum (FBS) and antibiotics (100 U/ml penicillin, 0.1 mg/ml streptomycin). At confluence, the medium was replaced with fresh growth medium, or serum-free medium, and collected 24 h later. MacCM was centrifuged at 200 g for 5 min, and the supernatant was stored at −20°C until use for preadipocyte studies. Control medium (serum-supplemented or serum-free, but not exposed to macrophages) was processed in parallel. Immunodepletion of PDGF was performed by incubating serum-free control medium and serum-free MacCM with 10 μg/ml of neutralizing anti-PDGF antibody (R&D Systems, Minneapolis, MN) or nonspecific goat IgG for 1 h at room temperature. Medium was subsequently incubated with protein A sepharose (∼10 μl/ml medium) for 30 min. Samples were then centrifuged at 3,000 g for 15 min, and supernatants were applied to preadipocyte cultures.
Preadipocyte signal transduction studies.
Low passage 3T3-L1 preadipocytes (ATCC) were grown in DMEM supplemented with 10% calf serum (CS) and antibiotics. Confluent 3T3-L1 preadipocytes were either placed in serum-supplemented MacCM or control medium with or without 20 ng/ml human recombinant PDGF (Calbiochem, Gibbstown, NJ or Millipore, Billerica, MA) for the time periods shown. Where indicated, preadipocytes were pretreated with 10 μM imatinib mesylate (kindly provided by Novartis, Basel, Switzerland) for 90 min. To interfere with PDGF action, MacCM was incubated with 0–20 μg/ml neutralizing anti-PDGF antibody (R&D Systems, Minneapolis, MN) for 1 h at room temperature before addition to preadipocyte cultures. Preadipocyte signaling was also assessed following exposure to serum-free MacCM immunodepleted of PDGF.
Immunoblot analysis of cell lysates and conditioned medium.
After stimulation, cells were lysed in 1× Laemmli buffer (18) supplemented with 5% β-mercaptoethanol, 1 mM sodium orthovanadate, 5 mM EGTA (pH 8.0), 50 mM sodium fluoride, and 5 mM sodium pyrophosphate. Protein was measured using the Dc Protein Assay (Bio-Rad, Hercules, CA) with bovine serum albumin (BSA) as a standard. Equal amounts of solubilized protein (10–50 μg) or equal volumes of immunoprecipitated protein were resolved by SDS-PAGE and transferred to a nitrocellulose membrane. Membranes were incubated for 1 h in 5% skim milk or 3% BSA to block nonspecific binding sites and then probed as indicated with the following primary antibodies directed against: Akt1 (C-20; goat polyclonal; 1:1,000), PDGF receptor (PDGFR)-β (rabbit polyclonal; 1 μg/ml; both from Santa Cruz Biotechnology, Santa Cruz, CA), phosphotyrosine (PY20, mouse monoclonal; 1:1,000; BD Biosciences, Mississauga, ON, Canada), ERK1/2 (1.0 μg/ml; Upstate Biotechnology, Charlottesville, VA), phospho-ERK1/2 (pERK1/2; Thr202/Tyr204; 1:1,000), phospho-Akt (pAkt; Ser473; rabbit polyclonal; 1:1,000; all from Cell Signaling Technology, Beverly, MA). This was followed by incubation with the appropriate horseradish peroxidase (HRP)-conjugated secondary antibody. Signal detection was performed using Immobilon Western chemiluminescence HRP substrate (Millipore). Relative intensity of the bands was quantified using AlphaEaseFC software (Alpha Innotech, San Leandro, CA) and expressed as integrated optical density (IOD) units.
Serum-free medium, control or conditioned by J774A.1 macrophages, was concentrated with a 3-kDa cut-off filter (Millipore). Media were centrifuged at 3,000 g for 6 h. Retentate was collected by inverting the column and centrifuging at 2,000 g for 10 min. The retentate was diluted in an equal volume of 2× Laemmli buffer, and an aliquot was resolved by SDS-PAGE under nonreducing conditions and immunoblotted with anti-PDGF antibody (2 μg/ml, R&D Systems).
Cells were lysed in PBS (pH 7.4) containing 1% Nonidet-40, 50 mM sodium fluoride, 5 mM sodium pyrophosphate, 0.1 mg/ml phenylmethylsulfonylfluoride, 200 mM sodium orthovanadate, 10 μg/ml aprotinin, 10 μg/ml leupeptin, 4 μg/ml benzamidine, and 1 mM β-glycerophosphate. Lysates (∼0.5 mg protein, assayed by Micro BCA protein assay, Pierce, Rockford, IL) were precleared with protein A-sepharose beads and incubated with 1 μg anti-PDGFRβ antibody complexed to protein A-sepharose beads. Immunoprecipitated proteins were washed and solubilized in 1× Laemmli containing 1 mM sodium orthovanadate. Immunoprecipitated proteins were subsequently analyzed by immunoblot analysis as described above.
Analysis of cell death.
Confluent 3T3-L1 preadipocytes were placed in control medium or MacCM containing 10% FBS or serum free. Where indicated, the PDGFR-selective inhibitor imatinib (10 μM) or 0.1% DMSO (vehicle control) was added to these media. After 6 h, floating cells were removed, and adherent cells were trypsinized and stained with 0.2% trypan blue dye. Viable cells (trypan blue exclusion) were counted in duplicate using a Neubauer hemacytometer.
To visualize and quantify apoptotic changes in cell nuclei, confluent 3T3-L1 preadipocytes, grown on coverslips, were exposed to the same media described above for 3 h. Cells were then fixed in 10% formaldehyde for 1 h, followed by staining for 10 min with 1 μg/ml Hoechst 33342 in a humid chamber. Individual coverslips were mounted onto glass slides using Moviol. Cells were visualized and photographed (×400 magnification) with a Zeiss Axioplan 2 microscope equipped with an Axiocam digital camera (Carl Zeiss, Toronto, Canada). Ten random fields were photographed for each of the three coverslips used per treatment. Counts of all stained nuclei and apoptotic nuclei were performed by two independent observers. Data are represented as percent apoptosis, calculated by dividing the total number of apoptotic nuclei by the total number of nuclei, multiplied by 100.
Annexin V binding was assessed using the Vybrant Apoptosis Assay kit no. 2 (Molecular Probes, Eugene, OR). Confluent 3T3-L1 preadipocytes were exposed to the same media for 3 h. Adherent cells were trypsinized and combined with the floating cell population. Cells were centrifuged at 500 g for 5 min, washed, resuspended, and stained with annexin V and propidium iodide (PI) for 15 min, as per the manufacturer's protocol. Annexin V binding and PI staining was quantified with a Beckman Coulter Epics XL-MCL flow cytometer.
Comparison of means was performed by ANOVA (P < 0.05 taken as significant), followed by the post hoc Newman-Keuls test to assess differences between individual means using GraphPad InStat v3.05 (GraphPad Software, San Diego, CA).
Macrophage-secreted products induce PDGFR signaling in 3T3-L1 preadipocytes.
To assess potential deregulated signaling in preadipocytes triggered by macrophage-derived factors, we conducted a time-course stimulation on 3T3-L1 preadipocytes exposed to J774A.1-MacCM. We observed a time-dependent increase in the tyrosine phosphorylation of a 180-kDa protein in preadipocytes exposed acutely (0–60 min) to MacCM (indicated by arrow in Fig. 1A, P < 0.05). To determine whether this might be due to phosphorylation of the PDGFR, we stimulated 3T3-L1 preadipocytes for 5 min with 20 ng/ml PDGF, J774A.1-MacCM, or control medium. Solubilized proteins were immunoprecipitated with an anti-PDGFRβ antibody, followed by antiphosphotyrosine immunoblot analysis. A 4.1-fold increase in PDGFR phosphorylation occurred upon treatment with MacCM (Fig. 1B, P < 0.05, MacCM vs. control).
We next investigated whether MacCM was capable of activating signaling molecules downstream from the PDGFR. A 15-min stimulation of 3T3-L1 preadipocytes was performed with either control medium, medium supplemented with PDGF, or MacCM, in the presence or absence of the PDGFR-selective inhibitor imatinib (3). MacCM stimulated the tyrosine phosphorylation of the PDGFR (Fig. 2A, P < 0.001) and phosphorylation of Akt and ERK1/2 (Fig. 2A, 5.3-fold increase in pAkt P < 0.05, 2.4-fold increase in pERK1/2, P < 0.05, MacCM vs. control). PDGF treatment produced similar results (Fig. 2A). The phosphorylation of Akt and of ERK1/2 by MacCM was inhibited by pretreatment with imatinib, indicating that this was a PDGF-dependent response (Fig. 2A). Under control conditions, Akt phosphorylation was not detected, and the basal level of ERK1/2 phosphorylation appeared to be independent of PDGFR activation, because it was not inhibited by imatinib (Fig. 2A).
To confirm, by another strategy, that acute MacCM-stimulated phosphorylation of Akt and ERK1/2 was mediated through PDGFR activation, we incubated MacCM with increasing concentrations of PDGF neutralizing antibody before exposure to 3T3-L1 preadipocytes. PDGF immunoneutralization of MacCM significantly decreased PDGFR phosphorylation (Fig. 2B), and this was accompanied by a significant inhibition of Akt and ERK1/2 phosphorylation (Fig. 2B).
MacCM-induced PDGFR signaling protects 3T3-L1 preadipocytes from apoptosis.
Since Akt and ERK1/2 signaling pathways have been implicated in cell survival, we therefore investigated whether MacCM-induced signaling could protect preadipocytes from cell death. MacCM was generated from J774A.1 or RAW 264.7 macrophages under serum-free conditions. Serum-free MacCM was then compared with serum-supplemented MacCM and serum-supplemented control medium with respect to their ability to preserve cell viability upon serum deprivation (growth factor withdrawal) of 3T3-L1 preadipocytes. The apoptotic response of 3T3-L1 preadipocytes to serum withdrawal has been reported previously (21, 22, 24). After 6 h of serum withdrawal, we observed a 50% reduction in the total number of viable cells, assessed by the enumeration of adherent cells excluding trypan blue dye (Fig. 3, A and B, control medium with serum vs. without, P < 0.001). In contrast, there was no decrease in the number of viable cells following the same time period of serum withdrawal when in the presence of serum-free MacCM from either macrophage cell model. Hoechst staining of nuclei following 3 h of serum deprivation confirmed this prosurvival effect (Fig. 3C). To further demonstrate the anti-apoptotic effect of MacCM, 3T3-L1 preadipocytes were stained with both annexin V and PI. A significant increase in annexin V staining was observed following serum withdrawal (Fig. 3D, annexin V+/PI− for serum-containing vs. serum-free control medium, P < 0.001). In contrast, there was no increase in annexin V staining for cells placed in serum-free MacCM. PI staining, an indicator of necrosis, was minimal and not significantly changed with any of the treatments, suggesting 3T3-L1 preadipocytes were strictly undergoing apoptosis when serum deprived.
To determine whether J774A.1-MacCM contained detectable amounts of PDGF, serum-free control medium or MacCM was immunoblotted with anti-PDGF antibody; human recombinant PDGF served as a standard. Bands (30–36 kDa) were detected in the PDGF and MacCM lanes (Fig. 4A). The slight mobility difference observed here has been reported previously by others comparing PDGF in conditioned medium to recombinant PDGF (8).
PDGF is known to promote cell survival (35). To assess whether PDGF alone is sufficient to protect 3T3-L1 preadipocytes from apoptosis induced by growth factor withdrawal, preadipocytes were placed in serum-free medium supplemented with increasing concentrations of recombinant PDGF. A significant and dose-dependent protection from serum withdrawal-induced cell loss was observed with concentrations of PDGF as low as 1 ng/ml (Fig. 4B, P < 0.05, PDGF vs. serum free).
To determine whether MacCM-mediated preadipocyte PDGFR activation was required for the prosurvival effect, 3T3-L1 preadipocytes were placed in serum-free MacCM in the presence of imatinib. As before, serum withdrawal in the absence of macrophage-secreted factors resulted in a 50% reduction in the number of viable cells (Fig. 4C, P < 0.01, serum-containing vs. serum-free control medium). The protection by macrophage-derived factors was lost in the presence of imatinib (Fig. 4C, P < 0.01, serum-free MacCM vs. serum-free MacCM + imatinib). Imatinib, added to serum-supplemented control medium, did not result in any 3T3-L1 preadipocyte death, and it did not aggravate cell loss when added to serum-free control medium (Fig. 4D). Hoechst staining of nuclei further indicated that imatinib blocked MacCM-dependent cell survival (Fig. 4E, P < 0.001).
We next sought to confirm the importance of PDGF signaling in preadipocyte survival by using PDGF-immunodepleted MacCM. After immunodepletion (Fig. 5A), MacCM was no longer capable of stimulating PDGFR tyrosine phosphorylation (Fig. 5B, P < 0.01). Incubation of serum-free MacCM with nonspecific IgG did not alter the ability of MacCM to completely prevent 3T3-L1 preadipocyte cell loss (Fig. 5C). In contrast, serum deprivation of preadipocytes in the presence of PDGF-immunodepleted MacCM led to 30% cell death (Fig. 5C, serum-free MacCM treated with IgG vs. αPDGF, P < 0.001). Additionally, there were significantly more apoptotic nuclei observed when serum deprivation was performed in PDGF-depleted MacCM compared with MacCM (Fig. 5D, P < 0.001).
Our studies demonstrate that MacCM contains PDGF and activates PDGFR signaling pathways in 3T3-L1 preadipocytes. We have shown that MacCM prevents 3T3-L1 preadipocyte apoptosis induced by growth factor withdrawal, and PDGF is sufficient and necessary for this effect. This novel interaction between macrophages and preadipocytes hints at the growing complexity of adipose tissue remodeling.
The recognition that obesity promotes an increase in the number of adipose tissue macrophages has initiated a new way of understanding adipose tissue inflammation and insulin resistance (31, 38, 39). Several potential roles for this population of macrophages have been proposed, such as direct release of pro-inflammatory cytokines into the circulation or inhibition of mature adipocyte insulin-regulated glucose and lipid metabolism (20, 27, 34).
Regulation of preadipocyte fate with respect to adipose tissue remodeling and expansion is thought to be important to maintain adipose tissue function and result in the metabolically healthy obese phenotype (13). Inhibition of adipogenic capacity and the resulting decreased formation of new adipocytes (hyperplasia) would predispose to exaggerated adipocyte hypertrophy and dysfunction (6, 13). Paracrine communication between macrophages and stromal progenitor preadipocytes has been proposed by us and others (4, 17, 23). The nature of such interactions may alter the functionality of adipose tissue. We have reported that MacCM from mouse J774 and human THP-1 macrophage cell lines inhibit the differentiation of mouse 3T3-L1 and human primary preadipocytes into adipocytes (4, 5, 40). Conditioned medium from human monocyte-derived macrophages and from human adipose tissue macrophages is similarly anti-adipogenic (15, 17). A recent report has also demonstrated the same phenomenon in a mouse model (33).
The effects of MacCM on preadipocyte proliferation have also been studied with both pro- and anti-proliferative effects described (15, 17, 23). The reasons for the discrepancy are not clear at present, but progress in this area is needed, as proliferation capacity would be expected to exert an important influence on the size of the preadipocyte pool in the stromal fraction.
Another influential process on preadipocyte number is apoptotic susceptibility, but there are no reports in the literature examining macrophage-secreted factors on preadipocyte survival. It has been shown, by us and others, that 3T3-L1 preadipocytes are susceptible to apoptosis induced by serum withdrawal (21, 22, 24). Therefore, we addressed this question using the established and widely used mouse J774A.1 macrophage, RAW264.7 macrophage, and 3T3-L1 preadipocyte cell line models. Our data show a clear, consistent, and potent anti-apoptotic effect of MacCM on serum-deprived preadipocytes. Although macrophages are known to produce many pro-inflammatory factors that might have been predicted to favor cell death, the survival effect we saw indicated to us that a macrophage-secreted growth factor might be implicated.
Our acute signaling preadipocyte studies with MacCM revealed that the PDGFR was activated, including stimulation of the classical downstream kinases, Akt and ERK1/2. We also demonstrated the presence of PDGF in MacCM, and that immuno-based neutralization of PDGF in the MacCM rendered it incapable of PDGFR activation.
PDGF activation of Akt, nuclear factor (NF)κB, and ERK1/2 has been shown to mediate its pro-survival activity in other cell types (1, 30, 37). Signaling through these pathways has been shown to regulate caspase activity and the expression of Bcl-2 family members (35). The effect of MacCM on these intracellular targets in serum-deprived preadipocytes remains to be identified.
PDGF gene expression has been shown to be elevated in alternatively activated macrophages (32), whereas treatment of macrophages with the Th1 cytokine, IFNγ, suppresses PDGF expression (16). Hypoxia, which is a feature of the obese adipose tissue (36), has also been shown to increase PDGF expression in macrophages (25).
Our study is limited to the use of macrophage and preadipocyte cell lines that are established as in vitro models for inflammation and metabolic research. It should be noted that macrophages are highly influenced by the local environment in vivo, with different classes of cytokines resulting in pro-inflammatory (classic) versus anti-inflammatory (alternate) states of macrophage activation (10). Indeed, some investigators describe adipose tissue macrophages in intermediate states between these two extremes (2, 19, 20). Furthermore, human preadipocytes have been reported to be less susceptible to apoptosis induced by serum deprivation in vitro (26). It will therefore be important to conduct similar experiments with human primary cell cultures to explore the impact of human macrophages on human preadipocyte apoptosis. As discussed above, it should be noted that adipogenesis studies using either cell lines or primary cell cultures have been very consistent.
To date, identifying the role of specific secreted macrophage products for any preadipocyte response has been elusive. Therefore, it is significant that we have discovered that PDGF is the critical factor for this newly described anti-apoptotic effect of MacCM. Our data show that PDGF is clearly both sufficient and required for the anti-apoptotic effect. We have demonstrated that blocking PDGF receptor activation by imatinib interfered with the ability of MacCM to prevent preadipocyte apoptosis. Imatinib, though selective for the PDGF receptor kinase, is also known to inhibit c-Abl and c-Kit (3). Neither of these two kinases have been implicated in preadipocyte survival responses. In any case, to more precisely implicate PDGFR, we immunodepleted the MacCM of PDGF by immunoprecipitation with PDGF antibody, an approach used by others (29). PDGF-immunodepleted MacCM was unable to stimulate tyrosine phosphorylation acutely and could no longer sustain preadipocytes subjected to serum withdrawal over the 6-h time period.
In conclusion, our findings suggest a novel role for macrophages in protecting preadipocytes from death. Furthermore, we have identified PDGF as a factor that is sufficient and necessary for this effect. These results add to a growing body of knowledge concerning macrophage-preadipocyte interactions and adipose tissue remodeling.
This work was supported by grant MOP-43850 from the Canadian Institutes for Health Research (to A. Sorisky) and a Heart and Stroke Foundation of Ontario Master's Studentship (to A. S. D. Molgat).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2009 the American Physiological Society