Pharmacological inhibition of TBK1/IKKε blunts immunopathology in a murine model of SARS-CoV-2 infection

Abstract

TANK-binding kinase 1 (TBK1) is a key signalling component in the production of type-I interferons, which have essential antiviral activities, including against SARS-CoV-2. TBK1, and its homologue IκB kinase-ε (IKKε), can also induce pro-inflammatory responses that contribute to pathogen clearance. While initially protective, sustained engagement of type-I interferons is associated with damaging hyper-inflammation found in severe COVID-19 patients. The contribution of TBK1/IKKε signalling to these responses is unknown. Here we find that the small molecule idronoxil inhibits TBK1/IKKε signalling through destabilisation of TBK1/IKKε protein complexes. Treatment with idronoxil, or the small molecule inhibitor MRT67307, suppresses TBK1/IKKε signalling and attenuates cellular and molecular lung inflammation in SARS-CoV-2-challenged mice. Our findings additionally demonstrate that engagement of STING is not the major driver of these inflammatory responses and establish a critical role for TBK1/IKKε signalling in SARS-CoV-2 hyper-inflammation.

Fig. 1. Idronoxil inhibits STING, MAVS and TRIF driven responses.

a Interferon [IFN]-β luciferase expression in HEK293T mSTING cells overexpressing cGAS treated O/N with the indicated compounds (n = 3 independent experiments). b IFN-stimulated response element [ISRE]-Luciferase expression in L929 cells after 7 h stimulation with the mouse STING agonist DMXAA at 20 μg/ml and the indicated compounds (n = 3 independent experiment for all except Quercetin n = 2 independent experiments). c IFN-β (right) and IP-10 (left) protein levels in THP-1 monocytes after 6 h stimulation with the human STING agonist GSK3 at 100 nM and treatment with indicated idronoxil (IDX) doses. Right: 2.5 μM IDX was used (n = 3 independent experiments). d RNA sequencing analysis of the top 20 murine ISGs in RNA lysates from Trex1^Q98X bone marrow-derived macrophages [BMDMs] treated O/N with 1.25 μM IDX (heat map of log₂ expression relative to the average across all samples; significant genes are indicated with an asterisk) (n = 3 independent animals [C1–C3]). NT is non-treated. e Immunoblot of wild-type [WT] immortalised bone marrow-derived macrophages [iBMDMs] stimulated with 50 μg/ml DMXAA (STING) for the indicated times ±2.5 μM IDX (1 representative blot of 3 independent experiments shown). Molecular weight markers are shown in kDa. A doublet of STING bands appears upon STING activation. Phosphorylated proteins are indicated with [P]. f Immunoblot of wild-type and Tbk1^KO iBMDMs stimulated with 200 ng/ml LPS (TLR4) ± 2.5 μM IDX treatment for the indicated time (1 representative blot of 3 independent experiments shown). Phosphorylated proteins are indicated with [P]. Molecular weight markers are shown in kDa. g IFN-β luciferase expression in p125HEK cells treated with indicated dose of IDX and stimulated O/N with 50 ng/ml of 3p-hpRNA (RIG-I agonist [RIG-Ia]) (n = 3 independent experiments). h RT-qPCR analyses of IFNB1/18S, RSAD2/18S and IFIT1/18S in RNA lysates from primary human bronchial epithelial cells [PBECs] from 3 independent donors treated with 5 μM IDX or 200 nM MRT, and stimulated 3 h with 1 μg/ml transfected polyI:C (MDA5/RIG-I), 25 μg/ml naked polyI:C (TLR3) or transfected with 2.5 μg/ml ISD70 (cGAS) (n = 3 independent donors). a–c, g, h Data are mean ± s.e.m. c One-way or (h) two-way ANOVA with uncorrected Fisher’s LSD (with single pooled variance) multiple comparisons are shown. a–c, g Non-linear regression analyses are shown. Exact p values are shown for all comparisons. Source data and detailed statistical analyses are provided as a Source Data file.

Fig. 2. Idronoxil inhibits formation of TBK1 signalling complexes.

a Interferon [IFN]-β luciferase expression in HEK293T cells expressing the indicated constructs after O/N incubation (n = 3 independent experiments). b Surface plasmon resonance [SPR] analysis of idronoxil (IDX) binding to recombinant human TBK1 (representative of n = 3 independent experiments). c qPCR amplification signal of kinase concentration after KinomeScan^TM assays, for the indicated IDX concentrations (averaged from n = 2 technical replicates). IKKe is IKKε. d In silico modelling of IDX binding (bottom—green) and STING C-terminal tail (top—yellow, PDB: 6nt9) to TBK1 dimer (the blue and red residues refer to the two different TBK1 units in the structure). Immunoblot of Sting^KO immortalised bone marrow-derived macrophages [iBMDMs] expressing mCitrine-STING, after 1 h stimulation with 50 μg/ml DMXAA (STING) and increasing doses of IDX (e), or 5 μM IDX or 100 nM MRT67307 [MRT] (f). e 1 representative blot of three independent experiments shown. f One representative blot of two independent experiments shown. Pink: pull-down of mCitrine-STING with anti-GFP antibody. Black: whole cell lysates. e, f Phosphorylated proteins are indicated with [P]. Molecular weight markers are shown in kDa. g IP-10 protein levels in wild-type (IC₅₀ = 1.6 μM), Tbk1^KO (IC₅₀ = 0.75 μM), Ikbkε^KO [Ikkε^KO](IC₅₀ = 2 μM) and wild-type [WT] iBMDM cells after O/N stimulation with 50 μg/ml DMXAA (STING) and increasing doses of IDX (n = 3 independent experiments). h HEK-TLR3 cells expressing NF-κB and IFN-β luciferase reporters were treated 6 h with 0.5 μg/ml poly(I:C) [pIC] (TLR3) in the presence of 5 μM IDX or 400 nM MRT (n = 3 independent experiments). a, g, h Data are mean ± s.e.m. a–c, g Non-linear regression analyses are shown. h Two-way ANOVA with uncorrected Fisher’s LSD (with single pooled variance) multiple comparisons are shown. Exact p values are shown for all comparisons. Source data and detailed statistical analyses are provided as a Source Data file.

Fig. 3. Idronoxil limits SARS-CoV-2-driven hyper-inflammation.

a Daily weights of K18-hACE2 mice after intranasal infection (day 0) with 10³ PFU SARS-CoV-2 (n = 6 animals were examined per group in one experiment). Idronoxil [IDX]/vehicle injections were performed on days 3, 4, and 5, with mice culled on day 6 post-infection. Veh is vehicle, SARS is SARS-CoV-2 infected, and Sham is non-infected. b Bronchoalveolar lavage fluid [BALF] cell subsets were enumerated following Modified Giemsa staining to distinguish cell populations (n = 6 animals were examined per group in one experiment). c Histology of fixed lung lobes with Modified Giemsa staining (1 representative image of 6 per group) and (d) associated inflammatory scores (n = 6 animals were examined per group in one experiment). e Collagen deposition around the primary airways determined by Sirius Red/Fast Green staining (1 representative image of 6 per group) and (f) associated measurement of the collagen deposition around the primary airway basement circumference (n = 6 animals were examined per group in one experiment). c, e Scale bars of 50 μm are indicated. g Cytokine quantification in lung homogenates (n = 6 animals were examined per group in one experiment, however one outlier in the IDX + SARS group was removed in these analyses after ROUT outlier analysis). h mRNA transcript levels in lung homogenates. Expression of indicated genes is shown relative to Hprt (n = 6 animals were examined per group in one experiment, however qPCR amplification failed for one animal in the Sham + vehicle group). a, b, d, f–h Data are mean ± s.e.m. a, g, h Two-way or (b, d, f) one-way ANOVA with uncorrected Fisher’s LSD (with single pooled variance) multiple comparisons are shown. Exact p values are shown for all comparisons except for (a) where “ns” is non-significant (p > 0.05). Source data and detailed statistical analyses are provided as a Source Data file.

Fig. 4. TBK1 inhibition limits SARS-CoV-2-driven hyper-inflammation.

a Daily weights of K18-hACE2 mice after intranasal infection (day 0) with 10³ PFU SARS-CoV-2. Idronoxil [IDX]/MRT67307 [MRT]/H151/vehicle injections were performed on days 3, 4, and 5, with mice culled on day 5 post-infection (n = 6 animals were examined per group in one experiment). Veh is vehicle, SARS is SARS-CoV-2 infected, and Sham is non-infected. b Bronchoalveolar lavage fluid [BALF] cell subsets were enumerated following Modified Giemsa staining to distinguish cell populations (n = 6 animals were examined per group in one experiment, however one outlier in the SARS + Vehicle group was removed in the Neutrophils analyses). c Histology of fixed lung lobes with associated inflammatory scores (n = 6 animals were examined per group in one experiment). d Alveolar thickness in the primary airways determined by histology (n = 6 animals were examined per group in one experiment). e Cytokine quantification in lung homogenates (n = 6 animals were examined per group in one experiment, however one outlier in each but the SARS + Vehicle group was removed in the CXCL1 analyses after ROUT outlier analysis). f mRNA transcript levels in lung homogenates. Expression of indicated genes is shown relative to Hprt (n = 6 animals were examined per group in one experiment). a–f Data are mean ± s.e.m. a Two-way or (b–f) one-way ANOVA with uncorrected Fisher’s LSD (with single pooled variance) multiple comparisons are shown. Exact p values are shown for all comparisons except for (a) where the comparisons for days 0, 1, 2 and 3 are provided in Source Data File, and where “ns” is non-significant (p > 0.05). Source data and detailed statistical analyses are provided as a Source Data file.