Antioxidant and food additive BHA prevents TNF cytotoxicity by acting as a direct RIPK1 inhibitor

The study investigates whether the widely used antioxidant butylated hydroxyanisole (BHA) protects against TNF-induced cell death by scavenging reactive oxygen species (ROS) or through another mechanism. TNF signaling can activate NF-κB/MAPK and induce cell death via RIPK1 kinase-dependent apoptosis or RIPK1/RIPK3/MLKL-mediated necroptosis. Prior studies often used BHA to argue a requirement for ROS in TNF-induced necroptosis, but the exact ROS source and mechanism remained unclear, with conflicting reports and evidence that some antioxidants do not block necroptosis. The authors hypothesize that BHA’s protective effects are not due to ROS scavenging but instead to direct inhibition of RIPK1, thereby questioning the necessity of ROS for TNF-driven necroptosis and apoptosis and reassessing conclusions drawn from BHA usage as an antioxidant probe.

Background literature highlights BHA, BHT, and TBHQ as synthetic phenolic antioxidants used industrially and experimentally to infer roles for ROS in signaling and cell death. Several studies reported ROS generation upon TNF stimulation and during necroptosis, suggesting roles for mitochondrial or NADPH oxidase-derived ROS, and proposed that BHA protection indicates a ROS requirement. One report suggested oxidation of RIPK1 cysteines is necessary for activation and necrosome assembly. However, other antioxidants often fail to block TNF-induced necroptosis, and BHA was proposed to act beyond ROS scavenging (e.g., affecting mitochondrial complex I and lipoxygenases). Additionally, mitochondria-depleted cells can still undergo necroptosis, challenging mitochondrial ROS necessity. The literature is thus conflicting on whether ROS are required for necroptosis and RIPK1 activation.

• Cell systems: Mouse embryonic fibroblasts (MEFs; Ripk3+/+ and Ripk3−/−), mouse dermal fibroblasts (MDFs), L929 cells, human HT-29 and BT549 cells; standard culture conditions. Mycoplasma testing performed; HT-29 authenticated.
• Reagents: TNF (mouse/human, FLAG-tagged), inhibitors (Nec-1s for RIPK1, GSK'872 for RIPK3, TPCA-1 for IKK, GSK8612 for TBK1/IKKε), apoptosis/necroptosis modulators (zVAD-fmk, CHX), ROS scavengers/antioxidants (BHA, 2-BHA, 3-BHA, BHT, TBHQ, NAC, DecylQ, α-tocopherol, Trolox, Ferrostatin-1), kinase inhibitors (Dabrafenib), mitochondrial and lipoxygenase inhibitors (rotenone, NDGA). Antibodies included anti-RIPK1, anti-pSer166/T169 RIPK1 (custom mouse, commercial human), anti-TRADD, anti-FADD, anti-caspase-8, anti-IκBα, loading controls.
• Cell death assays: Sytox Green uptake measured kinetically by plate reader or IncuCyte; percent death calculated relative to max fluorescence after Triton X-100 permeabilization. Conditions: TNF alone (L929), TNF+zVAD (MEFs, MDFs), TNF+IKKi+zVAD (HT-29), TNF+IKKi or TNF+TBK1i (RIPK1-dependent apoptosis), TNF+CHX, etoposide, staurosporine (RIPK1-independent apoptosis), ferroptosis (ML162, erastin).
• HSV1 necroptosis assay: IFNβ pretreatment to induce ZBP1, infection with HSV1 ICP6 RHIM mutant ± zVAD; tested effects of Nec-1s, BHA, GSK'872 and in Ripk3−/− MEFs.
• Biochemistry: TNFR1 complex I immunoprecipitation via FLAG-TNF pulldown; USP21 treatment to remove ubiquitin; immunoblot for RIPK1 autophosphorylation (S166/T169), TRADD, etc. Cytosolic pRIPK1 immunoprecipitation using anti-pS166/T169, with USP21 and phosphatase treatment. Complex IIb/necrosome IP via caspase-8 and immunoblot for RIPK1, FADD, caspase-8.
• In vitro kinase assays: Recombinant RIPK1 (aa 1–479) and RIPK3 (aa 1–439); ADP-Glo kinase assay to measure ATP consumption; tested dose-dependent effects of BHA, BHT, TBHQ, Nec-1s, Dabrafenib; control for assay interference by BHA in ADP-Glo luminescence without kinase.
• Molecular docking: MOE used to dock 2-BHA, 3-BHA, BHT into RIPK1 structures. PDB 4ITH (DLG-out/Glu-out with Nec-1s) and PDB 4NEU (DLG-out/Glu-in with 1-aminoisoquinoline). Site finder to define pockets; analyzed in PyMOL; assessed hydrogen bonding (V76, S161) and hydrophobic interactions (M67, L70, M92).
• In vivo models:
  • TNF-induced shock (SIRS): 10-week-old C57BL/6J females, 16 h fasting; oral gavage corn oil ± BHA or BHT (625 mg/kg) 1 h before i.v. TNF (15 µg per 20 g body weight). Monitored hypothermia and survival over 3 days; humane endpoints defined.
  • Sharpinpdm/cpdm chronic dermatitis model: Littermates fed 5 weeks with control diet or 0.7% w/w BHA or BHT diets starting at 4 weeks. Assessed gross phenotype, histology (H&E), TUNEL on organs, serum LDH and IL-6.
• Statistics: GraphPad Prism for standard analyses; for time-course cell death, linear mixed model with AR(1) covariance in Genstat; for in vivo temperature (two-way ANOVA with Tukey) and survival (log-rank Mantel-Cox). Sample sizes pre-determined (G*Power); randomization; blinding for data analysis.

• BHA specifically protects against RIPK1 kinase-dependent cell death:
  • Robust protection by BHA and Nec-1s in TNF-induced necroptosis in L929, MEFs (TNF+zVAD), MDFs (TNF+zVAD), and HT-29 (TNF+IKKi+zVAD). BHA did not protect RIPK1-independent necroptosis triggered by HSV1/ZBP1; RIPK3 inhibition (GSK'872) or Ripk3 knockout fully protected.
  • BHA mirrored Nec-1s protection in TNF-induced RIPK1-dependent apoptosis (with IKKi or TBK1i), but had no effect on RIPK1-independent apoptosis (TNF+CHX, etoposide, staurosporine).
• Antioxidants other than BHA do not protect RIPK1-dependent death:
  • BHT, NAC, DecylQ, α-tocopherol, Trolox, Ferrostatin-1 failed to protect TNF-induced RIPK1-dependent apoptosis or necroptosis, though they effectively blocked ferroptosis (ML162 or erastin), confirming functional antioxidant activity.
• BHA inhibits cellular activation of RIPK1:
  • BHA blocked RIPK1 autophosphorylation on S166/T169 both at TNFR1 complex I (under IKKi/TBK1i conditions) and in the cytosol during TNF+zVAD-induced necroptosis. Other antioxidants did not affect RIPK1 activation. BHA prevented assembly of complex IIb/necrosome (RIPK1-FADD-caspase-8 association).
• Direct RIPK1 inhibition by BHA:
  • In vitro ADP-Glo assays showed dose-dependent inhibition of RIPK1 by BHA, approximately 50% activity reduction at 100 µM; BHT had no effect. BHA did not inhibit RIPK3 (Dabrafenib did). BHA at 100 µM completely blocked cellular RIPK1 activation and cytotoxicity; partial effects at 50 µM; no effect at 10 µM.
• Structure-activity and docking:
  • Docking predicted 3-BHA (but not 2-BHA or BHT) binds RIPK1 in DLG-out/Glu-out conformation (PDB 4ITH), forming H-bonds with V76 and S161 and hydrophobic interactions with M67, L70, M92, analogous to Nec-1s binding. 2-BHA’s tert-butyl position impedes similar interactions; BHT lacks key H-bonding oxygen and is sterically bulkier.
  • Experimentally, 3-BHA showed superior inhibition of RIPK1 activation and RIPK1-dependent death compared to 2-BHA, while both isomers equally inhibited ferroptosis, indicating comparable ROS scavenging.
  • TBHQ, structurally similar to 3-BHA, inhibited RIPK1 in vitro (~60% reduction at 100 µM) and in cells, blocking RIPK1 activation and RIPK1-dependent apoptosis and necroptosis.
• In vivo effects:
  • Sharpinpdm/cpdm chronic model: BHA diet improved gross appearance but did not significantly reduce tissue TUNEL positivity, serum LDH, or IL-6 compared to control/BHT diets.
  • TNF-induced SIRS: Oral BHA (625 mg/kg) significantly protected mice from TNF-induced hypothermia and lethality versus corn oil or BHT controls; BHT had no benefit, supporting RIPK1 inhibition rather than ROS scavenging as the mechanism.

The findings show that BHA’s protection against TNF-induced necroptosis and apoptosis stems from direct RIPK1 inhibition, not ROS scavenging. This challenges the notion that ROS are required for TNF-driven necroptosis and RIPK1 activation, at least in the studied cell types. The specificity of protection to RIPK1 kinase-dependent modalities, lack of effect by multiple other antioxidants, and direct biochemical inhibition of RIPK1 strongly support this conclusion. Structural docking and isomer-specific effects (3-BHA versus 2-BHA; BHT inactive) elucidate the molecular basis for RIPK1 inhibition and enabled identification of TBHQ as another type III RIPK1 inhibitor. In vivo, oral BHA conferred protection in an acute TNF-induced SIRS model consistent with RIPK1 inhibition, whereas limited efficacy in the Sharpinpdm/cpdm chronic model suggests constraints such as disease severity and pharmacokinetics/bioavailability. Collectively, many prior studies using BHA as an antioxidant probe may need reinterpretation, as effects attributed to ROS depletion could have resulted from RIPK1 inhibition. The results also expand the repertoire of RIPK1 inhibitors and underscore RIPK1’s central role in inflammatory cell death.

This study demonstrates that BHA and the food additive TBHQ act as direct type III inhibitors of RIPK1. BHA protects cells from RIPK1 kinase-dependent apoptosis and necroptosis by blocking RIPK1 activation and complex IIb/necrosome assembly, without affecting RIPK1-independent death. Multiple other antioxidants do not replicate these effects, indicating ROS are dispensable for TNF-induced RIPK1-dependent cell death in the tested contexts. In vivo, oral BHA protects against TNF-induced lethal shock but shows limited benefit in a severe chronic inflammatory model. The work cautions against interpreting BHA’s effects solely as antioxidant-driven and suggests re-evaluation of prior studies using BHA to implicate ROS. Future research should: refine structure-based design of phenolic RIPK1 inhibitors; assess pharmacokinetics and tissue distribution of BHA/TBHQ relative to RIPK1 inhibition in vivo; explore efficacy in additional RIPK1-driven disease models (e.g., NEMO-deficient IEC colitis); and reassess ROS roles in necroptosis using tools not confounded by RIPK1 inhibition.

• Antioxidant panels were broad but cannot exclude cell type- or context-specific contributions of ROS to necroptosis or RIPK1 regulation.
• In silico docking provides a static view and may not fully capture RIPK1 kinase domain flexibility; 2-BHA showed partial inhibitory effects despite predicted poor docking.
• In vitro IC50 estimates for BHA are higher than cellular potency, likely reflecting lipophilicity and local concentration effects; precise intracellular pharmacokinetics were not measured.
• In the Sharpinpdm/cpdm model, limited efficacy may reflect disease severity and/or suboptimal bioavailability; systemic exposure levels of BHA after dietary administration were not quantified.
• Dietary BHA concentrations used in mice exceed regulatory limits for human food, limiting direct translational implications.