OTUD1 deubiquitinase regulates NF-κB- and KEAP1-mediated inflammatory responses and reactive oxygen species-associated cell death pathways

Deubiquitinating enzymes (DUBs) regulate numerous cellular functions by removing ubiquitin modifications. We examined the effects of 88 human DUBs on linear ubiquitin chain assembly complex (LUBAC)-induced NF-κB activation, and identified OTUD1 as a potent suppressor. OTUD1 regulates the canonical NF-κB pathway by hydrolyzing K63-linked ubiquitin chains from NF-κB signaling factors, including LUBAC. OTUD1 negatively regulates the canonical NF-κB activation, apoptosis, and necroptosis, whereas OTUD1 upregulates the interferon (IFN) antiviral pathway. Mass spectrometric analysis showed that OTUD1 binds KEAP1, and the N-terminal intrinsically disordered region of OTUD1, which contains an ETGE motif, is indispensable for the KEAP1-binding. Indeed, OTUD1 is involved in the KEAP1-mediated antioxidant response and reactive oxygen species (ROS)-induced cell death, oxeiptosis. In Otud1−/−-mice, inflammation, oxidative damage, and cell death were enhanced in inflammatory bowel disease, acute hepatitis, and sepsis models. Thus, OTUD1 is a crucial regulator for the inflammatory, innate immune, and oxidative stress responses and ROS-associated cell death pathways.

In this study, we examined the effects of 88 human DUBs on LUBAC-mediated NF-κB activation, and identified OTUD1 (also known as DUBA7) [11], as the most potent down-regulator. We further identified that OTUD1 is involved not only in TNF-αinduced apoptosis and necroptosis but also in KEAP-1-mediated oxidative stress response and reactive oxygen species (ROS)induced cell death, oxeiptosis.

MATERIALS AND METHODS
A detailed method is available in the Online Supplemental Materials.

RESULTS
OTUD1 is a negative regulator of canonical NF-κB signaling To comprehensively explore the DUBs involved in LUBACmediated NF-κB activation, we prepared 88 human DUB cDNAs and analyzed their effects by a luciferase assay in HEK293T cells (Fig. 1a). Taking the NF-κB activity induced by the expression of LUBAC alone as 100%, its co-expression with 16 DUBs, such as USP10, upregulated the NF-κB activity. In contrast, 34 DUBs significantly downregulated the LUBAC-induced NF-κB activation. Since OTUD6A, OTUD2, and OTUD1 had stronger inhibitory effects than those of the known LUBAC-suppressive DUBs, such as CYLD, OTULIN, and A20, we performed another NF-κB luciferase assay by expressing 10-fold higher amounts of LUBAC subunits than those in Fig. 1a. As a consequence, these DUBs dose-dependently suppressed the LUBAC-and/or TNF-α-induced NF-κB activation (Fig. 1b, Supplementary Fig. 1a, b). Since OTUD1 showed the most potent inhibitory effect, we focused on investigating its physiological functions.
Human OTUD1 is composed of an unknown N-terminal region (a.a. , an OTU domain (a.a. 291-446), which contains the active Cys320, and a ubiquitin-interacting motif (UIM, a.a. 457-481) (Fig. 1c) [11]. Although the N-terminal region lacks homology with known domains, it shares similarities with the amino acid sequence of the hypothetical protein Rv1157c of Mycobacterium tuberculosis (Supplementary Fig. 1c). Interestingly, the region contains an abundance of Ala (21.7%), Pro (16.9%), and Gly (9.6%). Therefore, we named the region as Ala-, Pro-, and OTUD1 suppresses canonical NF-κB activation through the catalytic activity and the N-terminal region. a Screening for DUBs that regulate LUBAC-mediated NF-κB activation. Effects of 88 human DUBs on LUBAC-induced NF-κB activation were analyzed by a luciferase assay. b Dose-dependent inhibition by OTUD1 on LUBAC-and TNF-α-induced NF-κB activation. Effects of increasing amounts (0.1, 0.3, and 1.0 μg) of OTUD1 were examined with co-expression of LUBAC or 6 h treatment with 10 ng/ml TNF-α in HEK293T cells. c Domain structure of wild-type (Wt) and mutants of OTUD1. APGR: Ala-, Pro-, and Gly-rich region; OTU ovarian tumor protease, UIM ubiquitin-interacting motif. d Effect of OTUD1 mutants on the LUBAC-induced NF-κB activity. The relative NF-κB activity induced in the presence of Wt or various mutants of OTUD1, and expression levels of OTUD1 and LUBAC subunits are shown. a, b, d Data are shown as mean ± SD by ANOVA post-hoc Tukey test (n = 3 or 4). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, NS not significant. e The N-terminal APGR region is disordered. Intrinsically ordered and disordered segments of OTUD1 were analyzed by DICHOT [12] (https://idp1.force.cs.is.nagoya-u.ac.jp/dichot/). f OTUD1 eluted in the high molecular weight fractions. Gel filtration analyses of lysates prepared from parental and OTUD1 −/− cells, and FLAG-OTUD1-Wt-and FLAG-ΔAPGR-expressing HEK293T cells were performed using a Superdex 200 column. Concentrated fractions were subjected to immunoblotting with the indicated antibodies. *Non-specific signal.
Gly-rich region (APGR). To identify the critical region of OTUD1 in NF-κB inhibition, we constructed various mutants of OTUD1 and performed luciferase assays (Fig. 1c, d, Supplementary Fig. 1d). The OTUD1-Wt strongly suppressed the LUBAC-and TNF-α-induced NF-κB activities and the catalytically inactive OTUD1-CA showed partial NF-κB suppression. The combined mutation with the catalytically inactive APGR-deletion (ΔAPGR-CA) completely abolished the NF-κB-suppressive effect of OTUD1. These results suggested that both the catalytic activity and the APGR play a role in NF-κB suppression.
The protein disorder prediction by DICHOT [12] suggested that APGR is an intrinsically disordered low complexity domain (Fig.  1e). Moreover, the endogenous and transiently expressed FLAGtagged OTUD1 (calculated molecular weight: 51 kDa) in HEK293T cells eluted in broad fractions in a gel filtration analysis, mainly at approximately 330-240 kDa (Fig. 1f). In contrast, FLAG-ΔAPGR (calculated molecular weight: 22 kDa) predominantly eluted at monomer fractions of <44 kDa. These results suggested that the APGR functions as a protein-interaction site.

OTUD1 down-regulates inflammatory cytokine-induced canonical NF-κB activation
The OTUD1 gene is composed of a single exon in both humans and mice, and we constructed OTUD1-knockout (OTUD1 −/− ) HeLa and HEK293T cells and Otud1 −/− -mice ( Supplementary Fig. 2a, b). The genetic depletion of OTUD1 showed no effect on the expression of LUBAC and NF-κB signaling factors ( Supplementary  Fig. 2c). Moreover, OTUD1 is ubiquitously expressed in various mouse tissues ( Supplementary Fig. 2d), and intracellular amounts of K48-linked, K63-linked, and pan-ubiquitin chains were not affected by the genetic ablation of Otud1 in mouse embryonic fibroblasts (MEFs) (Supplementary Fig. 2e).
To examine the cellular functions of OTUD1, we analyzed the TNF-α-induced canonical NF-κB pathway. The genetic ablation of OTUD1 in HEK293T cells enhanced the TNF-α-induced NF-κB luciferase activity ( Supplementary Fig. 2f). In OTUD1 −/− -HeLa cells, TNF-α-induced phosphorylation of IκBα and p65, hallmarks of NF-κB activation, was upregulated as compared to that in the parental cells (Fig. 2a). Moreover, the FLAG-TNF-α precipitation indicated that OTUD1 is recruited to TNFR upon TNF-α stimulation, and the polyubiquitinations of RIP1 and LUBAC in the TNFR signaling complex I were enhanced in OTUD1 −/− -cells as compared to the parental cells (Fig. 2a, Supplementary Fig. 2g). TANK-binding kinase 1 (TBK1) and IKKε reportedly prevent TNF-induced cell death by RIP1 phosphorylation [17]. Importantly, the recruitment and phosphorylation of TBK1 and IKKε were reduced in the TNFRsignaling complex I from OTUD1 −/− cells ( Supplementary Fig. 2h).
Similarly, upon stimulation with IL-1β, the phosphorylation of NF-κB factors and the subsequent degradation of IκBα were enhanced in OTUD1 −/− cells as compared to the parental cells (Fig. 2c). In contrast, the phosphorylation of JNK, a MAP kinase, was not affected in OTUD1 −/− -cells. When K63-polyubiquitinated proteins were captured from cell lysates by K63-TUBE, enhanced high molecular weight smear migrations of interleukin-1 receptorassociated kinase 1 (IRAK1), HOIP, and SHARPIN were detected in IL-1β-treated OTUD1 −/− -HeLa cells (Fig. 2d). In contrast, the M1-TUBE analysis indicated that the M1-ubiquitination of these factors was not increased (Fig. 2e), suggesting that OTUD1 regulates the IL-1β-induced K63-deubiquitination of IRAK1 and LUBAC. After IL-1β-treatment, the mRNA and protein levels of NF-κB targets were enhanced in OTUD1 −/− -cells, as compared to those in parental cells ( Supplementary Fig. 2j, k).
To further investigate the effect of OTUD1 on signal transduction and transcription, we performed RNA-seq analysis. A principal component analysis clearly demonstrated that the OTUD1deficiency affected the transcription of a set of genes (Fig. 2f). The gene ontology analyses revealed the significant upregulation of inflammatory responses and NF-κB-related factors in OTUD1 −/− -HeLa cells (Fig. 2g, h, Supplementary Fig. 3a). These results clearly indicated that OTUD1 is a negative regulator for the IL-1β-induced canonical NF-κB activation pathway.
We further examined the role of OTUD1 using splenic B cells from Wt-and Otud1 −/− -mice, and confirmed that the Otud1deficiency did not affect the CD40-or B cell receptor-mediated induction of NF-κB target genes ( Supplementary Fig. 3b, c). Furthermore, lymphotoxin β-mediated non-canonical NF-κB activation, which is demonstrated by the intranuclear translocation of p52 and RelB, was not affected in Otud1 −/− -MEFs ( Supplementary  Fig. 3d).

OTUD1 activates IFN antiviral signaling
We next examined the antiviral pathway using Otud1 +/+ -and Otud1 −/− -MEFs. Upon stimulation with poly(dA:dT), which activates the IFN antiviral pathway [19], the phosphorylation of TBK1 and IRF3 and the expression of IRF3-target genes were reduced in Otud1 −/− -MEF cells as compared to the Otud1 +/+ -MEFs ( Fig. 3a, b). Similarly, the TLR3 ligand poly(I:C)-induced Fig. 2 OTUD1 downregulates inflammatory cytokine-induced canonical NF-κB activation. a The enhanced ubiquitination and recruitment of NF-κB signaling factors to TNFR complex I. Parental and OTUD1 −/− -HeLa cells were stimulated with 1 μg/ml FLAG-TNF-α for the indicated periods, and cell lysates and anti-FLAG immunoprecipitates were subjected to immunoblotting with the depicted antibodies. b K63-linked polyubiquitination of NF-κB activators is enhanced in OTUD1 −/− cells. Parental and OTUD1 −/− -HeLa cells were stimulated with 20 ng/ml TNF-α for the indicated periods, and then pulled down by K63-TUBE. Samples were subjected to immunoblotting with the indicated antibodies. c The enhanced IL-1β-induced NF-κB activation in OTUD1 −/− cells. Parental and OTUD1 −/− cells were stimulated with 1 ng/ml IL-1β, and analyzed as in a. d, e Alterations of K63-and M1-linked ubiquitination of NF-κB activators in OTUD1 −/− -HeLa cells. Parental and OTUD1 −/− cells were stimulated with 1 ng/ml IL-1β for the indicated periods, and pulled down by K63-TUBE (d) or M1-TUBE (e). Samples were immunoblotted with the indicated antibodies. f, g RNA-seq analysis. Parental and OTUD1 −/− -HeLa cells were stimulated with 10 ng/ml IL-1β for 1 and 3 h. The cells were lysed and subjected to a transcriptome-wide expression analysis using their extracted total RNA. The principal component analysis (f) and heatmap analysis of significantly varied 34 mRNAs, using the cut-off value of 10 (g) was performed. h The enhanced inflammatory responses in OTUD1 −/− -HeLa cells. Taking the cut-off value of 3.0, pathway analyses of 175 up-regulated genes were performed by MSigDB Hallmark 2020, and are listed in ascending order of P-value. phosphorylation of IRF3 and the expression of IRF3-targets were decreased in Otud1 −/− -bone marrow-derived macrophages (BMDMs) and MEFs (Fig. 3c, d, Supplementary Fig. 4a). Upon infection with Sendai virus (SeV), a single-stranded RNA virus, Otud1 −/− -MEFs and -BMDMs showed reduced expression of IRF3-target genes as compared to Otud1 +/+ -cells (Fig. 3e, f). Indeed, RNA-seq analysis of poly(I:C)-treated MEFs revealed that the Otud1-deficiency downregulated the transcription of a set of IFN genes (Fig. 3g, h), and gene ontology analyses demonstrated that the Otud1-deficiency affected the expression of IFN and the RIG-I-like receptor signaling pathway (Fig. 3i,  Supplementary Fig. 4b). Collectively, these results indicated that OTUD1 is a positive regulator for the type I IFN antiviral pathway.
After an 8 h treatment with TNF-α + cycloheximide (TC), the trypan blue-positive dead cells were significantly increased in OTUD1 −/− -HeLa and HEK293T cells (Fig. 4a). The cell survival analysis indicated that OTUD1 −/− -HEK293T cells are more sensitive to cell death than the parental cells after TC-treatment (Fig. 4b). Indeed, the cleavages of PARP, caspase 3, and caspase 8, hallmarks of the apoptosis, were enhanced in TC-treated OTUD1 −/− -HEK293T cells as compared to those in parental cells (Fig. 4c, d). The effective complex II formation, as shown by the enhanced association of caspase 8 with FADD, was detected in OTUD1 −/− -HEK293T cells (Fig. 4d). These results suggested that OTUD1 has an inhibitory effect on the TNF-αinduced apoptosis.
Interestingly, the restoration of OTUD1-Wt, but not the catalytically inactive mutant, rescued ROS resistance and cell viability (Fig. 5l, m), suggesting that the DUB activity of OTUD1 is necessary to resist oxeiptosis. OTUD1 directly bound with KEAP1, and PGAM5 coprecipitated with OTUD1 only in the presence of KEAP1 (Supplementary Fig. 5d). Although the co-precipitation of endogenous AIFM1 with FLAG-OTUD1 was detected by the MS analysis (Fig. 5a, Supplementary Fig. 5a), the co-precipitation of exogenous OTUD1 with AIFM1 was not detected (Supplementary Fig. 5d). Whereas, in Otud1 −/− -MEFs, increased intranuclear AIFM1 was detected concomitantly with the reduced mitochondrial AIFM1 ( Supplementary  Fig. 5e), which may affect mitochondrial oxidative phosphorylation and ROS generation [27]. These results indicated that OTUD1 binds and regulates the K63-ubiquitination of KEAP1 under basal conditions and is important for the efficient antioxidative stress response, ROS production, and oxeiptosis. The increased H 2 O 2mediated ROS-production and reduced cell viability in Otud1 −/ − -MEFs than that in Otud1 +/+ -MEFs were prevented by broad antioxidant scavengers, such as N-acetylcysteine (NAC) and butylated hydroxyanisole (BHA), but not by a mitochondria-targeted superoxide dismutase mimetic, Mito-TEMP ( Supplementary Fig. 5f, g). These results suggested that cytosolic ROS scavengers ameliorate the increased ROS levels and cell death in Otud1-deficient cells.

OTUD1 suppresses inflammatory and oxidative stress responses in vivo
A recent report showed that OTUD1 inhibits colonic inflammation in vivo [28]. We confirmed that significant shortening of the colon, and higher diarrhea and fecal blood scores were observed in a DSS-administrated ulcerative colitis-like inflammatory bowel disease (IBD) model of Otud1 −/− -mice ( Supplementary Fig. 6a, b). The middle and distal colon regions in DSS-treated Otud1 −/− -mice showed more severe damage, characterized by multiple focal dropouts of entire crypts, inflammatory cell infiltration, and edema, as compared with the DSS-treated Otud1 +/+ mice (Fig. 6a). Histologic scores of colitis were significantly higher in the middle and distal colon of DSS-treated Otud1 −/− -mice, as compared with DSS-treated Otud1 +/+ mice (Fig. 6b). Moreover, the expressions of NF-κB target genes were upregulated in the distal colon from Otud1 −/− -mice (Fig. 6c). Importantly, we detected the increased 8-OHdG-positive oxidative DNA damage and TUNEL-positive cell death in the colon of DSS-treated Otud1 −/− -mice (Fig. 6d, e). Furthermore, p65 staining was predominantly observed in the cytoplasm of the colonic epithelial cells and the nuclei of a few inflammatory cells (Supplementary Fig. 6c, Arrows) in the lamina propria of the colonic mucosa in the control mice. Nuclear p65positive colonic epithelial cells ( Supplementary Fig. 6c, Arrowheads) were observed in the remaining colonic mucosa of DSS-treated Otud1 +/+ -and Otud1 −/− -mice. Moreover, increased numbers of nuclear p65-positive colonic epithelial cells and inflammatory cells in the DSS-treated Otud1 −/− -mice were consistent with the severity of colitis. These findings indicated that the canonical NF-κB activation is increased in DSS-induced colitis in the Otud1 −/− -mice, as compared with the Otud1 +/+ -mice. In contrast, we did not detect any drastic effects on K63-, K48-and pan-ubiquitination in the colonic lysates from Otud1 +/+ -and Otud1 −/− -mice, with or without DSS-treatment ( Supplementary Fig. 6d). DSS-treated Otud1 −/− -mice also showed splenomegaly with increased TUNEL-positive splenocytes (Supplementary Fig. 6e, f). These results indicated that OTUD1 suppresses the NF-κB-mediated inflammatory and ROS-induced oxidative damage responses, as well as cell death through specific deubiquitination in an in vivo colitis model.
In contrast to LPS/GalN, LPS challenge alone induces systemic inflammation and sepsis with increases in oxidative stress and NF-κB activation [32,33]. When mice were intraperitoneally administered LPS, we found that the survival of Otud1 −/− -mice was significantly shorter as compared to that of Otud1 +/+ -mice ( Supplementary Fig.  7d). Histopathological findings in the livers are shown in Supplementary Table 1 and Supplementary Fig. 7e. In male mice, mild fatty changes characterized by increased microvacuoles in the cytoplasms of hepatocytes were observed in all LPS-treated Otud1 +/+ -and Otud1 −/− -mice, but not in the Otud1 +/+ and Otud1 −/− controls ( Supplementary Fig. 7e, upper panels). In the female mice, the incidence of hepatocellular death (necrosis and/or apoptosis), predominantly localized in the midzonal area, was increased in the LPS-treated Otud1 −/− -mice as compared with the LPS-treated Otud1 +/+ -mice, albeit without statistical significance (Supplementary Table 1, Supplementary Fig. 7e, lower panels). This suggested that the female Otud1 −/− -mice exhibited higher susceptibility to LPSinduced hepatotoxicity than their wild-type counterparts. Indeed, the AST activities were significantly elevated in LPS-treated Otud1 −/ − -mice, as compared to the LPS-treated Otud1 +/+ -mice (Supplementary Fig. 7f). These results indicated that the genetic ablation of Otud1 enhances septic shock.
OTUD1 is associated with prognosis of kidney cancer Finally, to investigate the involvement of OTUD1 in human diseases, we analyzed cancer databases. We found that low levels of OTUD1 expression are significantly associated with a poor prognosis in renal clear cell carcinoma patients (Fig. 7a). Importantly, the expressions of OTUD1 and KEAP1 were significantly correlated in the normal kidney cortex; however, the positive correlation was lost in the tumor (Fig.  7b). In contrast, the expression levels of OTUD1 and NF-κB1 were correlated in both the normal and tumor kidney cortexes. Thus, we concluded that the orchestration of OTUD1 with KEAP1 functions to suppress kidney cancer in humans, and therefore, OTUD1 is a critical regulator to maintain homeostasis.

DISCUSSION
OTUD1 was initially found as a biomarker of thyroid cancer [34] that reportedly stabilizes the p53 tumor suppressor [35]. OTUD1 Fig. 5 OTUD1 regulates KEAP1-mediated oxidative stress response and ROS-associated cell death. a Identification of OTUD1-interacting proteins. FLAG-tagged OTUD1 was expressed in HEK293T cells and immunoprecipitated with an anti-FLAG antibody, and the coimmunoprecipitants were then analyzed by MS. Proteins with an abundance ratio >50, and values for CUL3 and PGAM5 are shown. b OTUD1 physiologically interacts with KEAP1. HEK293T cell lysates and anti-OTUD1 immunoprecipitates were subjected to immunoblotting with the indicated antibodies. c KEAP1 in OTUD1 −/− cells eluted at a lower molecular weight as compared to that in parental cells. Gel filtration fractions from parental and OTUD1 −/− cells, as shown in Fig. 1f, were immunoblotted with an anti-KEAP1 antibody. d OTUD1 regulates K63ubiquitination of KEAP1. Cell lysates and K63-TUBE precipitates from Otud1 +/+ -and Otud1 −/− -MEFs were immunoblotted with the indicated antibodies. e APGR of OTUD1 is required for KEAP1-binding. Wt-and mutants of Myc-OTUD1 were expressed with HA-KEAP1 in HEK293T cells. Cell lysates and anti-Myc-immunoprecipitates were immunoblotted with the indicated antibodies. f The Kelch domain and the C-terminal region (CTR) of KEAP1 is responsible for OTUD1-binding. A similar analysis as in e was performed using Wt-and mutants of HA-KEAP1 and Myc-OTUD1. g OTUD1 contains an ETGE motif in APGR. Localization of the ETGE motif in OTUD1, and amino acid sequence alignment with NRF2 are shown [26]. h: human, m: mouse, c: chicken, z: zebrafish. h The ETGE motif in OTUD1 is indispensable for KEAP1-binding. Myc-tagged Wt-or ETGE motif-deleted mutant of OTUD1 was expressed with HA-KEAP1-Wt as indicated, and cell lysates and anti-HA immunoprecipitates were immunoblotted with the depicted antibodies. i Enhanced hydrogen peroxide-induced ROS generation in Otud1-deficient cells. Otud1 +/+and Otud1 −/− -MEFs were treated with 0.3 mM H 2 O 2 for the indicated periods, and the intracellular ROS levels were analyzed by a DCFH-DA assay. j Enhanced expression of NRF2 target genes in Otud1 −/− -MEFs. MEFs were treated with or without 0.3 mM H 2 O 2 for 3 h, and qPCR analyses were performed. k Reduced cell viability in Otud1 −/− -MEFs under oxidative stress. MEFs were treated with 0.3 mM H 2 O 2 for the indicated periods, and cell viability was analyzed by Celltiter Glo assay. l DUB activity of OTUD1 is necessary to suppress ROS production. OTUD1-Wt or active site mutant (CA) was restored into Otud1 −/− -MEFs, and ROS levels were analyzed by a DCFH-DA assay after treatment with or without 0.3 mM H 2 O 2 for 2 h. m DUB activity of OTUD1 protects cells from H 2 O 2 -induced death. A similar treatment as in l was performed with or without 0.1 mM H 2 O 2 for 8 h, and cell viability was analyzed by a Celltiter Glo assay. Data are shown as mean ± SD by t-test of each time point (n = 4, i, k) or ANOVA post-hoc Tukey test (n = 4, j, l, m). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, NS not significant.
also upregulates the expression of p21 and Mdm2, thus accelerating apoptosis [35], and cleaves K33-linked polyubiquitin chains from the TGF-β pathway inhibitor SMAD7 [15]. Furthermore, OTUD1 is induced by RNA viruses, which gives rise to the upregulation of Smurf1 [36]. OTUD1 reportedly regulates the Hippo pathway [37]. Furthermore, quite recently OTUD1 was found to suppress IBD by removing the K63 ubiquitin chain from RIP1 [28]. Collectively, these findings indicate that OTUD1 functions as a crucial regulator of cancer progression, antiviral host defense responses, and inflammatory responses.
Although the TNF-α-induced NF-κB activation was upregulated in OTUD1 −/− -cells, the activation of TBK1 and IKKε were suppressed ( Supplementary Fig. 2), and in contrast to the previous report [16], we determined that IRF3-mediated IFN antiviral signaling was downregulated by the deficiency of Otud1 (Fig. 3). Since TBK1 activation and IRF3 phosphorylation were suppressed in Otud1 −/− -MEFs, OTUD1 seems to be primarily involved in the upstream TBK1/IKKε activation in the type I IFN production pathway (Fig. 7c).
In the presence of excess amounts of ROS, KEAP1 induces cell death named oxeiptosis through binding with PGAM5 and AIFM1 [25]. Moreover, upon TNF-α-induced necroptosis, PGAM5 binds necrosomes, composed of RIP1-RIP3-MLKL, which mediate the activation of the mitochondrial fission factor Drp1 and mitochondrial fragmentation [44]. A recent report found that OTUD1 is involved in caspase-dependent and -independent apoptosis by facilitating the intranuclear translocation of AIFM1 and the degradation of MCL1 [27]. We showed that OTUD1 is a regulator of the KEAP1-PGAM5-AIFM1 axis, which affects ROS generation and cell death pathways upon oxidative and inflammatory responses (Figs. 5, 7c and Supplementary Fig. 5).
NF-κB and NRF2/KEAP1 are redox-sensitive transcription factors that cooperate in oxidative stress and inflammatory responses and cell death pathways [45]. The intracellular levels of ROS, such as hydrogen peroxide, are regulated by the activities of thioredoxin reductase 1 and glutathione disulfide reductase [46,47], and OTUD1 seems to be an additional crosstalk regulator of the NF-κB and NRF2/KEAP1 pathways. Furthermore, TNF-α reportedly induces ROS production through the activation of NADPH oxidase 1 complex (Nox1) [48], and RIP1 is necessary for ROS generation at the initiation of caspase-independent cell death [49,50]. In this study, we showed the drastic increase in ROS levels by the TC-and TCZ-treatments, which accompanied the marked cell death in Otud1 −/− -MEFs (Fig.  7c, Supplementary Fig. 5). Free radicals induce oxidative damage, which causes genotoxicity, oxidations, and cell death. Together, KEAP1-NRF2 and NF-κB co-operate the oxidative stress and inflammatory responses, and dysfunctions of these systems are associated with various disorders, such as neurodegeneration and inflammatory diseases [51,52]. This study has demonstrated that OTUD1 is a central regulatory factor in these signal pathways, and the genetic ablation of OTUD1 is associated with inflammatory bowel disease, hepatitis, sepsis, and kidney cancer (Figs. 6, 7 and Supplementary Figs. 6, 7) [16,28]. Therefore, OTUD1 is a critical drug discovery target for the treatment of these diseases.

DATA AVAILABILITY
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.