Selective Induction of Cell Death in Human M1 Macrophages by Smac Mimetics Is Mediated by cIAP-2 and RIPK-1/3 through the Activation of mTORC

This article is featured in Top Reads, p.

Macrophages are generally described as either “classically activated” M1 macrophages or “alternatively activated” M2 macrophages, which differ in their phenotypic and functional characteristics and represent two extreme ends of the macrophage polarization spectrum (1, 2). M1 macrophages can be generated from monocyte-derived macrophages (MDMs; MDM-M0) by cytokines, such as TNF-α and IFN-γ, or by intracellular pathogens and bacterial LPS (1, 3). They exhibit enhance phagocytic capacity and Ag presentation ability, and secrete high levels of proinflammatory cytokines, such as TNF-α, IL-12, IL-6, and IL-1β, and toxic molecules, such as inducible NO synthase–dependent reactive oxygen intermediates and reactive nitrogen intermediates (1, 3). In contrast, cytokines, such as IL-4, IL-13, IL-10, or TGF-β, induce polarization of M2 macrophage subsets, namely, M2a, M2b and M2c (4, 5). In general, IL-4 and/or IL-13 polarize macrophages toward the M2a subset, whereas the M2b subset can be generated after stimulation with immune complexes and TLR agonists or IL-1 receptor ligands (4). Lastly, the M2c subset can be generated by glucocorticoids and/or IL-10 (4, 6). Macrophage subsets are highly plastic (7–9) and are known to play a role in disease pathogenesis (3, 10, 11). During infection, the tissue microenvironment drives macrophage polarization toward an M1 phenotype to control microbial infection and further induce inflammatory responses (8, 12). After successful clearance, the microenvironment drives macrophages toward an M2 anti-inflammatory stage, which helps in resolving the inflammatory responses and promotes wound healing and tissue repair (8, 9). Excessive inflammation or inability to resolve inflammation may damage the surrounding tissues and promote the development of diseases characterized by chronic infection or autoimmune phenomena. For example, M1 signature has been shown to be associated with human colon cancer, inflammatory bowel disease, rheumatoid arthritis, tuberculoid leprosy, active tuberculosis, and Helicobacter pylori gastritis (13–16). Conversely, M2 polarization is observed in lepromatous leprosy, Whipple’s disease, and chronic rhinosinusitis (16–18).

Macrophages in general are resistant to spontaneous and induced apoptosis that is critical to maintain cell survival and homeostasis (19, 20). The mechanisms governing macrophage resistance to cell death have been attributed to the differential expression of proapoptotic and antiapoptotic genes/proteins, including Bcl-2 and inhibitors of apoptosis proteins (IAPs) family members (20, 21). The IAPs family includes xIAP, cellular IAP (cIAP)-1, and cIAP-2, and they tightly regulate cell survival and immune functions downstream of TNF receptor-1 and other death receptors (22). The second mitochondria-derived activator of caspases (SMAC) mimetics (SMs), the inhibitors of IAP proteins, have been employed to investigate the role of IAP proteins in cell survival and immune regulation (23). SMs are small peptides that mimic the N-terminal four-amino acid sequence of the endogenous SMAC, which bind to XIAP and block cIAP-1/2 activity, resulting in caspase activation and enhanced proteasomal degradation, and eventually repression of antiapoptotic IAPs function and cell death (23).

We have previously characterized human macrophage subsets M0, M1, M2a, M2b, and M2c using a nonexhaustive panel of surface markers and cytokine production (4, 7). We have also shown that polarization of M0 into M1 macrophages rendered them highly susceptible to SM-induced cell death, whereas polarized M2a, M2b, and M2c subsets were resistant (7). Both apoptosis and necroptosis were induced in SM-induced cell death in M1 macrophages. In this article, we extended our investigations to understand the molecular mechanism governing SM-induced cell death during differentiation into M1 macrophages and in polarized M1 macrophages.

The isolation of human primary monocytes and the generation of MDMs (MDM-M0) have been previously described (4, 21). In brief, blood was collected in heparinized syringes and subjected to density gradient centrifugation using Lymphoprep solution (STEMCELL Technologies, Vancouver, BC, Canada) at 1600 rpm for 45 min. The PBMCs were collected from the buffy coat and suspended in serum-free medium (Iscove’s modified DMEM 1×; WISENT, St-Bruno, QC, Canada) at a concentration of 2.5 × 10⁶ cells/ml. The cells were then cultured for 3 h to facilitate the adherence of monocytes, and the nonadherent cells were washed off. The adherent monocytes were cultured for 7 d in complete medium, which consists of Iscove’s modified DMEM 1× supplemented with 10% FBS (Sigma-Aldrich, St. Louis, MO), 1% penicillin and streptomycin (Pen-Strep; R&D Systems, Minneapolis, MN), along with 10 ng/ml human recombinant M-CSF (R&D Systems). On the fourth day, cells were washed and replaced with fresh complete medium. The purity of macrophages was assessed on day 8 by measuring CD14⁺ expression by flow cytometry. M0 macrophages contained less than 2% T cells, B cells, and NK cells. To generate different macrophage subsets, we stimulated M0 macrophages with either IFN-γ (20 ng/ml; Thermo Scientific, Rochester, NY) for 48 h to generate M1 macrophages or with IL-10 (10 ng/ml; R&D Systems) to generate the M2c subset. The cell line SNB-75 was cultured in complete medium and in petri dishes until reaching 90% confluency. Cells were then detached, resuspended in medium, and seeded into 12-well culture plates and incubated overnight to conduct the experiments.

Human macrophage subsets were washed twice with serum-free medium and incubated with different concentrations of SMAC mimetics (SM-LCL161) purchased from Active Biochem (catalog number [Cat. no.] A-1147; Kowloon, Hong Kong) for 48 h. For experiments performed to study the effect during M1 differentiation (prepolarization) or in polarized M1 (postpolarization), the following inhibitors were added for 1 h either before or after IFN-γ stimulation and followed by SM-LCL161 treatment for 48 h. All chemical inhibitors were reconstituted and stored based on the recommendation from the manufacturers. The Jak inhibitor, RUX (Cat. no. A3012; ApexBio, Houston, TX), NF-κB–IκB kinase (IKK) inhibitor (PS-1145; Cat. no. P6624), mammalian target of rapamycin complex (mTORC) inhibitor (rapamycin; Cat. no. R0395), and S6 kinase inhibitor (PF-4708671; Cat. no. PZ0143) were all purchased from Sigma-Aldrich (St. Louis, MO). The protein kinase R (PKR) inhibitor (C16; Cat. no. 527450), PI3K inhibitor (LY294002; Cat. no. 440202), ERK inhibitor (PD-98059; Cat. no. 513000), JNK inhibitor (SP-600125; Cat. no. 420119), and p38 MAPKs inhibitor (SB-203580; Cat. no. 559389) were all purchased from Calbiochem (San Diego, CA). The mTORC inhibitor, torin (Cat. no. 4247), and Akt inhibitor (MK-2206; Cat. no. 11593) were purchased from Tocris Bioscience (Oakville, ON, Canada) and Cayman Chemicals (Ann Arbor, MI), respectively.

Cell death was determined by intracellular propidium iodide (PI) staining as described previously (21). In brief, adherent cells were gently scraped and transferred into new tubes. Cells were centrifuged at 2200 rpm followed by the addition of methanol (350 μl) and incubated for 15 min at 4°C in the dark. A mixture of 25 μl PI solution (Sigma-Aldrich) and 25 μl RNase (Roche Diagnostics, Laval, Québec, Canada) was added to each pellet and incubated for 1 h at 4°C. The PI-stained cells were analyzed using PI channel in Flow cytometer-BD LSRFortessa Cell Analyzer (BD Life Sciences, Mississauga, ON, Canada). The histograms were analyzed using FlowJo Software version 10.0.7 (BD Biosciences, San Jose, CA) and plotted using GraphPad Prism 5.0 software (San Diego, CA).

Kinetic live-cell imaging was carried out with exogenous YOYO-1 iodide (Cat. no. Y3601; Thermo Fisher Scientific) or caspase-3/7 reagent (Cat. no. 4440; ESSEN Bioscience, Ann Arbor, MI) and monitored using IncuCyte Zoom live-cell imaging (ESSEN Bioscience) as described previously (7). The caspase-3/7 reagent directly binds to cells undergoing apoptosis by intrinsic and extrinsic pathways. The instrument allows visualization of the fluorescent signals and quantifies the intensity of signal/cell while the cells are still in culture at 37°C. Cells were washed with serum-free medium, and YOYO-1 iodide was added at a 1 μl:1 ml ratio. Plates were then incubated in IncuCyte Zoom, and images were automatically taken every 2 h for 48 h. Images were collected, processed, and analyzed by IncuCyte Zoom software, and data were presented using GraphPad Prism 5.0 software.

Caspase-8 activity was determined by flow cytometry by using caspase-8 (active) FITC staining kit (Cat. no. AB65614; Abcam, Toronto, ON, Canada) as described previously (7). In brief, cells were treated with SM-LCL161 followed by addition of caspase-8 reagent (FITC-IETD-FMK) for 1 h at 37°C. Caspase-8 activity was assessed by using a BD LSRFortessa X20 flow cytometer and FL-1 channel. The histograms were plotted using FlowJo version 10.0.7 software.

Western blotting was carried out as previously described (21). In brief, 40–50 μg/μl total protein lysates were collected and subjected to 10% SDS-PAGE gel. The gel was transferred to a polyvinylidene difluoride membrane (Bio-Rad Laboratories, Mississauga, ON, Canada). All membranes were blocked with 5% skim milk for 30 min and probed with primary Abs against the following proteins: phospho-Jak1 (Cat. no. 3331), NF-κB p100/p52 (Cat. no. 4882), receptor-interacting protein kinase (RIPK)-1 (Cat. no. 4926/3493), phospho-PKR (Cat. no. 3076), phospho-S6 Kinase (Cat. no. 9234), cIAP-2 (Cat. no. 3103), cIAP-1 (Cat. no. 7065), RIPK-3 (Cat. no. 13526), phospho-Akt (Cat. no. 4056), MLKL (Cat. no. 14993), phospho-JNK (Cat. no. 4668), IFN regulatory factor-1 (IRF-1; Cat. no. 8478), phopho-ERK1/2 (Cat. no. 9101), phospho-P38 (Cat. no. 9215), Mcl-1 (Cat. no. 5453), Fas (Cat. no. 4233), IκBα (Cat. no. 4812/9242), phospho-IKKα/β (Cat. no. 2697), Bcl-xL (Cat. no. 2764), Bcl-2 (Cat. no. 4223), Bax (Cat. no. 5023), caspase-9 (Cat. no. 9502), and TRAIL (Cat. no. 3219). All Abs were purchased from Cell Signaling Technology (Danvers, MA). Membranes were then probed with goat anti-rabbit or anti-mouse secondary Abs IgG conjugated to HRP (Cat. no. 172-1019 and Cat. no. 170-6516; Bio-Rad Laboratories). For total proteins evaluation, membranes were stripped and probed with either anti–β-actin or anti-GAPDH Abs (Cell Signaling Technology). Protein bands were developed using the Amersham ECL Western blotting detection reagent (Cat. no. RPN2235; Cedarlane, Burlington, ON, Canada), visualized using Syngene Chemi Genius 2 Bio Imaging System (Syngene, Frederick, MD), and processed in GeneSnap software, version 6.08 (Syngene). The cIAP-1, cIAP-2, RIPK-1, RIPK-2, and Bax bands were quantified from at least three different samples using ImageJ (National Institutes of Health, Bethesda, MD) and normalized to their respective GAPDH bands.

M0, M1, and M2c supernatants were collected after 48 h of polarization with cytokines or LPS stimulation. TNF-α was quantified using TNF-α DuoSet ELISA kit (Cat. no. DY210-05; R&D System) as per the manufacturer’s protocol. In brief, a 96 well-culture microplate (Corning Incorporated, Corning, NY) was coated with anti-human TNF-α capture Ab for overnight at room temperature, washed, and blocked with reagent diluent for 1 h at room temperature. Next, 100 μl of each sample and the reconstituted standard TNF-α (15.6–1000 pg/ml) were incubated overnight at room temperature followed by incubation with detection Ab (100 μl/well) for 2 h. Then, 100 μl substrate solution was added before the enzymatic reaction was stopped using 50 μl/well liquid stop solution. The microplate was read at 490 nm using iMark Microplate reader (Bio-Rad Laboratories) and analyzed using microplate manager 6 software.

M0, M1, or M2c macrophages were stimulated with recombinant human TNF-α (rhTNF-α; 5, 10, 20 ng/ml; Cat. no. 210-TA; R&D Systems) for 2 h followed by SM-LCL161 (1 μM) treatment for 48 h and analysis of cell death by PI staining and flow cytometry. To neutralize endogenous TNF-α, we incubated M1 macrophages with human FcR blocking buffer (Cat. no. 130-059-901; Miltenyi Biotec, Auburn, CA) for 20 min at 4°C. Macrophages were then treated with different concentrations of neutralizing LEAF Purified anti–TNF-α Abs (1.5, 5, and 15 μg/ml; Cat. no. 502804; BioLegend, San Diego, CA) or the control LEAF Purified Mouse anti-human IgG1, κ Isotype Ctrl (Cat. no. 400153; BioLegend) for 2 h. Cells were then treated with SM-LCL161 (0.5 μM) for 48 h and analyzed for cell death by PI staining and flow cytometry. To determine the biological activity of neutralizing anti–TNF-α Abs, we treated SNB-75 cells with complete medium containing different concentrations of rhTNF-α, anti–TNF-α Abs, or rhTNF-α and anti-TNF-α Abs together for 2 h followed by SM-LCL161 (3 μM) treatment for 48 h. SNB-75 cells were then analyzed for cell death by IncuCyte Zoom as described above.

Transfection of macrophages with small interfering RNA (siRNA) was performed as optimized earlier (24). M0 macrophages were cultured with 600 μl of serum-free, antibiotic-free medium (X1 DMEM supplemented with 10% FBS) for 3 h. IRF-1 siRNA or the negative control siRNAs (Cat. nos. 4392420 and 4390843, respectively; Thermo Fisher Scientific) were purchased. The transfection reagent Lipofectamine RNAiMax (Cat. no. 13778075; Thermo Fisher Scientific Invitrogen) was diluted in 200 μl of Opti-MEM medium. The diluted siRNA and diluted Lipofectamine RNAiMax reagent were then mixed in a 1:1 ratio and incubated for 5 min at room temperature to allow formation of the siRNA complex. The siRNA mixtures were then added to cells and incubated for 24 h. Cells were stimulated with IFN-γ for 48 h. Subsequently, cells were washed and treated with SM-LCL161 (1 μM) treatment. After 48 h, cells were collected, and cell death was assessed by intracellular PI staining and flow cytometry.

The statistical analyses were determined using GraphPad Prism 5.0 software (San Diego, CA). Data from all experiments were analyzed using the Mann–Whitney U test, and p values <0.05 were considered to be statistically significant. Data are presented as mean ± SD. The p values are represented as follows: *p < 0.05, ** p < 0.01, ***p < 0.001, and ****p < 0.0001.

Healthy donors involved in the study gave informed written consent, and the protocol for obtaining blood samples was approved by the Review Ethics Board of the Ottawa General Hospital and the Children’s Hospital of Eastern Ontario (Ottawa, ON, Canada).

We have previously demonstrated that polarization of M0 macrophages toward the M1 state rendered them highly susceptible to SM-LCL161–induced cell death, whereas M2a, M2b, and M2c differentiated subsets were resistant, with M2c being the most resistant (7). In this study, we confirmed our previous findings with emphasis on M1, M0, and M2c subsets. MDMs (M0) were polarized for 48 h into M1 and M2c subsets with IFN-γ and IL-10, respectively, and then treated with SM-LCL161 for 48 h. Analysis of cell death was assessed by intracellular PI staining and flow cytometry. As expected, M1 macrophages were highly sensitive to SM-LCL161 with a significant reduction in cell survival compared with untreated M1 or SM-LCL161–treated M0 and M2c macrophages (Fig. 1A). The levels of SM-induced cell death were dependent on the concentration of SM-LCL161 used and the duration of the presence of SM-LCL161 with M1 macrophages (Fig. 1B, 1C). Analysis of antiapoptotic cIAP-1 and cIAP-2 proteins expression revealed that untreated M0 and M2c macrophages expressed similar levels of cIAP-1 and cIAP-2, while untreated M1 macrophages exhibited a significantly higher expression of both cIAPs proteins (Fig. 1D, 1E). Despite these differences in protein expressions, SM-LCL161 treatment caused a significant degradation of cIAP-1/2 across all subsets (Fig. 1D, 1E). These results suggest that IAPs signaling is differentially regulated in M0, M1, and M2c macrophages, but that SM-induced cell death is independent of the overall expression levels of cIAP-1 and cIAP-2. Moreover, SM-induced cell death may require a second signal in M1 macrophages because degradation of cIAPs is not sufficient to induce cell death.

We have previously shown that SM-induced cell death in M1 macrophages is mediated by both extrinsic and intrinsic pathways of apoptosis (7). In this article, we confirm our results and show the involvement of both intrinsic and extrinsic apoptosis pathways by measuring the activation of caspase-3/7, -8, and -9 in SM-induced apoptosis in M1 macrophages. There was a significant increase in caspase-3/7 activities in M1 macrophages on treatment with 0.5 or 1.0 μM SM-LCL161 that was inhibited by prior treatment with z-VAD.fmk as measured by the IncuCyte live imaging (Fig. 2A). This increase in caspase-3/7 activity was detected as early as 4 h and reached a peak by 10 h after SM-LCL161 treatment (Fig. 2A). For caspase-9, the Western immunoblots showed an increase in cleaved caspase-9 bands on treatment with 0.2, 0.5, and 1 μM SM-LCL161 in M1 macrophages (Fig. 2B). The caspase-8 activity was assessed by measuring its cleavage activity by flow cytometry. M1 macrophages treated with either 0.2, 0.5, or 1 μM SM-LCL161 revealed significantly higher caspase-8 activity compared with the untreated cells as shown in the representative flow histogram (Fig. 2C). Staurosporine was used as a positive control to induce caspase-8 activity.

Because M1 macrophages are generated by IFN-γ, we determined whether IFN-γ–mediated signaling confers the sensitivity of M1 cells to SM-induced cell death and whether blocking IFN-γ signaling would reverse this effect. For this, M0 macrophages were treated with various concentrations of Jak inhibitor (ruxolitinib [RUX]) for 2 h before IFN-γ stimulation. Blocking Jak signaling by RUX significantly reduced SM-induced cell death in M1 macrophages in a dose-dependent manner (Fig. 3A). RUX alone did not affect cell survival in the absence of SM-LCL161. The biological activity of RUX was confirmed by showing inhibition of IFN-γ−induced Jak-1 phosphorylation (Fig 3B).

IFN-γ induces the expression of transcription factor IRF-1 that plays a key role in IFN-γ–induced cell death in various cancer cells (25, 26). To determine whether IRF-1 also plays a role in SM-induced cell death in M1 macrophages, we first evaluated the expression of IRF-1 in M0, M1, and M2c macrophages. M1 macrophages had increased IRF-1 expression compared with M0 and M2c subsets, and SM-LCL161 did not affect IRF-1 expression in M1 macrophages (Fig. 4A). To understand the role of IRF-1 in SM-mediated cell death, we transfected M0 macrophages with specific IRF-1 siRNA for 24 h and then stimulated them with IFN-γ for 48 h. The transfected macrophages were treated with SM-LCL161 for another 48 h before assessment of cell death. IFN-γ–stimulated M0 macrophages treated with SM-LCL161 in the presence and absence of nontargeting control siRNAs demonstrated comparable levels of cell death. The nontargeting or IRF-siRNAs alone did not affect cell death in these macrophages. Interestingly, IRF-1 silencing significantly reduced SM-induced cell death in IFN-γ–stimulated macrophages compared with the untransfected macrophages or macrophages treated with control nontargeting siRNAs (Fig. 4B). The IRF-1 siRNA downregulated IRF-1 expression compared with untransfected macrophages or macrophages transfected with nontargeting control siRNAs (Fig. 4C).

IFN-γ also induces the expression of IFN-induced, dsRNA-activated PKR, which is responsible for inhibiting viral protein synthesis, TNF-α, and Fas-induced apoptosis or RIPK-1/RIPK-3–mediated necroptosis (27, 28). Thus, IFN-γ–induced differentiation into M1 state may induce PKR activity that may mediate SM-induced cell death. The involvement of PKR was investigated using its specific C16 inhibitor that inhibits the apoptotic PKR/eukaryotic translation initiation factor 2α signaling pathway (29). M0 macrophages were treated with various concentrations of C16 for 2 h before stimulation with IFN-γ for 48 h followed by treatment with SM-LCL161 for another 48 h and determination of cell death by either PI staining and flow cytometry or IncuCyte PI staining (Supplemental Fig. 1A, 1B). C16 at any concentration did not rescue SM-induced cell death in IFN-γ–stimulated M0 macrophages. The biological activity of C16 was confirmed by showing reduction in IFN-γ–induced PKR phosphorylation (Supplemental Fig. 1C). Overall, SM-induced cell death during differentiation into M1 macrophages is regulated by the Jak/STAT pathway through the activation of IRF-1 transcription factor and independently of PKR activation.

Because cIAPs are key regulators of MAPK pathways (30), the role of ERK, JNK, and p38 MAPK was determined by employing their specific inhibitors PD-98059, SP-600125, and SB-203580, respectively (31). To determine the involvement of MAPKs during differentiation, we treated M0 macrophages with various concentrations of the above-mentioned MAPK inhibitors for 2 h followed by IFN-γ stimulation for 48 h. Alternatively, polarized M1 macrophages were treated with MAPK inhibitors for 2 h. Macrophages were then challenged with SM-LCL161 for another 48 h, followed by assessment of cell death. None of the MAPK inhibitors at any concentration affected SM-induced cell death during differentiation into M1 macrophages (Supplemental Fig. 2A–C, left) or into M1 polarized macrophages (Supplemental Fig. 2A–C, right). The biological activity of these inhibitors was also determined by showing inhibition of ERK, p38, and JNK phosphorylation compared with the control, IFN-γ/LPS–stimulated macrophages (Supplemental Fig. 2D).

The PI3K/Akt-mTORC pathway plays a key role in cell activation, proliferation, and survival (32, 33). To determine the involvement of the PI3K/Akt pathway, we used PI3K and Akt inhibitors, LY294002 and MK0266, respectively. M0 macrophages were treated with either PI3K or Akt inhibitor for 2 h followed by IFN-γ stimulation. Similarly, the polarized M1 macrophages were treated with either PI3K or Akt inhibitors for 2 h. Subsequently, the macrophages were challenged with SM-LCL161 for 48 h followed by assessment of cell death. Neither PI3K inhibitor (Supplemental Fig. 3A) nor Akt inhibitor (Supplemental Fig. 3B) at any concentration added before polarization or in M1 polarized macrophages affected SM-induced cell death. The activity of PI3K inhibitors was confirmed by showing inhibition of LPS-induced Akt phosphorylation (Supplemental Fig. 3C, left), whereas that of Akt inhibitor was confirmed by showing inhibition of IFN-γ–induced Akt phosphorylation in M0 macrophages (Supplemental Fig. 3C, right).

The mTORC pathway plays a central regulatory role in many cellular processes, such as cell metabolism, cell survival, differentiation, proliferation, and immune responses (34). The mTORC protein complex is a serine/threonine kinase that acts downstream of phosphoinositide-dependent kinase-1 in the PI3K pathway and functions as a nutrient/energy sensor and controls protein synthesis and cell growth (34). The mTORC links with other proteins and serves as a core component of two distinct protein complexes, mTORC-1 and mTORC-2, which regulate different cellular processes (34). The role of mTORC complexes was investigated by using the chemical inhibitors rapamycin, which inhibits mTORC-1, and torin, which inhibits both complexes (35, 36). Because ribosomal p70S6 kinase is a downstream target of the mTORC-1 complex (34), we also investigated the role of p70S6K in the SM-induced cell death. The p70S6K inhibitor (PF-4708671) targets S6K downstream of mTORC-1 responsible for protein synthesis (37). To determine the involvement of the mTORC pathway, we treated M0 macrophages with various concentrations of mTORC inhibitors torin, rapamycin, or PF-4708671 for 2 h followed by IFN-γ stimulation for 48 h. The macrophages were treated with SM-LCL161 for another 48 h followed by assessment of cell death. Blocking of the mTORC pathway by torin (Fig. 5A, left) and rapamycin (Fig. 5B, left) and of the S6K pathway by PF-4708671 (Fig. 5C, left) at any concentration added before polarization significantly reduced SM-induced cell death. The biological activity of torin, rapamycin, and PF-4708671 was determined by treating macrophages with different concentrations of these inhibitors followed by stimulation with IFN-γ or LPS for 30 min followed by Western blotting for S6K phosphorylation. Torin, rapamycin, and PF-4708671 inhibited S6K phosphorylation in a dose-dependent manner as compared with the positive control, IFN-γ–, or LPS-stimulated macrophages in the absence of the inhibitors (Fig. 5A–C, right). Because rapamycin inhibits mTORC-1 and torin inhibits both mTORC-1 and mTORC-2, the results suggest that the SM-induced cell death during differentiation into M1 macrophages is mediated by the activation of the mTORC-1 pathway and its downstream S6K signaling pathway.

To determine the involvement of the mTORC1/2-S6K pathway in SM-induced cell death of polarized M1 macrophages, we polarized M0 macrophages into M1 state by IFN-γ followed by treatment with various concentrations of mTORC inhibitors, torin and rapamycin, or S6K inhibitor for 2 h. The macrophages were then treated with SM-LCL161 for another 48 h before assessment of cell death. Interestingly, only torin protected polarized M1 macrophages against SM-induced cell death (Fig. 6A). Because rapamycin failed to rescue macrophage from SM-induced cell death (Fig. 6B), the results suggest that SM-induced cell death in M1 polarized macrophages is regulated by the mTORC-2 pathway and independent of the mTORC-1/S6K pathway (Fig. 6C).

It has been shown that IRF-1 is involved in the differentiation of M1 macrophages (38, 39). Impairment of IFN-γ signaling through the specific inhibitors for Jak/STAT and IRF pathways may thus inhibit M1 differentiation, leading to inhibition of SM-induced cell death in M1 macrophages. Because IFN-γ activates the mTORC pathway, and its impairment by mTORC inhibitors, rapamycin and torin, inhibited SM-induced cell death (Figs. 5, 6), it was pertinent to determine whether mTORC regulates SM-induced cell death through mTORC-mediated M1 differentiation. For this, we analyzed M1-specific surface markers CD80, CD86, and HLA-DR after treatment of M0 macrophages with rapamycin or torin and IFN-γ stimulation for M1 differentiation. IFN-γ stimulation induced M1 differentiation as evidenced by enhanced expression of CD80, CD86, and HLA-DR (Fig. 7). However, pretreatment with either rapamycin or torin failed to inhibit M1 differentiation because these inhibitors did not affect IFN-γ–induced expression of CD80, CD86, and HLA-DR (Fig. 7). These results suggest that IFN-γ–activated mTORC mediates SM-induced cell death in M1 macrophages independent of IFN-γ–induced M1 differentiation.

SM-induced cell death in M1 macrophages and the contrasting resistance of M0/M2c macrophages to SM-LCL161 may be because of differential expression of proapoptotic and antiapoptotic proteins. High levels of cIAP-1/2 proteins in M1 macrophages and their similar degradation by SM-LCL161 do not explain the resistance of M0/M2c subsets and susceptibility of M1 macrophages to SM-induced cell death (Fig. 1D, 1E). Thus, SM-induced degradation of cIAP-1/2 proteins alone may not be enough to cause cell death in M1 macrophages and may require the involvement of other proapoptotic, antiapoptotic, or IAP-associated proteins. Therefore, we analyzed the expression of proapoptotic Bax and Fas; antiapoptotic Bcl-2 family of proteins, namely, Bcl-2, Mcl-1, and Bcl-xL; IAP proteins, namely, cIAP-1 and cIAP-2; and IAP-associated proteins, such as RIPK-1 and RIPK-3, in M0, M1, and M2c macrophages in the presence and absence of various concentrations of SM-LCL161. As expected, SM-LCL161 caused degradation of both cIAPs proteins in M0, M1, and M2c macrophages (Fig. 8A). However, SM-LCL161 did not affect the expression of proapoptotic Fas, TRAIL proteins, or antiapoptotic Bcl-2, Bcl-xL, or Mcl-1 in M0, M1, or M2c macrophages, although M1 macrophages exhibit significantly higher expression of these proteins (Fig. 8A). In addition, SM-LCL161 caused accumulation of proapoptotic Bax protein in M1 macrophages, but not in M0 and M2c macrophages (Fig. 8B).

Interestingly, analysis of the IAP-associated proteins RIPK-1 and RIPK-3 revealed that SM-LCL161 treatment caused degradation of both RIPK-1 and RIPK-3 proteins in M1 macrophages, whereas the expression of these proteins was not altered in M0 and M2c macrophages (Fig. 8A). These results suggest that SM-LCL161-induced cell death in M1 macrophages may be associated with the simultaneous degradation of cIAP-1, cIAP-2, RIPK-1, and RIPK-3. We have previously shown that SM-induced cell death in M1 macrophages involves both apoptosis and necroptosis pathways by showing inhibition of SM-induced cell death after treatment of M1 macrophages with zVAD.fmk (caspase inhibitor) and necrosulfonamide (necroptosis inhibitor) (7). Furthermore, SM treatment caused degradation of RIPK-1, RIPK-3, and MLKL proteins that was restored after pretreatment with zVAD.fmk (caspase inhibitor) and Necrostatin-1 and/or necrosulfonamide (necroptosis inhibitors) (7). In this article, we confirm these results by showing that SM-LCL161 degraded RIPK-1 and RIPK-3 in M1, but not in M0 and M2c macrophages.

Because mTORC inhibitor torin rescued SM-induced cell death, we predicted that SM-induced cIAP-1/2 and RIPK-1/3 degradation would be restored after treatment of M1 macrophages with torin. Therefore, the expression of cIAP-1, cIAP-2, and other effector molecules, including RIPK-1, RIPK-3, and Bax, after blocking mTORC by torin was investigated. SM-LCL161 treatment for 24 h caused degradation of cIAP-1, cIAP-2, RIPK-1, and RIPK-3. Blocking mTORC either during differentiation into M1 macrophages (Fig. 9A, 9B) or in polarized macrophages by torin significantly prevented the degradation of cIAP-2, RIPK-1, and RIPK-3 (Fig. 10A, 10B). Interestingly, cIAP-1 was not affected by blocking the mTORC pathway. As described above, SM-LCL161 treatment significantly enhanced the expression of proapoptotic Bax protein in M1 macrophages. Treatment of macrophages with torin either during differentiation into M1 macrophages (Fig. 9C) or in polarized M1 macrophages (Fig. 10C) significantly inhibited Bax expression in SM-treated macrophages to levels similar to untreated macrophages. Overall, the results suggest that SM-induced cell death is regulated by cIAP-2, RIPK-1, and RIPK-3 through the activation of the mTORC pathway during differentiation into M1 macrophages and in polarized M1 macrophages.

Endogenously produced TNF-α has been shown to mediate SM-induced cell death in various cell lines (40, 41). In several cancer cell lines, SMs induce the production of TNF-α through the activation of noncanonical NF-κB, where then it synergizes with SMs to induce cell death (40, 42). Therefore, we first determined the levels of TNF-α produced by M1 macrophages either spontaneously or after treatment with SM-LCL161. For this, M0 and polarized M1 and M2c macrophages were cultured for 48 h followed by TNF-α measurement in the supernatants. All three subsets produced TNF-α spontaneously, with M1 macrophages producing significantly higher levels compared with the M0 and M2c subsets (Fig. 11A). Alternatively, macrophage subsets were washed and then treated with SM-LCL161 for 48 h followed by TNF-α measurement in the supernatants. Contrary to cancer cell lines, SM-LCL161 at any concentration did not induce TNF-α production in either M0, M1, or M2c macrophages compared with the untreated macrophages (Fig. 11B–D).

Next, we investigated whether enhancing TNF-α production by LPS would augment SM-LCL161–induced cell death. As seen in (Fig. 11E, LPS induced high levels of TNF-α production (>1000 pg/ml) in all three subsets after 24 h of stimulation, with M1 macrophages producing the highest levels (∼3000 pg/ml). LPS-stimulated M0 and M2c macrophages were still resistant to cell death induced by any concentration of SM-LCL161 (Fig. 11F, 11G). Interestingly, stimulation with LPS did not lead to a significant increase in cell death in M1 macrophages after treatment with any concentration of SM-LCL161 (Fig. 11H).

In addition to inducing TNF-α expression, LPS stimulation can induce a number of other changes in macrophages that may be responsible for resistance to cell death in M0 and M2c macrophage subsets. To directly determine the involvement of TNF-α in macrophage cell death, we treated M0, M2c, and M1 macrophages ex vivo with rhTNF-α for 2 h before SM-LCL161 treatment. TNF-α–treated M0 and M2c macrophages were still resistant to SM-induced cell death (Fig. 11I, 11J), and addition of rhTNF-α (5–20 ng/ml) failed to significantly increase cell death in M1 macrophages compared with that observed after SM-LCL161 treatment alone (Fig. 11K).

To further rule out the involvement of TNF-α, we neutralized endogenous TNF-α with anti-TNF-α neutralizing Abs. M1 macrophages treated with anti–TNF-α neutralizing Abs for 2 h before treatment with SM-LCL161 were still sensitive to SM-induced cell death (Fig. 11L, Supplemental Fig. 4A). Addition of anti–TNF-α Abs at any concentration (1.5–15 μg/ml) did not significantly reduce SM-induced cell death compared with that observed with isotype-matched control Abs (Fig. 11L). Representative histograms showing the effect of anti–TNF-α Abs on SM-induced cell death are shown (Supplemental Fig. 4A). The biological activity of neutralizing anti–TNF-α Abs was determined by using a standard human SNB-75 glioblastoma cell line. SM-LCL161 induced cell death in SNB-75 cells only when it was combined with rhTNF-α. The addition of anti–TNF-α Abs (15 μg/ml) significantly reduced SM-induced cell death in SNB-75 cells compared with the cell death observed with isotype-matched control Abs (15 μg/ml; Supplemental Fig. 4B). Overall, SM-LCL161 did not induce TNF-α production, and SM-induced cell death in primary human M1 macrophages appears to occur independent of TNF-α.

Because SMs degrade cIAP-1/2 and IAPs regulate NF-κB signaling (42, 43), we wanted to evaluate whether SM-LCL161 induces cell death through NF-κB signaling in M1 macrophages. cIAPs ablation by SMs promotes a noncanonical pathway of NF-κB activation in cancer cells leading to enhanced TNF-α secretion, which after interaction with its receptor leads to degradation of IκBα and subsequent release of p65 and p50 subunits of NF-κB (41, 42). To determine the effect of SM-LCL161 on the canonical pathway of NF-κB activation, we measured the level of IκBα by Western blotting. For this, LPS stimulation was used as a positive control. In contrast with cancer cells (30, 41), SM-LCL161 alone did not affect IκBα expression in M1 macrophages. As expected, LPS stimulation decreased IκBα levels, indicative of NF-κB activation (Fig. 12A).

Noncanonical regulation of NF-κB depends on NF-κB–inducing kinase (NIK), which promotes partial processing of the p100 subunit to its p52 mature form (44). In resting cells, this pathway is subdued via continuous NIK degradation by cIAP-1/2 (42, 45). SMs treatment results in NIK accumulation and subsequent activation of the noncanonical NF-κB pathway in cancer cells (41, 42, 46). To determine the effect of SM on noncanonical NF-κB signaling, we treated M1 macrophages with 0.5 and 1 μM SM-LCL161 as described above. LPS stimulation was again used as a positive control. Levels of p100 to p52 were evaluated by Western blotting. In contrast with cancer cells (41, 42), SM-LCL161 alone did not affect p100 processing in macrophages. LPS stimulation induced the expression of p100 and its subsequent processing to p52 (Fig. 12B). The results suggest that SM-LCL161 alone does not activate either the canonical or noncanonical pathway of NF-κB in M1 macrophages.

To determine whether blocking the NF-κB pathway during differentiation into M1 macrophages or in polarized M1 macrophages would rescue SM-induced cell death, we treated M0 macrophages with various concentrations of the NF-κB inhibitor PS-1145, an IKK inhibitor, for 2 h followed by IFN-γ stimulation for 48 h. Alternatively, M1 macrophages first polarized by IFN-γ were treated with PS-1145 for 2 h. Subsequently, cells were treated with SM-LCL161 for another 48 h followed by assessment of cell death. Blocking of NF-κB by PS-1145 either before or after polarization into M1 macrophages did not rescue SM-LCL161–induced cell death in M1 macrophages (Fig. 12C). The biological activity of PS-1145 was confirmed by inhibition of IFN-γ–induced phospho-IKK expression (Fig. 12D). Overall, the results suggest that selective induction of cell death in human M1 macrophages by SMs may be mediated by cIAP-2, RIPK-1, and RIPK-3 degradation through mTORC activation and independent of TNF-α and NF-κB pathways (Fig. 13).

In this study, we investigated the mechanism governing SM-induced cell death during differentiation into M1 macrophages and in polarized M1 macrophages. IFN-γ–induced differentiation conferred the sensitivity of M1 cells to SM-induced cell death through the Jak/STAT and IRF-1 pathways and the mTORC-1-S6K pathways. However, SM-induced cell death in polarized M1 macrophages is regulated by the mTORC-2 pathway. Also, mTORC-1 regulated SM-induced cell death independent of M1 differentiation. In addition, SM-induced cell death is mediated by cIAP-2, RIPK-1, and RIPK-3 degradation through the activation of mTORC in polarized M1 macrophages, as well as during M1 differentiation. Finally, in contrast with cancer cell lines, SM-induced cell death is independent of endogenously produced TNF-α, as well as canonical and noncanonical NF-κB pathways in M1 macrophages.

IFN-γ plays a key role in immunity against intracellular pathogens, inflammation, and cancer (47, 48). It is primarily produced by activated Th cells and NK cells (49) and mediates polarization of M0 into M1 macrophages (1, 4). Thus, IFN-γ produced by immune cells may play a central role in driving M1 polarization in vivo. The mechanism by which IFN-γ induces susceptibility to SM-induced cell death in M1 macrophages remains unknown. Here, we show that SM-induced cell death is dependent on IFN-γ–induced signaling pathways because inhibiting the Jak/STAT pathway by specific inhibitors and IRF-1 silencing restored survival of M1 macrophages. Although PKR, an IFN-induced kinase, plays a key role in promoting IFN-γ–induced cell death (27, 28), SM-induced cell death in M1 macrophages was found to be PKR independent.

IFN-γ also activates PI3K/Akt-mTORC pathways (50). PI3K or Akt pathways were not involved in SM-induced cell death either during differentiation or in polarized macrophages. However, blocking the mTORC-S6K pathway revealed interesting results. Our results suggest that mTORC-1 is sufficient to mediate SM-induced cell death during differentiation into M1 macrophages. In contrast, SM-induced cell death in polarized M1 macrophages is regulated by a distinct mTORC-2 pathway independent of the mTORC-1-S6K pathway involved in SM-induced cell death during IFN-γ–induced differentiation of M0 into M1 macrophages. The mTORC pathway plays a central regulatory role in many cellular processes, such as cell metabolism, survival, differentiation, proliferation, and immune responses (34). Unlike mTORC-1, which is extensively studied in cancer studies, the role of mTORC-2 is not completely defined (36). Activated mTORC-1 targets several downstream proteins, such as ribosomal protein S6K, eukaryotic translation initiation factor 4E binding proteins, and sterol regulatory element–binding protein 1/2 involved in protein and lipid synthesis (34, 51). In contrast, mTORC-2 controls other proteins, such as serum and glucocorticoid-induced protein kinase-1, and protein kinase C-α that contributes to cell survival, metabolism, and cytoskeleton organization (34, 52–54). Both mTORC complexes share overlapped functions and regulate several proteins involved in metabolism, cell survival, immune cell activation, and differentiation (34, 37). Our results suggest a differential requirement of mTORC1-S6K in SM-induced cell death occurring during differentiation and of mTORC-2 in polarized M1 macrophages based on the differential activity of mTORC-1 and mTORC-2 inhibitors. However, the precise role of mTORC-1 and mTORC-2 needs to be discerned by the selective deletion of the adaptor protein Raptor (mTORC-1) and Rictor (mTORC-2) (54–56). Alternatively, the role of mTORC-1 could be determined by activating mTORC-1 via silencing the endogenous mTORC-1 inhibitor tuberous sclerosis complex 2 (56).

One of the ubiquitous roles of mTORC is to promote anabolic processes such as protein synthesis accompanied by increased protein degradation and elevated levels of active proteasomes through the increased expression of genes encoding proteasome subunits (34, 57–59). SMs cause degradation of cIAP-1/2 through their E3 ubiquitin ligase activity. Depletion of cIAPs results in the formation of the ripoptosome, autoubiquitination, and subsequent proteasomal degradation, leading to caspase activation and apoptosis (42, 60). Our results show SMs caused degradation of cIAP-2, RIPK-1, and RIPK-3 that was prevented by mTORC-1/2 inhibitors, rapamycin and torin, suggesting that mTORc regulates SM-induced cell death in M1 macrophages through the proteasomal pathway. However, further studies are needed to understand the involvement of the proteasomal pathway, including the ubiquitination of these proteins and their stabilization employing proteasomal inhibitors such as MG132 in SM-induced cell death in M1 macrophages.

IRFs play an important role in macrophage differentiation, regulation of maturation, phenotypic polarization, and their function (38, 39). Among the nine IRFs, IRF-1 is involved in commitment of M1 macrophages, whereas IRF-4 controls polarization into M2 macrophages (39, 61, 62). Because IFN-γ activates IRF through the Jak/STAT pathway (63, 64), IFN-γ–induced differentiation into M1 macrophages may thus be regulated by the Jak/STAT–IRF-1 pathway. In this article, we demonstrated that the IFN-γ–activated Jak/STAT and IRF-1 pathway mediates SM-induced cell death in M1 macrophages. As shown by us and others (65), IFN-γ activates the mTORC pathway. Because mTORC also mediates SM-induced cell death in M1 macrophages, we determined whether mTORC-1/2 mediates M1 differentiation that may be required for SM-induced cell death. IFN-γ–activated mTORC did not regulate M1 differentiation, which suggested that IFN-γ–activated mTORC-1/2 is sufficient to cause SM-induced cell death independent of M1 differentiation.

Our results show that SM-induced cell death in M1 macrophages is not attributed to the expression levels of IAPs alone and requires the involvement of IAP-associated signaling molecules or other members of the proapoptotic or antiapoptotic Bcl-2 family. We have shown that Bcl-2 proteins do not regulate SM-induced cell death because the expression levels of Bcl-2, Mcl-1, Bcl-xL, and other proteins, such as Fas and TRAIL, were not affected after SM-LCL161 treatment. As expected, the levels of proapoptotic Bax were increased in M1 macrophages after SM-LCL161 treatment, in contrast with the M0 and M2c macrophages wherein the levels of Bax were not affected.

The RIPK-1 and RIPK-3 enzymes play a key role in several cellular processes, including apoptosis and necroptosis (66). In the context of TNF-α, its interaction with the TNF receptor leads to the recruitment of the adaptor, TRADD, which recruits RIPK-1 and two ubiquitin ligases, TRAF2 and clAP-1, which eventually form the TNF receptor-1 complex-I leading to NF-κB activation (67, 68). In the absence of IAPs and NF-κB, complex-I is modified to form a complex-IIa that eventually culminates in apoptosis (67, 69). In the absence of IAPs and the caspases, complex-IIb is assembled in the cytosol (67, 69), which allows RIPK-1 and RIPK-3 to phosphorylate each other, leading to the formation of necrosome (67, 69). The necrosome recruits MLKL, which is phosphorylated by RIPK-3, and this complex translocates to the plasma membrane, leading to pore formation and eventually necroptosis (70, 71). We have previously shown that SM-LCL161 induced apoptosis, as well as necroptosis, in M1 macrophages (7). Interestingly, evaluation of IAPs and IAP-associated kinases revealed that RIPK-1 and RIPK-3 were degraded in M1, but not in M0 and M2c, macrophages on treatment with SM-LCL161. The fact that the mTORC inhibitor, torin, restored the expression of cIAP-2, RIPK-1, and RIPK-3 after SM-LCL161 treatment suggests that SM-induced cell death in M1 macrophages is regulated by cIAP-2, RIPK-1, and RIPK-3 through the activation of mTORC1/2. Moreover, RIPK-1 and RIPK-3 may cooperate with cIAP-2 in the differential regulation of SM-induced cell death in M1 macrophages (Fig. 13).

As mentioned earlier, cIAP-1 and cIAP-2 also play an essential role in NF-κB activation (30, 42). In the context of TNF-α, interaction of TNF-α with its receptor leads to the activation of both canonical and noncanonical NF-κB pathways through the recruitment of various proteins, including cIAP-1/2, TRAFs, and RIPK-1 (46, 72). In cancer cell lines, SMs alone activate both canonical and noncanonical NF-κB pathways by promoting TNF-α secretion and stabilizing NIK, respectively (41, 42, 46). In contrast with cancer cells, we show that SMs alone did not activate either the canonical or noncanonical NF-κB pathway in M1 macrophages. Moreover, these results are in conformity with our observations using the NF-κB inhibitor clearly indicating that NF-κB did not mediate SM-induced cell death in M1 macrophages.

SM-mediated killing has been in general attributed to endogenously produced TNF-α in cancer cells (40). Our results show that M1 macrophages either alone or after LPS stimulation produced high levels of TNF-α, SM-LCL161 did not induce TNF-α production, rhTNF-α did not enhance SM-induced cell death, and neutralization of endogenous TNF-α by anti–TNF-α Abs did not restore cell survival. These results suggest that the SM-induced cell death in M1 macrophages is regulated by a mechanism independent of endogenously produced TNF-α. In accordance with our findings, there are reports showing that SM induces cell death in a TNF-α–independent manner in some cell types (73–75). The different results on SM-induced TNF-α production, cell death, and canonical and noncanonical NF-κB activation in cancer cell lines and primary macrophages may be because of different cell types. SM-sensitive cell lines secrete TNF-α in response to cIAPs ablation and consequently become susceptible to TNF-α–induced apoptosis (40). However, M1 macrophages produce significant amounts of TNF-α without stimulation, and we did not detect enhanced TNF-α secretion after SMs treatment. This is in keeping with the fact that SMs treatment causes apoptosis of sensitive cancer cells (40, 76), but not of primary macrophages (21, 77).

In summary, our results show that susceptibility of M1 macrophages to SM-induced cell death is dependent on IFN-γ–mediated differentiation through the Jak/STAT pathway and subsequent activation of IRF-1. In addition, SM-induced cell death occurs after simultaneous degradation of cIAP-2, RIPK-1, and RIPK-3 through the activation of the mTORC signaling pathway. The inflammatory M1 macrophages have been implicated in many diseases, including rheumatoid arthritis, inflammatory bowel disease, and multiple sclerosis (3, 8, 13, 78). Thus, selective induction of cell death in M1 macrophages by agents blocking the IAPs and/or RIPK-1/3 would have potential therapeutic benefits with regard to macrophage-associated inflammatory diseases. However, killing of M1 macrophages by SM in diseases such as cancer may have an adverse impact on their ability to mount an effective innate response against various infectious agents and worsen immune deficiency.

We acknowledge the generous help provided by the nurses in collecting blood samples. We also acknowledge the healthy donors for providing blood samples.

The authors have no financial conflicts of interest.