Activation of TNFR2 sensitizes macrophages for TNFR1-mediated necroptosis

Macrophages express TNFR1 as well as TNFR2 and are also major producers of tumor necrosis factor (TNF), especially upon contact with pathogen-associated molecular patterns. Consequently, TNF not only acts as a macrophage-derived effector molecule but also regulates the activity and viability of macrophages. Here, we investigated the individual contribution of TNFR1 and TNFR2 to TNF-induced cell death in macrophages. Exclusive stimulation of TNFR1 showed no cytotoxic effect whereas selective stimulation of TNFR2 displayed mild cytotoxicity. Intriguingly, the latter was strongly enhanced by the caspase inhibitor zVAD-fmk. The strong cytotoxic activity of TNFR2 in the presence of zVAD-fmk was reversed by necrostatin-1, indicating necroptotic cell death. TNFR1- and TNF-deficient macrophages turned out to be resistant against TNFR2-induced cell death. In addition, the cIAP-depleting SMAC mimetic BV6 also enforced TNF/TNFR1-mediated necroptotic cell death in the presence of zVAD-fmk. In sum, our data suggest a model in which TNFR2 sensitizes macrophages for endogenous TNF-induced TNFR1-mediated necroptosis by the known ability of TNFR2 to interfere with the survival activity of TRAF2-cIAP1/2 complexes.

Tumor necrosis factor (TNF) is a pleiotropic cytokine that occurs as a type II transmembrane protein but can be released from the plasma membrane by proteolytic processing. 1 Membrane-bound and soluble TNF both contain the characteristic carboxy-terminal TNF homology domain, which is responsible for self-assembly into trimeric molecules and receptor binding. Membrane-bound and soluble TNF strongly interact with two receptors, TNFR1 and TNFR2, but the two forms of TNF are differentially effective in receptor activation. 1 Whereas membrane-bound TNF activates TNFR1 and TNFR2 efficiently, soluble TNF is sufficient for TNFR1 activation but largely inactive upon binding to TNFR2. TNFR1 belongs to the death receptor subgroup of the TNF receptor family and can trigger apoptosis and necroptosis. [2][3][4] However, cell death induction by TNFR1 is typically efficiently antagonized by concomitant activation of the cytoprotective classical NFκB pathway and/or ubiquitous expression of anti-apoptotic proteins. 1,2 The latter involve FLIP proteins which generally inhibit death receptor-induced caspase-8 activation but also complexes containing TRAF2, cIAP1 and cIAP2 which specifically interfere with caspase-8 activation in context of TNFR1 signaling. [2][3][4] Worth mentioning, TRAF2-cIAP1/2 complexes also mediate K63-linked ubiquitination of RIP1 in the TNFR1 signaling complex, thereby facilitating TNFR1mediated activation of the classical NFκB pathway. Indeed, TNFR1 signaling is predominantly pro-inflammatory as TNFR1-induced cell death is blocked as long as the aforementioned protective mechanisms are not impaired.
In contrast to TNFR1, TNFR2 contains no cytoplasmic death domain. Upon ligand binding, TNFR2 recruits TRAF2 and various TRAF2-associated proteins, such as TRAF1, cIAP1 and cIAP2, but also interacts with other signaling proteins independently of TRAF2. 1,5 TNFR2 activation has been linked to a variety of immune regulatory functions, which, in contrast to the activities of TNFR1, often result in antiinflammatory effects. 6 Murine models shed light on the complex interplay of the TNFR1-TNFR2 system in vivo, demonstrating additive, synergistic or even antagonistic effects. At the cellular level, several mechanisms for the cross-talk between TNFR1 and TNFR2 have been identified. 1 Besides the obvious competition for ligand binding, TNFR1 and TNFR2 can induce, for example, autocrine TNF production in a cell type-specific manner. 1 In context of TNFR1 activation by soluble TNF, subsequent induction of membrane-bound TNF results in costimulation of TNFR2, thereby converting the initially transient activation into sustained autocrine signaling. In addition, TNFR1 and TNFR2 compete for the cytoplasmic pool of TRAF2-cIAP1/2 complexes. By depletion and/or degradation of TRAF2, TNFR2 is capable to modulate TNFR1 signaling. 1 Moreover, TNFR2 but not TNFR1, stimulates the alternative NFκB pathway by triggering proteolytic processing of the inactive p100/RelB dimers into active p52/RelB NFκB complexes. 7 Notably, TNFR2-induced alternative NFκB signaling can be enhanced by TNFR1-mediated induction of p100 and RelB expression via the classical NFκB pathway. 7 In macrophages, the complexity of the TNF-TNFR1/2 system is especially relevant. Macrophages on one hand coexpress TNFR1 and TNFR2 and are on the other hand a pathophysiologically important source of TNF, for example, in response to a variety of pathogen-associated molecular patterns (PAMPs). TNF not only acts as a macrophagederived effector molecule, but in an autocrine fashion also controls macrophage activation and survival, as seen for example during infection with mycobacteria. [8][9][10][11][12][13][14][15][16][17] However, the molecular mechanisms of TNF-induced cell death in macrophages are incompletely understood and were, therefore addressed in our study. Using macrophages genetically deficient for TNFR1, TNFR2 or TNF together with TNFR1and TNFR2-specific TNF variants, we show that TNFR2 activation sensitizes macrophages for TNFR1-mediated necroptosis triggered by autocrine produced TNF and provide evidence that this is related to TNFR2-induced depletion/ degradation of TRAF2-cIAP1/2 complexes.

Results
TNFR2 stimulation triggers necroptosis in zVAD-fmk treated macrophages. For our study, we generated macrophages from HoxB8-immortalized murine myeloid progenitor cells (MPCs). Typical murine macrophage surface markers such as F4/80 and CD11b were detectable 5 days after differentiation into macrophages was initiated (Figure 1a). Flow cytometry revealed significant TNFR2 expression on macrophages, whereas there was only poor expression of this receptor on MPCs (Figure 1a). TNFR1 was not detectable by flow cytometry. However, functional studies with human TNF, which does not bind to murine TNFR2 18 and consequently selectively stimulates TNFR1 in murine cells, revealed robust TNFR1 responsiveness as seen by activation of the classical NFκB pathway and upregulation of the NFκB target genes TRAF1 and p100 (Figure 1b). To determine the individual contribution of TNFR1 and TNFR2 to TNF-induced cell death of macrophages, we selectively stimulated the two receptors using soluble human TNF and a nonameric TNF fusion protein with specificity for TNFR2 (TNC-sc(mu)TNF80). Importantly, cells were also challenged with the TNFR1-and TNFR2-specific ligands in the presence of the pan-caspase inhibitor zVAD-fmk, which can unleash potential necroptotic activity by blocking caspase-8-mediated cleavage of RIP1. 19,20 Irrespective of zVAD-fmk, human TNF-mediated TNFR1 activation had no cytotoxic effects, neither in HoxB8immortalized MPCs, nor in macrophages (Figure 1c). In macrophages, TNFR2 activation using TNC-sc(mu)TNF80 resulted in a mild cell death response, which, however, was strongly enhanced in the presence of zVAD-fmk ( Figure 1c). In accordance with the fact that TNFR2 is not or only poorly expressed on murine MPCs (Figure 1a), TNC-sc(mu)TNF80 had no cytotoxic effect on this cell type (Figure 1c). Macrophages were rescued from the cytotoxic activity of the zVADfmk/TNC-sc(mu)TNF80 mixture by treatment with the RIP1 inhibitor necrostatin-1 (nec-1, ref. 21 Figures 2a and b). Since RIP1 has a crucial role in necroptotic signaling cascades this indicates that cell death induction by zVAD-fmk/TNC-sc(mu) TNF80 is due to necroptosis. Expectedly, TNC-sc(mu)TNF80 showed no cytotoxic effect on M-CSF differentiated SCF-ER-Hoxb8 immortalized macrophages derived of TNFR2-deficient mice (Figures 2a and c).
TNFR2 stimulation triggers autocrine TNF-production and TNFR1-mediated necroptosis in murine macrophages. Notably, TNFR2 neither interacts with RIP1 nor with caspase-8, but has the ability to modulate the activity of these molecules in context of TNFR1 signaling by restricting the availability of TRAF2-cIAP1 and TRAF2-cIAP2 complexes. 1,[22][23][24] We previously demonstrated in tumor cell lines that TNFR2-mediated depletion/degradation of cytoplasmic TRAF2-cIAP1/2 complexes not only antagonizes TNFR1-induced RIP1-mediated activation of the classical NFκB pathway, but also enhances TNFR1-induced cell death. 23,24 Given that macrophages express TNF and that TNFR1 is a prominent and potent trigger of necroptosis, 2-4 we evaluated whether TNF and TNFR1 are involved in induction of necroptosis by zVAD-fmk/TNC-sc(mu)TNF80. For this purpose, macrophages derived from Hoxb8immortalized murine MPCs of TNF-and TNFR1-deficient mice ( Figure 3a) were subjected to TNFR1/TNFR2 stimulation experiments. In both cell types, co-treatment with zVADfmk and TNC-sc(mu)TNF80 showed no significant cytotoxic effect (Figures 3b and c) indicating that TNFR2-induced necroptosis in macrophages is due to sensitization for endogenous TNF-induced TNFR1-mediated necroptosis. Moreover, exclusive stimulation of TNFR2 resulted in induction of TNF, which was even enhanced in the presence of zVAD-fmk (thus under necroptotic conditions) ( Figure 3d). Together, these data raised the possibility that TNFR2/zVADfmk triggers TNFR1-mediated necroptosis in macrophages by inducing TNF production and concomitant interference with the inhibitory activity of TRAF2-cIAP1/2 complexes on RIP1-dependent necroptosis. Indeed, we found a strong reduction of TRAF2 levels after TNFR2 stimulation in the presence and absence of zVAD-fmk and/or nec-1 ( Figure 3e). Notably, TRAF2 levels remained low in necrostatin-1protected cells following zVAD-fmk/TNC-sc(mu)TNF80 challenge, indicating that the TNFR2-triggered loss of TRAF2 was not an epiphenomenon of ongoing necroptosis ( Figure 3e). TNFR2 stimulation did not affect cIAP1 levels ( Figure 3e). cIAP2 expression in macrophages is too low to allow detection by standard western blotting protocols.
BV6 triggers TNF/TNFR1-mediated necroptosis in murine macrophages. To further substantiate the idea that TNFR2 enhances TNFR1-induced necroptosis in macrophages by limiting the availability of TRAF2-cIAP complexes for TNFR1 stimulated by co-expressed TNF, we investigated the effects of the SMAC mimetic BV6 on macrophage viability. BV6 is a symmetric low-molecular weight compound that interacts with two molecules of XIAP, cIAP1 or cIAP2. 25 In the case of cIAP1 and cIAP2, this results in dimerization-induced K48-linked auto-ubiquitination and subsequent proteasomal degradation. 25 cIAP1 expression levels were severely reduced in BV6-treated macrophages irrespective of zVADfmk or necrostatin-1 treatment, while there was no effect on TRAF2 (Figure 4a). In line with an anti-necroptotic activity of TRAF2-cIAP complexes (and similar to TNFR2 stimulation),  BV6 triggered cell death in macrophages in the presence of zVAD-fmk in a dose-dependent manner (Figure 4b). zVADfmk/BV6-induced cytotoxicity was largely abolished by necrostatin-1 treatment (Figure 4c), again indicating necroptotic cell death. Moreover, zVAD-fmk/BV6-induced necroptosis was abrogated in macrophages derived from Hoxb8immortalized murine MPCs of TNF or TNFR1 knockout mice, but remained unaffected in TNFR2 knockout macrophages (Figures 4d-f). This strengthened the idea that zVAD-fmk/BV6induced necroptosis finally emanates from the TNF/TNFR1 axis. Collectively, our data suggest an anti-necroptotic role of TRAF2-cIAP complexes, along with capsase-8, in macrophages to prevent suicidal autocrine TNF-TNFR1 signaling.
TNFR2 is involved in LPS-induced necroptosis in murine macrophages. TLR3 and TLR4 are not only major inducers of TNF production in macrophages but have also the potential to trigger necroptosis in this cell type. [26][27][28][29] To evaluate whether TNFR2 has a role in TLR-induced necroptosis of macrophages, we performed initial experiments with agonists of TLR3 and TLR4. Although the TLR3 agonist poly(I:C) showed no cytotoxic activity in macrophages differentiated from HoxB8-immortalized MPCs, TLR4 stimulation with LPS in the presence zVAD-fmk resulted in robust cell death (Figures 5a and b). This cytotoxic effect was severely reduced in TNFR2 knockout macrophages indicating a role of the TNF system in TLR4-induced necroptosis. At the first glance, this is contradictory to a former publication arguing for TNF-independent necroptosis by LPS in macrophages, [27][28][29] but another publication reported partly reduced zVAD/LPSinduced cell death. 30 Moreover, a recent study demonstrated that the role of TNF in LPS-induced necroptosis in macrophages is dependent on the LPS concentration used for stimulation (high-concentrations TNF-independent, lowconcentrations TNF-dependent). 31

Discussion
Macrophages are perhaps the best studied primary cell type capable to undergo necroptotic cell death. Indeed, macrophages are strongly responsive to the majority of natural inducers of necroptosis including TNF, interferons and various viral and bacterial PAMPs. 26,27,32,33 Macrophages activated by PAMPs, for example, LPS, are an integral part of the host's first line of defense against invading pathogens. Some bacteria evade macrophage control by inducing cell death of infected macrophages. This occurs by various mechanisms reaching from caspase-dependent apoptosis over IL-1βinduced pyroptosis to RIP1/RIP3-mediated necroptosis. 34 Bacteria-triggered necroptosis in macrophages has been demonstrated for various bacteria producing pore forming Although the pore forming toxin producing bacteria trigger necroptosis by membrane disruption and ion dysregulation, S. Typhimurium-induced macrophage necroptosis was attributed to IFNβ induction and TNFR1/R2-independent RIP1/RIP3mediated necroptosis. 27 The latter is in line with studies demonstrating TLR3-and TLR4-induced necroptosis of macrophages. 28,29,37 Despite the prototypic nature of TNFinduced necroptosis and the overwhelming role of the TNF-TNFR system in macrophage biology, the potential of macrophages to undergo TNF-induced necroptosis has not been evaluated yet in detail. Especially, the specific role of the two TNF receptors in TNF-induced necroptosis in macrophages remained unclear. We, therefore, elucidated in our study the interplay of endogenous TNF with TNFR1 and TNFR2 as well as the individual contribution of the two TNF receptors to TNF-induced necroptosis in murine macrophages.
Our initial observation that exclusive exogenous triggering of TNFR2 in the presence of the caspase-inhibitor zVAD-fmk is fully sufficient to induce robust necroptotic cell death in murine macrophages (Figures 1c and 2) was at the first glance surprising. TNFR2, in contrast to TNFR1, does not interact with RIP1 and has also not been recognized so far as a trigger of necroptotic cell death, although a 'deadly potential' of TNFR2 has been observed before. More than a decade ago, we and others demonstrated that a very few tumor cell lines undergo apoptosis in response to TNFR2 stimulation due to endogenous TNF production and subsequent TNFR1 activation. 38,39 We, therefore, analyzed the involvement of endogenous TNF and TNFR1 in macrophage necroptosis following TNFR2 activation. Indeed, it turned out that (i) TNFR2 induces TNF under necroptotic conditions in macrophages ( Figure 3d) and (ii) that both TNF and TNFR1 deficiency protects against TNC-sc(mu)TNF80/zVAD-fmktriggered necroptosis (Figures 3b and c), indicating that autocrine TNF/TNFR1 signaling is also operative in this scenario. Although the results from the TNF-and TNFR1deficient macrophages nicely fit into the concept that TNFR2 kills macrophages in the presence of zVAD-fmk by induction of TNF and subsequent triggering of TNFR1-mediated necroptosis, this straightforward model cannot be complete. The crucial unanswered question is why selective exogenous stimulation of TNFR1 is not sufficient to trigger macrophage necroptosis in a similar fashion as TNFR2-induced endogenous TNF. In this context, it is important to note that the discussed apoptosis-inducing activity of TNFR2 is not only based on TNF-induction with subsequent TNFR1 activation but also needs a second, per se non-toxic mechanism sensitizing for TNFR1-induced apoptosis. 38,40 The ability of TNFR2 to sensitize for TNFR1-induced apoptosis represents a general TNFR1-TNFR2 cross-talk mechanism and is not only of relevance in cells where triggering of TNFR2 alone results in cell death but also in cell lines and cell types where TNFR2 stimulation alone is not apoptotic. 23,24,[41][42][43][44][45][46] Indeed, a sensitizing effect of TNFR2 on TNFR1-induced cell death has also already been observed in cells undergoing TNFR1-induced necrosis. 47,48 Thus, the superior necroptosis-triggering activity  of TNFR2 might indicate that this sensitizing mechanism is of special relevance for TNF-induced cell death in macrophages. At the molecular level, TNFR2-induced enhancement of TNFR1-induced apoptosis has been traced back to the ability of TNFR2 to deplete/degrade TRAF2-cIAP1 and TRAF2-cIAP2 complexes which in context of TNFR1 signaling inhibit complex II-mediated caspase-8 activation. Since TRAF2 and the cIAPs also inhibit necroptosis, depletion and/or degradation of these molecules may also crucially contribute to zVAD-fmk/TNFR2-induced necroptosis in macrophages. In accordance with this idea, we observed significant reduction of TRAF2 levels after TNFR2 stimulation (Figure 3e). cIAPs fulfill in macrophages an anti-necroptotic function. This has already been recognized upon treatment with the SMAC mimetics/IAP antagonists SM-164 and compound A. 49,50 The effects of SMAC mimetics in macrophages resemble those described for TNFR2 activation in several ways as (i) both trigger degradation of cIAP1 and cIAP2, thereby sensitizing cells for TNFR1-induced cell death and (ii) both activate transcription factors of the NFκB family enabling (in a cell-type dependent manner) autocrine TNF production. In line with this notion, we found that a mixture of zVAD-fmk and BV6, but not BV6 alone, triggered TNF-and TNFR1-dependent necroptosis (Figures 4d and e). In accordance with the fact that BV6 and TNFR2 control TNFR1induced necroptosis at the level of TRAF2 and the cIAPs, TNFR2 deficiency showed no effect on zVAD-fmk/BV6induced necroptosis (Figure 4f). Noteworthy, triggering of macrophage necroptosis by LPS, which stimulates TNF production, also revealed a TNFR2-dependent component (Figure 5b) suggesting that there is crosstalk between TNF induction and direct TLR4-triggered TRIF-mediated necroptosis. Taken together our data suggest that TNFR2, due to its ability to deplete TRAF2-cIAP complexes, is a (patho) physiological modulator of TNFR1-induced necroptosis in macrophages.

Materials and Methods
Compounds, antibodies and reagents. The SMAC mimetic BV6 was synthesized as described in ref. 51  For stimulation of TLR3 and TLR4, high-and low-molecular weight poly(I:C) and LPS form the mouse TLR1-9 agonist kit (# tlrl-kit1mw) from InvivoGen (San Diego, Ca, USA) have been used. Human TNF was a kind gift from Prof. Daniela Männel (University of Regensburg, Germany). The strongly agonistic murine TNFR2specific nonameric murine TNF variant TNC-sc(mu)TNF80 was produced and characterized as described elsewhere (under revision elsewhere). To control TNCsc(mu)TNF80 for its LPS content, the latter was determined with the Pierce LAL Chromogenic Endotoxin Quantitation Kit (Thermo Fisher Scientific) as described in the manual of the manufacturer. If necessary LPS has been removed using the Pierce High Capacity Endotoxin Removal Resin packed in columns according to the recommendations of the supplier. Afterwards samples were controlled for successful LPS removal.

FACS analysis.
To detect cell-surface expressed proteins, cells were harvested and resuspended in PBS. PE-labeled antibodies of the specificity of interest and corresponding isotype control antibodies were added at the dilution recommended by the supplier for 0.5 h on ice. After removal of unbound antibodies (3 × wash with PBS) cells were analyzed with a FACSCalibur (BD Biosciences, Heidelberg, Germany) following standard procedures.
Western blotting. MPCs were seeded in 6-well plates and were differentiated to macrophages. The latter were then subjected to stimulation with the reagents of interest and whole cell lysates were prepared. For this purpose, the cells were scraped into ice-cold PBS with a rubber policeman, pelleted by centrifugation and after resuspension in 4 × Laemmli sample buffer (8% (w/v) SDS, 0.1 M dithiothreitol, 40% (v/v) glycerol, 0.2 M Tris, pH 8.0) supplemented with phosphatase inhibitor cocktail II (Sigma) lysed by sonification (ten pulses) and heating for 5 min at 96°C. Remaining insoluble debris was removed by centrifugation and the resulting whole cell lysates were subjected to SDS-PAGE. Separated proteins were transferred from the gel to nitrocellulose and after blocking of nonspecific binding sites by incubation in Tris-buffered saline with 0.1% (v/v) Tween 20 and 5% (w/v) dry milk, immunoblotting was performed with the primary antibodies of interest, horseradish peroxidaseconjugated secondary antibodies (Dako, Hamburg, Germany) and the ECL western blotting detection reagents and analysis system (Thermo Fisher Scientific).
qPCR. Total RNA was isolated using the RNeasy mini kit from Qiagen (Valencia, CA, USA) according to the protocol of the manufacturer. Two micrograms of total RNA were transcribed into complementary DNA using the high-capacity cDNA reverse transcription kit (Applied Biosystems, Carlsbad, CA, USA). Tnf mRNA levels were quantified using the TaqMan mouse Tnf (Mm00443258_m1) gene expression TNFR2-dependent necroptosis in macrophages D Siegmund et al assay (Applied Biosystems) and an ABI Prism 7900 sequence detector (Applied Biosystems). qRT-PCR reactions were performed in triplicates for each sample of an experiment and normalized to the housekeeping gene Hprt1 (Mm00446968_m1). mRNA levels were calculated using the SDS 2.1 software (Applied Biosystems).
Evaluation of cellular viability. MPCs or macrophages derived thereof (2 × 10 4 /well) were cultivated overnight in 100 μl culture medium in 96-well plates and stimulated the next day in triplicates with the reagents of interest. After 36 h cell viability was determined using the MTT assay or crystal violet staining. For normalization of viability each experiment not only included untreated cells, but also cells that had been treated with a cytotoxic mixture (1 μg/ml Fc-CD95L, 20 μM CHX, 1% (w/v) sodium azide) triggering maximal cell death. The average of the triplicate of cells treated with the cytotoxic mixture was defined as 0% viability and the average of the triplicate of the untreated cells was set to 100% viability. Each measurement was finally normalized according to these values and the averages of the various triplicates were considered as a data point for further analyses using the One-way ANOVA Bonferroni's multiple comparison test function of the GraphPad