Introduction

Autoinflammation describes a group of inherited, mostly monogenic disorders with recurrent fever and systemic inflammation in the absence of identifiable infectious agents (Manthiram et al, 2017). The post-translational modification of proteins by ubiquitin plays an essential role in the regulation of immune signaling pathways, in particular in the innate immune response (Zinngrebe et al, 2014). Variants in genes involved in ubiquitination and its reversal process, deubiquitination, have been identified as underlying cause of a new category of autoinflammatory diseases (Aksentijevich & Zhou, 2017; Beck & Aksentijevich, 2019).

Ubiquitination links ubiquitin molecules to substrate proteins, or to one another via the C-terminal carboxyl group of the donor ubiquitin and one of the seven internal lysine (K) residues or the N-terminal methionine (M) 1 of the acceptor ubiquitin. This results in a total of eight different inter-ubiquitin linkage types (Spit et al, 2019). M1-linkages, also known as linear ubiquitin linkages, are assembled in a head-to-tail fashion (Kirisako et al, 2006) by a tripartite protein complex called linear ubiquitin chain assembly complex (LUBAC) consisting of Shank-Associated RH Domain-Interacting Protein (SHARPIN), Heme-Oxidized IRP2 Ubiquitin Ligase 1 (HOIL-1), and HOIL-1-Interacting Protein (HOIP) (Gerlach et al, 2011; Ikeda et al, 2011; Tokunaga et al, 2011).

Ubiquitin linkages are disassembled by so-called deubiquitinating enzymes (DUBs). In 2013, the OTU-deubiquitinase with linear linkage specificity (OTULIN; also known as FAM105B or Gumby) was identified to specifically bind to and hydrolyze linear ubiquitin linkages assembled by LUBAC (Keusekotten et al, 2013; Rivkin et al, 2013).

Dysregulation of linear ubiquitin linkages is associated with numerous human diseases, including immune disorders, cancer, and neurodegeneration (Jahan et al, 2021). Variants affecting OTULIN’s catalytic activity, resulting in increased linear ubiquitin linkages, cause embryonic lethality in mice (Rivkin et al, 2013; Heger et al, 2018). Moreover, two independent groups identified a surplus of linear ubiquitin linkages in humans due to homozygous variants in OTULIN to result in an autoinflammatory disease: OTULIN-Related Autoinflammatory Syndrome (ORAS) or Otulipenia (Damgaard et al, 2016; Zhou et al, 2016). Eight patients with ORAS carrying homozygous missense or premature stop variants in the OTULIN gene have been identified to date (Damgaard et al, 2016, 2019, 2020; Zhou et al, 2016; Nabavi et al, 2019) (Table EV1). All reported patients were born prematurely, showed first signs of disease within weeks after birth, and suffered from fever, nodular panniculitis, failure to thrive, diarrhea, and arthritis accompanied by increased levels of leukocytes, neutrophils, and C-reactive protein (CrP) (Damgaard et al, 2016, 2019, 2020; Zhou et al, 2016; Nabavi et al, 2019). One ORAS patient carrying a homozygous missense mutation in OTULIN (Damgaard et al, 2016) additionally suffered from steatosis and hepatocyte degeneration with abnormal liver values (Damgaard et al, 2020) suggesting that functioning OTULIN is also essential for liver health. This is further supported by the fact that mice with liver-specific deletion of OTULIN show a similar disease phenotype with liver inflammation and apoptosis ultimately leading to formation of hepatocellular carcinoma (Damgaard et al, 2020; Verboom et al, 2020).

In the present study, we identified a 7-year-old boy with compound heterozygosity in OTULIN carrying two different heterozygous variants with one variant on each allele of the OTULIN gene. He suffered from an atypical form of ORAS with late-onset manifesting as a fulminant autoinflammatory episode with sterile abscess formation in different organs including skin, lung, and spleen. By performing structural and biochemical analyses, OTULIN gene deletion and reconstitution experiments with different OTULIN variants in a heterologous cell system and by assessing response of patient-derived fibroblasts and B cells to immune stimuli, we provide characterization of the combined impact of the two different OTULIN variants on OTULIN’s function on both, molecular and functional levels.

Results Sterile abscess formation in a patient with compound-heterozygous missense variants in the OTULIN gene

A 7-year-old male patient of Greek origin was admitted with abdominal pain and subfebrile temperatures. The boy’s psychomotor development was age-appropriate, and he was obese (body weight: 36.7 kg, height: 1.29 m, body mass index (BMI): 22.1 kg/m2 (97th age-specific BMI percentile)). He had previously suffered from a pneumonia at the age of 6 months, an appendicitis at the age of 6 years, and a gluteal abscess which had been difficult to treat. Initially, he presented with leukocytosis (25.5 G/l; normal range: 4.5–13.5 G/l), neutrophilia (14.93 G/l; normal range: 1.8–8 G/l), and highly elevated levels of CrP (241 mg/l; normal range: < 10 mg/l) (Fig 1A). Shortly after admission, he developed spiking fevers with continuously increasing inflammatory parameters (Fig 1A). Treatment with broad-spectrum antibiotics did not influence the course of systemic inflammatory response syndrome. During further course, the patient developed inflammatory lesions on the left and right wrists and the right ankle (Fig 1B). Total body magnetic resonance imaging (MRI) further revealed abscess formation in the left lower pulmonary lobe, in the left axilla, and in the spleen (Fig 1C). The patient was transferred to intensive care unit (ICU) and underwent the following surgical procedures: debridement of lesions on the wrists, axillary dissection, partial resection of lung and pancreas, and splenectomy. Pus was drained from multiple sites of inflammation; however, biopsies and smears remained sterile (Appendix Table S1). All blood cultures, stool cultures, throat and anal swabs, and tracheal fluids remained sterile (Appendix Table S2). Histopathological analysis of the skin (Fig 1D) revealed massive inflammatory infiltrates of the corium, predominantly consisting of granulocytes, monocytes, and macrophages. In the lung, we found partially necrotizing infiltrates with neutrophils, and also, the spleen showed signs of inflammation and necrosis (Fig 1D). Eosinophils or giant cells were not detected. A monoclonal antibody directed against actin to visualize small blood vessels showed disruption of vessel walls by inflammatory cells in all three organs (Fig 1D). Although the patient’s urine was positive for Pneumococcal antigen (Appendix Table S2), gram-positive bacteria were not detectable in biopsies of lung, spleen, and skin (Appendix Table S3).

Figure 1. Sterile abscess formation in a patient with compound-heterozygous missense variants in the OTULIN gene

Blood parameters of patient are depicted. Patient’s skin alterations are depicted. T2-weighted MR images in coronal plane show abscess formation in the left lower pulmonary lobe (upper panel) and in spleen and left axilla (lower panel). Histological sections of patient biopsies stained with hematoxylin and eosin (left panel) or with an actin antibody (right panel). Scale bars: skin: left 1,000 µm, right 100 µm; lung: left 500 µm, right 250 µm; spleen: left 1,000 µm, right 75 µm. Whole Exome Sequencing (WES) and targeted Sanger sequencing identified compound heterozygous variants in OTULIN at position c.258 G > A, p.M86I on the paternal allele and at position c.500 G > C, p.W167S on the maternal allele. The pedigree is depicted. The patient (II-2) is the second child of non-consanguineous parents.

Since no infectious agent was identified and broad-spectrum antibiotics had not improved the patient’s condition, an autoinflammatory syndrome was suspected, and additional treatment with corticosteroids was started on day 14 of in-patient stay (Fig 1A). This resulted in a decline of body temperature and CrP levels and in marked improvement of the patient’s condition. Liver enzymes such as aspartate transaminase (AST), alanine transaminase (ALT), or gamma-glutamyltransferase (GGT) were elevated at this time and returned to normal in the further course of the disease (Fig EV1A). Alkaline phosphatase (AP) (Fig EV1A), but also total bilirubin, prothrombin, and activated partial thromboplastin were within normal range at all times, and an ultrasound examination of the liver during the autoinflammatory episode showed no abnormalities (Fig EV1B, upper panel). During recovery, the patient developed a pneumothorax and suffered from several bleeding duodenal ulcer requiring application of endoclips. No signs of pathology apart from uncharacteristic inflammation in the gastric antrum (Appendix Table S3) were found in biopsies. Two months following admission, the patient was discharged in good condition.

Click here to expand this figure.

Figure EV1. Abnormal liver function test and development of steatosis hepatis grade II

The course of the patient’s liver enzymes AST (aspartate transaminase) and ALT (alanine transaminase) (upper panel) or GGT (γ-glutamyltransferase) and AP (alkaline phosphatase) (lower panel) are depicted along the time axis with the upper limit of normal indicated by red and blue lines for AST and ALT and the normal range in shades of red and blue for GGT and AP, respectively. Liver ultrasound B-mode images of the patient during the initial phase of autoinflammation and about 2 and 5 years thereafter display diffuse increase of liver echogenicity and finally slightly impaired appearance of portal vein wall and diaphragm indicative of steatosis hepatis grade II.

To identify the underlying cause of the severe and life-threatening autoinflammatory episode in this patient, we performed whole exome sequencing (WES) and targeted Sanger sequencing revealing compound heterozygous missense variants in exon 3 (c.258G>A; p.M86I) and exon 5 (c.500G>C; p.W167S) (Fig 1E) of the OTULIN gene, respectively. WES revealed no other homozygous, compound heterozygous or pathogenic variants likely to explain the observed disease phenotype (Appendix Tables S4–S7; for filtering strategy see Appendix Fig S1). The patient inherited the p.M86I variant from his father, whereas his mother is a heterozygous carrier of the p.W167S mutation. The patient is the second child born to non-consanguineous parents (patient II-2; Fig 1F). Both, his parents and siblings are clinically well.

The compound-heterozygous missense variants p.M86I and p.W167S affect OTULIN protein expression and function

Missense or premature stop variants in OTULIN cause ORAS (Damgaard et al, 2016, 2019; Zhou et al, 2016; Nabavi et al, 2019). All published disease-causing variants in OTULIN are homozygous (Damgaard et al, 2016, 2019; Zhou et al, 2016; Nabavi et al, 2019) (Fig EV2A). As the patient’s phenotype differed considerably from published patients with ORAS (Table EV1), we assessed whether the identified compound heterozygosity in the OTULIN gene affected OTULIN protein expression and/or function. OTULIN protein expression was diminished in patient-derived fibroblasts and B cells compared to control (Fig 2A). As loss of OTULIN was shown to result in downregulation of LUBAC components in B cells (Damgaard et al, 2016; Heger et al, 2018), we also determined protein expression of SHARPIN, HOIL-1, and HOIP. Expression of the different LUBAC components remained stable in patient-derived cells (Fig 2A). To assess whether the diminished OTULIN expression was due to compromised antibody binding to OTULIN variant p.M86I, we applied a second commercially available antibody and confirmed equal detection of wildtype (OTULINWT) and variant (OTULINM86I or OTULINW167S) OTULIN protein overexpressed in A549 OTULIN KO cells (Fig EV2B). OTULIN mRNA expression was unchanged in patient-derived fibroblasts and B cells (Fig EV2C). The homozygous OTULIN mutation p.G281R was reported to result in diminished protein expression due to increased OTULIN degradation via the proteasome (Damgaard et al, 2019). The half-life of wildtype and variant OTULIN was similar (Fig EV2D). However, neither co-incubation with the proteasome inhibitor Bortezomib (BTZ) nor with Bafilomycin A1 (Baf A1), an inhibitor of lysosomal protein degradation, was capable of stabilizing recombinant wildtype and variant OTULIN protein upon incubation with CHX (Fig EV2E). MCL-1, a protein with high turnover (Wu et al, 2020), and LC3-II, a marker of autophagosomes (Yoshii & Mizushima, 2017), served as positive controls for BTZ and Baf A1, respectively (Fig EV2E).

Click here to expand this figure.

Figure EV2. Diminished OTULIN protein expression in patient-derived cells is not due to reduced mRNA expression or increased degradation via the proteasome or the lysosome

OTULIN protein is depicted. OTULIN antibodies by Abcam (antigen corresponding to AA 60–158) or by Cell Signaling (recombinant fragment surrounds S76 without spanning M86) were tested in parallel on A549 OTULIN KO cells transfected with different OTULIN constructs as indicated. One representative of two independent experiments is shown. Relative mRNA expression of OTULIN in fibroblasts and B cells with two different primer pairs is depicted. Data are presented as mean ± SD of six independent experiments; dots represent individual experiments performed in three technical replicates; ns, non-significant, unpaired t-test. A549 OTULIN KO cells were transfected with the different OTULIN constructs as indicated. The following day, cells were treated with 50 µg/ml cycloheximide (CHX) for the indicated times, harvested and analyzed by Western blot for OTULIN protein expression. Tubulin served as loading control. One representative of three independent experiments is shown. A549 OTULIN KO cells were transfected with the different OTULIN constructs as indicated. The following day, cells were treated with 50 µg/ml CHX alone or in combination with 1 µM Bortezomib (BTZ) or 1 µM Bafilomycin A1 (Baf A1) for 8 h or left untreated (DMSO). Expression of the proteins indicated was analyzed by Western blot. One representative of three independent experiments is shown.

Figure 2. The compound-heterozygous missense variants p.M86I and p.W167S affect OTULIN protein expression and function

Expression of indicated proteins was determined by Western blot in patient-derived and control fibroblasts and B cells and A549 OTULIN WT and KO cells. FLAG-tagged tandem ubiquitin binding entity (TUBE) assay was performed to pull down linear ubiquitin linkages in patient-derived and control fibroblasts and B cells. One representative (A, B) out of three independent experiments is shown.

Source data are available online for this figure.

One function of OTULIN is the cleavage of linear ubiquitin linkages (Verboom et al, 2021; Weinelt & van Wijk, 2021). Numerous studies demonstrated that OTULIN downregulation or knockout as well as the expression of certain OTULIN variants lead to accumulation of linear ubiquitin linkages in cells (Fiil et al, 2013; Keusekotten et al, 2013; Rivkin et al, 2013; Elliott et al, 2014; Draber et al, 2015; Damgaard et al, 2016; Hrdinka et al, 2016; Zhou et al, 2016; van Wijk et al, 2017; Heger et al, 2018). Thus, we next assessed the level of linear ubiquitin linkages in patient-derived and control cells. To enrich for linear ubiquitin, a FLAG-tagged tandem ubiquitin binding entity (TUBE) reagent was used to pull down linear ubiquitin linkages and associated proteins (Fig 2B). Importantly, linear ubiquitin chains and associated LUBAC components SHARPIN and HOIP were increased in both, patient-derived fibroblasts and B cells (Fig 2B).

Taken together, these results show that the compound heterozygous variants identified by WES in the OTULIN gene compromised OTULIN’s expression and DUB activity in patient-derived cells.

Pathogenic potential of OTULIN variants p.M86I and p.W167S

The potential pathogenicity of the two novel gene variants identified in this study was further analyzed using Ensembl Variant Effect Predictor (McLaren et al, 2016). Sorting Intolerant From Tolerant (SIFT) (Sim et al, 2012) and PolyPhen-2 (Adzhubei et al, 2010) algorithms predicted the maternal, rather than the paternal, variant to affect OTULIN protein function (Table 1). Apart from one intronic variant (Nabavi et al, 2019), all published homozygous variants in the OTULIN gene (Damgaard et al, 2016, 2019; Zhou et al, 2016) were located in the OTU domain of OTULIN containing its catalytic activity (Keusekotten et al, 2013). In agreement, the two novel variants identified in our study are also localized in the OTU domain (Fig EV2A). In the OTULIN 3D-structure (3ZNV, 3ZNZ; Keusekotten et al, 2013), M86 and W167 are, however, located on different ends of the protein (Fig 3A). The catalytic core of OTULIN is composed of N341, H339, and C129 (Fig 3A; pale cyan) (Rivkin et al, 2013). The crystal structure of this catalytic triad was shown to exist in two alternate conformations: “active” and “inhibited” (Keusekotten et al, 2013). The coordination of N341 is a key event in OTULIN activation (Keusekotten et al, 2013). One of the residues coordinating the N341 side chain is Y91 (Keusekotten et al, 2013) (wheat in Fig 3B). Y91, in turn, is involved, through its hydroxyl group, in an intense network of interactions between different OTULIN residues, among them its catalytic residues H339 and N341 (Fig 3B). This network is important for changing OTULIN’s conformation from “inhibitory” to “active” and co-ordinates the catalytic triad residue N341, aligning it to H339 (Keusekotten et al, 2013). Importantly, the aromatic ring of Y91 approaches the sulfur center of M86 to a distance of 4.1 Å (Fig 3B; sulfur in yellow). The sulfur center of sulfur-containing amino acids (M, C) and aromatic side chains of Y, W, or F are involved in close (< 5 Å) and frequent contacts in proteins (sulfur-arene interactions) (Meyer et al, 2003). This interaction is lost in the mutant M86I (Fig 3B; mutant in pale green). We hypothesize that this loss alters the conformation of the catalytic core of OTULIN, enhancing the proportion of the inactive conformation, and thereby reducing the enzyme’s turnover number (kcat). Consistent with this notion is the observation that replacement of Y91 by F resulted in a 20-fold reduction of kcat while not affecting KM (Keusekotten et al, 2013). Moreover, when in complex with M1-diubiquitin (3ZNZ), Y91 is one of OTULIN’s S1' contact sites with the F4 patch of proximal ubiquitin (Fig 3C). The interaction between Y91 and M86, which is lost in the mutant M86I, might, thus, also play a role in binding of OTULIN to linear ubiquitin (Fig 3C). The position of the maternal variant W167 is in close proximity to OTULIN’s helical arm containing the S1 contact site with the I44 patch of distal ubiquitin. This helical arm comprises or adjoins the three residues Y244, L272, and G281, replaced in the known homozygous variants of OTULIN (Damgaard et al, 2016, 2019, 2020; Zhou et al, 2016) (Fig 3D). The mutation of W167 to S results in a replacement of the bulky, apolar indole side chain of W167 by a much smaller, polar hydroxymethyl side chain (Fig 3D; close-up). It is likely, therefore, that the W167S replacement affects the orientation of the helical arm's S1 site, made up of L259, A262, and R263 toward the I44 patch of distal ubiquitin, thereby indirectly interfering with OTULIN’s binding (KM) to linear ubiquitin (Fig 3D).

Table 1. Compound heterozygous variants in the OTULIN gene in a patient with autoinflammation and sterile abscess formation. Nucleotide alteration CDS position AA alteration domain SIFT PolyPhen Paternal allele Chr5: 14678818G>A c.258G>A p.M86I OTU 0.11 0.31 Maternal allele Chr5: 14687661G>C c.500G>C p.W167S OTU 0 0.998 CDS, coding sequence; AA, amino acid; SIFT, Sorting Intolerant From Tolerant (< 0.05 = deleterious); PolyPhen, Polymorphism Phenotyping (> 0.908 “Probably Damaging”, 0.446–0.908 “Possibly Damaging”, < 0.446 “Benign”).

Figure 3. Pathogenic potential of OTULIN variants p.M86I and p.W167S

The overall structure of OTULINWT (3ZNV) is depicted with its catalytic triad (C129, H339, N341; pale cyan) and positions of variants identified in this study (M86, W167). Close-up of M86 (wheat) and M86I (pale green) with the interaction network around OTULIN’s catalytic center. Close-up of OTULINC129A and its interaction with the proximal ubiquitin (Ub). M86, Y91, E95, and R122 of OTULIN and their interaction with each other and with the F4 patch of the proximal Ub are depicted. Interactions (dotted red lines) and water molecules (light blue) < 5 Å are shown. Residues of the catalytic triad (A129, H339, N341) in pale cyan. OTULINW167S in complex with di-Ub. Contact sites of W167 (L237, I241, L273, and R274) in pale yellow. Positions of known homozygous OTULIN substitutions (Y244, L272, G281) in pale green. I44 patch of distal Ub in magenta and its contact sites L259, A262, and R263 in pale yellow. Close-up of W167 and S167: loss of size and gain of polarity from W to S.

Collectively, this structural analysis suggests that both substitutions, M86I and W167S, may interfere with binding of OTULIN to linear ubiquitin (Fig 3C and D). Moreover, M86I might additionally reduce OTULIN’s catalytic turnover number (kcat) (Fig 3B).

OTULIN variants compromise binding of OTULIN to linear ubiquitin and differentially affect OTULIN’s intrinsic stability and catalytic activity

To assess how the variants p.M86I and p.W167S affect OTULIN’s intrinsic thermal stability, we purified recombinant wildtype (OTULINWT) and variant OTULIN (OTULINM86I and OTULINW167S) and determined their melting points (Tm) by means of differential scanning fluorimetry (Fig 4A). While OTULINWT and OTULINM86I had a similar Tm (OTULINWT: 53.7°C, OTULINM86I: 54°C), OTULINW167S unfolded at a significantly lower temperature (Tm 47.4°C) in this assay (Fig 4A). These results are in line with the structural observation that W167 is part of the protein’s hydrophobic core and indicates that replacement with a smaller and more hydrophilic side chain of serine (W167S) influences its integrity. In contrast, the structurally more conservative mutation M86I appears to affect the overall conformation of the protein much less.

Figure 4. OTULIN variants compromise binding of OTULIN to linear ubiquitin and differentially affect OTULIN’s intrinsic stability and catalytic activity

Biochemical characterization of recombinant OTULIN variants by means of differential scanning fluorimetry (DSF) (A) and surface Plasmon resonance (SPR) measurements (B). DSF measurements with OTULINWT, OTULINM86I, or OTULINW167S. Melting temperatures (Tm) are calculated from five independent experiments with standard deviations. *P = 0.011; ****P = 6.64 × 10−18, unpaired t-test. SPR measurements and steady-state binding curves with calculated dissociation constants (Kd) after injection of a concentration series of OTULINC129A, OTULINC129A/M86I, or OTULINC129A/W167S to CM5-immobilized di-ubiquitin chains. Linear ubiquitin linkages isolated from A549 OTULIN KO cells by M1 TUBE assay were incubated with increasing concentrations of recombinant OTULINWT, OTULINM86I, and OTULINW167S or catalytically inactive OTULINC129A as control for 1 h. Afterward, samples were subjected to analysis by Western blot for the indicated proteins. Images are representative of three independent experiments.

To measure the effect of the M86I and W167S mutations on the binding affinity of OTULIN to linear di-ubiquitin chains, surface Plasmon resonance (SPR) measurements were performed. To prevent ubiquitin cleavage during SPR measurements, a catalytically inactive mutant (OTULINC129A) was used. OTULINC129A bound linear di-ubiquitin chains with a Kd of 3.3 µM in this assay, while lower affinities were determined for both variants OTULINC129A/M86I and OTULINC129A/W167S (4.9 µM and 5.1 µM, respectively) (Fig 4B).

To further evaluate to which extent OTULIN’s catalytic activity was impaired by the two variants, we performed a DUB assay using recombinant OTULINWT, OTULINM86I, and OTULINW167S on endogenous linear ubiquitin linkages isolated from A549 OTULIN KO cells (Fig 4C). Whereas OTULINWT was able to cleave all linear ubiquitin chains, residual amounts of linkages were still present in samples incubated with OTULINW167S. Moreover, also in samples treated with OTULINM86I, we detected remains of linear ubiquitin linkages, although to a lesser extent than in OTULINW167S-treated samples. In summary, we find binding of OTULIN to linear ubiquitin to be compromised by both variants; however, protein stability and catalytic activity is most affected by the OTULIN variant p.W167S.

Parental variants both contribute to defective hydrolysis of linear ubiquitin linkages in patient-derived cells, but to different extents

To validate these biochemical findings in a cellular context, we assessed the contribution of both parental variants to defects in hydrolysis of linear ubiquitin linkages in a heterologous cell system. Plasmid vectors encoding wildtype OTULIN (OTULINWT), the paternal p.M86I variant (OTULINM86I) and the maternal p.W167S variant (OTULINW167S) were transiently expressed in A549 OTULIN KO cells. Plasmid vectors encoding both parental variants (OTULINM86I;W167S) and the published variant p.L272P (OTULINL272P) (Damgaard et al, 2016; Zhou et al, 2016) were used for comparison. OTULINWT and OTULINM86I were more efficient in hydrolysis of linear ubiquitin linkages than OTULINW167S or OTULINM86I;W167S when re-expressed in OTULIN-deficient cells (Fig 5A). The amount of linear ubiquitin in cells transfected with OTULINL272P was similar to empty vector (EV)-transfected cells (Fig 5A) indicating that OTULIN’s catalytic activity was most compromised by this variant. The compromised DUB activity of OTULINW167S compared to OTULINWT and OTULINM86I was stable over a range of different DNA concentrations (Fig EV3A). As the double-mutant OTULINM86I;W167S was less efficient in cleaving linear ubiquitin linkages than a mix of OTULINM86I and OTULINW167S (Fig EV3B), we reasoned that OTULINM86I might be able to substitute for the defect of OTULINW167S. Indeed, defective hydrolysis of linear ubiquitin linkages by OTULINW167S was rescued by co-expression of OTULINWT and, importantly, also by OTULINM86I (Fig 5B), at least when equal amounts of OTULINM86I and OTULINW167S were present in cells. When we next transfected A549 OTULIN KO cells with the two constructs OTULINM86I and OTULINW167S in different proportions (Fig 5C), we found a surplus of OTULINW167S to result in increased linear ubiquitin linkages suggesting that OTULINM86I can rescue the defect of OTULINW167S only to some extent. To investigate the regulation of linear ubiquitin linkages by the different OTULIN variants in a setting that is close to the natural situation, we assessed linear ubiquitin linkages in EBV-transformed B cell lines from patient, mother, father and healthy control (Fig 5D). Surprisingly, not only the patient’s B cells but also the mother’s and the father’s B cells had a surplus of linear ubiquitin linkages, although to a lesser extent (Fig 5D). OTULIN protein expression was again diminished in patient-derived B cells, but, interestingly, also in B cells derived from parents, as compared to control (Figs 5D and EV3C; see Fig 2 for comparison), whereas levels of OTULIN mRNA were similar (Fig EV3D). Together, these results suggest that OTULIN’s catalytic activity is compromised by both parental variants, with the maternal variant p.W167S, however, more severely impairing hydrolysis of linear ubiquitin linkages than the paternal variant p.M86I.

Figure 5. Parental variants both contribute to defective hydrolysis of linear ubiquitin linkages in patient-derived cells, but to different extents

A–C. A549 OTULIN KO cells were transfected with empty vector or different OTULIN constructs as indicated and analyzed by Western blot. D. FLAG-tagged tandem ubiquitin binding entity (TUBE) assay was performed in B cells from control, patient, mother, and father to pull down linear ubiquitin linkages. One representative (A–D) of three independent experiments is shown.

Source data are available online for this figure.

Click here to expand this figure.

Figure EV3. The maternal OTULIN variant p.W167S more severely impairs hydrolysis of linear ubiquitin linkages than the paternal OTULIN variant p.M86I

A, B. A549 OTULIN KO cells were transfected with different OTULIN constructs as indicated. After 24 h, cells were lysed and subjected to analysis by Western blot. One representative of three independent experiments is shown. C. OTULIN protein expression in B cells derived from control, patient, mother, and father is shown. Blots (left panel) are representative of five independent experiments analyzed by densitometry (right panel). Data are presented as mean ± SD (dots represent individual experiments); ***P = 0.0008, **P = 0.0069, ordinary one-way ANOVA, Dunnett’s multiple comparisons test. D. Relative mRNA expression of OTULIN in B cells derived from control, patient, mother, and father determined with two different primer pairs is depicted. Data are presented as mean ± SD of four independent experiments; dots represent individual experiments performed in three technical replicates; ns, non-significant, ordinary one-way ANOVA, Dunnett’s multiple comparisons test. TNF signaling is altered in patient-derived cells

Both, mice and humans with defects in OTULIN suffer from autoinflammation mediated by TNF receptor 1 (TNFR1) (Damgaard et al, 2016, 2019; Heger et al, 2018). Importantly, patients with ORAS have successfully been treated with anti-TNF therapy (Damgaard et al, 2016, 2019; Zhou et al, 2016) (Table EV1). Thus, we investigated whether the compound heterozygosity in OTULIN affected TNF signaling in patient-derived cells. We found that TNF stimulation resulted in increased formation of linear ubiquitin linkages in patient-derived fibroblasts as compared to control (Fig 6A). Surprisingly, increased presence of linear ubiquitin linkages and LUBAC components was detectable in the TNFR1-signaling complex (SC) in patient-derived cells as compared to control (Fig 6B) although it was recently shown that recruitment of LUBAC to the TNFR1-SC is compromised when OTULIN’s catalytic activity is perturbed (Heger et al, 2018). OTULIN itself was not recruited to the TNFR1-SC (Fig 6B) in line with published data (Draber et al, 2015; Elliott et al, 2016; Hrdinka et al, 2016). Of note, OTULIN compound heterozygosity only marginally affected nuclear factor kappa B (NF-κB) and mitogen-activated protein kinase (MAPK) signaling activation upon TNF stimulation (Fig 6C). Secretion of interleukin-6 (IL-6) upon stimulation with TNF, however, was lower in patient-derived fibroblasts compared to control (Fig 6D). Lipopolysaccharide (LPS) stimulation resulted in diminished activation of NF-κB signaling (Fig EV4A) along with reduced secretion of IL-6 (Fig EV4B) in patient-derived fibroblasts as compared to control. Moreover, IκBα degradation induced by cluster of differentiation 40 ligand (CD40L) was more pronounced in control than in patient-derived B cells (Fig EV4C). These data are supported by the fact that IκBα, assessed by cycloheximide (CHX) chase analysis, was more stable in patient-derived B cells (Fig 6E) and fibroblasts (Fig EV4D) as compa