Phenytoin inhibits necroptosis

Abstract

Receptor-interacting protein kinases 1 and 3 (RIPK1/3) have best been described for their role in mediating a regulated form of necrosis, referred to as necroptosis. During this process, RIPK3 phosphorylates mixed lineage kinase domain-like (MLKL) to cause plasma membrane rupture. RIPK3-deficient mice have recently been demonstrated to be protected in a series of disease models, but direct evidence for activation of necroptosis in vivo is still limited. Here, we sought to further examine the activation of necroptosis in kidney ischemia-reperfusion injury (IRI) and from TNFα-induced severe inflammatory response syndrome (SIRS), two models of RIPK3-dependent injury. In both models, MLKL-ko mice were significantly protected from injury to a degree that was slightly, but statistically significantly exceeding that of RIPK3-deficient mice. We also demonstrated, for the first time, accumulation of pMLKL in the necrotic tubules of human patients with acute kidney injury. However, our data also uncovered unexpected elevation of blood flow in MLKL-ko animals, which may be relevant to IRI and should be considered in the future. To further understand the mode of regulation of cell death by MLKL, we screened a panel of clinical plasma membrane channel blockers and we found phenytoin to inhibit necroptosis. However, we further found that phenytoin attenuated RIPK1 kinase activity in vitro, likely due to the hydantoin scaffold also present in necrostatin-1, and blocked upstream necrosome formation steps in the cells undergoing necroptosis. We further report that this clinically used anti-convulsant drug displayed protection from kidney IRI and TNFα-induces SIRS in vivo. Overall, our data reveal the relevance of RIPK3-pMLKL regulation for acute kidney injury and identifies an FDA-approved drug that may be useful for immediate clinical evaluation of inhibition of pro-death RIPK1/RIPK3 activities in human diseases.

Fig. 1. MLKL mediates septic and ischemic injury.

a Survival after injection of recombinant human TNFα into wt, RIPK3-ko or MLKL-ko mice. b−c siRNA-mediated knockdown of RIPK3 or MLKL protects murine renal tubular cells (MCT) from TNFα/TWEAK/IFNγ(TTI) + zVAD-fmk (zVAD)-induced necroptosis 24 h after induction of cell death. Western blots for RIPK3 and MLKL indicate the efficiency of the siRNA-mediated knockdown. d−g Head-to-head comparison of RIPK3-deficient mice to MLKL-deficient mice in the model of renal ischemia-reperfusion injury (IRI). Eight-week-old RIPK3-ko and MLKL-ko mice were subjected to 30 min of renal pedicle clamping before the onset of reperfusion. Histological changes (d, scale bars = 50 µm) were quantified (e) 48 h later by employing a renal tubular damage score (TDS, see Methods for details), and functional markers of acute kidney injury (serum urea (f) and serum creatinine (g)) were measured. No statistically significant differences were detected between RIPK3-ko and MLKL-ko mice in any of these models. Statistical significance was indicated: n.s.: not statistically significant, p < 0.05 (*), p < 0.01 (**), p < 0.001 (***), n = 10−17 per group

Fig. 2. Detection of the human relevance of necroptosis and the role of MLKL.

a Human kidney transplant biopsies obtained from patients 4 days following ischemia-reperfusion injury were stained for pMLKL. b Immunofluorescence of a renal biopsy sample taken from a patient with diagnosed crystallopathy. c Freshly isolated hand-picked kidney tubules from either wt or MLKL-ko mice were perfused with the ferroptosis-inducing compound erastin and video-monitored for morphological changes (compare Video S1). Ballooning cells are indicated as spikes on the time charts corresponding to each experiment. Change of color of the time chart indicates the onset of synchronized tubular necrosis (STN) of all cells of the tubules. Note that MLKL-ko mice exhibit highly significantly less ballooning cells and longer time to the onset of STN. d High-resolution intravital microscopy of MLKL-ko and wild-type littermates was employed to measure e velocity and f flow in the peritubular capillaries. In addition, mean arterial pressure (MAP) was quantified. Note the strong reduction in peritubular flow. Statistical significance was indicated: n.s.: not statistically significant, p < 0.05 (*), p < 0.01 (**), p < 0.001 (***). Scale bars = 50 µm

Fig. 3. A cell-based screen identifies phenytoin as an inhibitor of necroptosis.

a, b HT29 cells were treated with TNFα/smac mimetic/zVAD-fmk to induce necrosulfonamide (NSA)-sensitive necroptosis over 4, 16 and 24 h. Cell cultures were split for FACS analysis (a) and western blotting (b). Note that cells that exhibit pMLKL-positivity in western blots (16 h with NSA) did neither lose plasma membrane integrity nor became positive for annexin V between 16 and 24 h. c Hypothesis-based screen for clinically available inhibitors of plasma membrane channels based on 7-AAD/annexinV positivity in L929 cells 48 h following TZ-treatment. Nec-1, Nec-1s, and ponatinib served as positive controls. d Verification of phenytoin in HT29 cells as an inhibitor of TSZ-induced necroptosis in direct comparison with Nec-1s. e Phenytoin prolongs survival after TNFα-induced SIRS. e−h Phenytoin significantly protects from renal ischemia-reperfusion injury. Statistical significance was indicated: n.s.: not statistically significant, p < 0.05 (*), p < 0.01 (**), n = 8 mice per group unless otherwise indicated. Scale bars = 50 µm

Fig. 4. 5-bezyl hydantoin (5bh) prevents necroptosis in vitro and in vivo.

a Murine tubular cells (MCTs) were treated with TNFα/TWEAK/IFNγ plus zVAD (compare Fig. 1b) in the presence of 500 µm 5-bh and PI exclusion was measured. b, c Similar to phenytoin, and necrostatin-1 (which represents another member of the hydantoin family), 5-bh also protects against morphological criteria of renal damage in a standard model of renal IRI. In line with this finding, functional markers of acute kidney injury either demonstrated a non-significant trend towards protection (d) or reached statistical significance in the case of the most often used marker of serum creatinine (e). Statistical significance was indicated when p < 0.05 (*) and p < 0.01 (**), n = 7−8 per group unless otherwise indicated. Scale bars = 50 µm

Fig. 5. Phenytoin prevents the formation of the necrosome.

a HT29 cells were treated with TSZ for 4 h and stained for human pMLKL. Note that phenytoin prevents the phosphorylation of MLKL as effectively as Nec-1s. b Screen for kinase inhibition by phenytoin reveals no major inhibition of any of the kinases investigated, including RIPK1 and RIPK3. c RIPK1 autophosphorylation assay as described in detail in the Methods section. 250 µm phenytoin result in 43% inhibition of RIPK1 autophosphorylation; Nec-1s serves as positive control. d HT-29 cells were treated with TSZ and phenytoin as indicated and NP40 soluble and insoluble fractions were separated before RIPK1 western blotting. e RAW cells were treated with LPS/zVAD-fmk (L/Z), phenytoin, Nec-1s and the RIPK3-inhibitor GSK872 as indicated. NP40 soluble and insoluble fractions were separated and stained for RIPK1 (short and long exposure), RIPK3 and pMLKL. f, g Primary wild-type MEFs were treated with TNF-α (50 ng/ml) in the presence of CHX (200 ng/ml) and zVAD (50 μm), with or without pre-treatment with indicated doses of phenytoin or RIPK1 inhibitor GSK’963 (indicated doses in (f), 5 µm in g) or Nec-1 (50 μm) or RIPK3 inhibitor GSK’872 (5 μm). pMLKL western blotting and RIPK3 immunoprecipitations were performed on lysates after 6 h and examined for necrosome formation by immunoblotting for RIPK1. h L929 cells were seeded on 100 mm plate and then stimulated with hTNF (10 ng/ml) in the presence/absence of zVAD-fmk (10 μm) and phenytoin (500 μm) for 2 and 4 h. Cells were then lysed in NP-40 lysis buffer (0.2% NP-40, 20 mm Tris, 150 mm NaCl and 10% glycerol, at pH 7.5) for 30 min in ice. Samples were centrifuged at 20,000 × g for 15 min and supernatants were incubated with FADD-specific antibody (M-19, Santa Cruz Biotechnology) overnight at 4 °C. Protein A/G beads (Santa Cruz Biotechnology) were added for a further 3 h. The beads were then washed five times with cold lysis buffer and FADD-associated proteins were eluted following incubating the beads in SDS-PAGE loading buffer at 70 °C for 20 min. i, j NIH3T3 + RIPK3-2xFV were incubated in absence or presence of 0.125 mm phenytoin in combination with 10 ng/ml of TNF plus 25 µm zVAD (g) or 2 nm AP-1 (homodimerizer; AP-20187) (h). Cell death was monitored by Sytox Green uptake by using an incucyte Kinetic Live Cell Imager. Alternatively, cell death was prevented by using 30 µm of Nec-1s (RIPK1 inhibitor) or 1 µm GSK’872 (RIPK3 inhibitor), respectively. k KBGFP#2 cells were seeded on six-well plate and then stimulated with hTNF (10 ng/ml) in the presence/absence of phenytoin (either 250 μm or 500 μm) for 24 h. Cells were then trypsinized and washed once with cold PBS. The GFP expression was analyzed by flow cytometry

Fig. 6. An epiphenomenal Ca⁺⁺-current downstream of pMLKL is sensitive to dantrolene.

TNFα/zVAD/SMAC (TZS)-mimetic-induced necroptosis of HT29 cells results in a whole cell current that is sensitive to necrosulfonamide (NSA) and peaks 16 h after induction (a). The NSA-sensitive Ca⁺⁺-current is sensitive to chloride-depletion (b). TZS-induced necroptosis is accompanied by a small-scale Ca⁺⁺-current within 3-4 h after induction of cell death (c) which is sensitive to dantrolene and NSA (d), but dantrolene does not interfere with combined positivity to annexin V and 7-AAD (compare screen in Fig. 3c). In contrast to dantrolene, however, the phenytoin prevents the cell from dying, but does not prevent the ionomycin-induced Ca⁺⁺-signal (e)