OTULIN deficiency in ORAS causes cell type-specific LUBAC degradation, dysregulated TNF signalling and cell death

Abstract

The deubiquitinase OTULIN removes methionine-1 (M1)-linked polyubiquitin signals conjugated by the linear ubiquitin chain assembly complex (LUBAC) and is critical for preventing TNF-driven inflammation in OTULIN-related autoinflammatory syndrome (ORAS). Five ORAS patients have been reported, but how dysregulated M1-linked polyubiquitin signalling causes their symptoms is unclear. Here, we report a new case of ORAS in which an OTULIN-Gly281Arg mutation leads to reduced activity and stability in vitro and in cells. In contrast to OTULIN-deficient monocytes, in which TNF signalling and NF-κB activation are increased, loss of OTULIN in patient-derived fibroblasts leads to a reduction in LUBAC levels and an impaired response to TNF Interestingly, both patient-derived fibroblasts and OTULIN-deficient monocytes are sensitised to certain types of TNF-induced death, and apoptotic cells are evident in ORAS patient skin lesions. Remarkably, haematopoietic stem cell transplantation leads to complete resolution of inflammatory symptoms, including fevers, panniculitis and diarrhoea. Therefore, haematopoietic cells are necessary for clinical manifestation of ORAS Together, our data suggest that ORAS pathogenesis involves hyper-inflammatory immune cells and TNF-induced death of both leukocytes and non-haematopoietic cells.

Keywords: TNF signalling; cell death; deubiquitinases; inflammatory disease; ubiquitin.

Figure 1. Mutations in OTULIN in a new case of OTULIN‐related autoinflammatory syndrome (ORAS)

1. A

   Segregation of the inflammatory symptoms (filled symbols) and the c.841G>A substitution in OTULIN in the affected kindred. ○, females; □, males; double lines, consanguineous relationship. Probands I.2 and I.3 are monozygotic twins. Roman numerals indicate generations.

2. B

   Schematic representation of the symptoms and clinical presentation of patient III.2.

3. C

   OTULIN DNA sequence chromatograms showing the homozygous single base substitution (c.841G>A, p.Gly281Arg, arrowhead). Data are representative to two independent experiments.

Figure EV1. Structural analysis of OTULIN^{G 281R} (related to Fig 2)

1. A

   Multiple sequence alignment of the sequence surrounding Gly281 in OTULIN's catalytic domain. The Gly281Arg mutation is indicated by a red arrowhead. Residues interacting with the proximal (orange), distal (maroon) or both Ub (yellow) moieties are indicated.

2. B

   Putty representation of the α11‐α12 loop incorporating the Gly281Arg mutation from the OTULIN^G281R structure (PDB: 6I9C). An increased radius of the cartoon and increase in colours yellow‐red represent increased crystallographic temperature, B factors, and reflect the increased disorder of the α11‐α12 loop relative to the rest of the structure (comparable B factors are only found at the N and C termini where there is no secondary structure).

3. C

   Electron density for the same region as in (B). A weighted 2Fo‐Fc map is shown contoured at 1σ. Residues from the α11‐α12 loop are shown as stick representation. Owing to the increased motility of the α11‐α12 loop, no interpretable electron density was observed for the side chains of Gly218Arg and His282 and only the Cα and backbone amide could be confidently fitted into the electron density map. Other residues in the region had interpretable electron density and several water molecules in the region were sufficiently ordered to be modelled (red cross).

4. D

   Tryptophan fluorescence upon thermal unfolding of OTULINcat^WT and OTULINcat^G281R (1.0 mg/mL) measured by nanodifferential scanning fluorimetry (nano‐DSF). Apparent melting temperatures (T _m) are indicated (dashed lines). Data are representative to two independent experiments.

Figure 2. Mutation of Gly281 to Arg unfolds OTULIN's substrate‐binding site and reduces its catalytic activity

1. A

   Overall structure of OTULIN^G281R (left, PDB: 6I9C) and OTULIN^C129A bound to M1‐linked diUb (middle, PDB: 3ZNZ; Keusekotten et al, 2013) and a superimposition of the two structures (right). The Cα positions of Gly281/Arg281 are shown as spheres.

2. B

   Close‐up view of the M1‐linked Ub‐binding site from OTULIN^G281R (light blue) and OTULIN's catalytic domain (blue) with M1‐linked diUb bound (proximal Ub, orange; distal Ub, red) showing a clash between proximal Ub and His282 upon G281R mutation (arrow).

3. C

   Affinity measurements by fluorescence polarisation (FP) with recombinant catalytically inactive OTULIN^C129A or OTULIN^C129A/G281R and FlAsH‐labelled M1‐linked diUb. Data represent mean ± SD of one experiment performed in technical triplicate. Data are representative of three independent experiments. a.u., arbitrary units. K _D, dissociation constant.

4. D

   M1‐linked tetraUb hydrolysis by recombinant OTULIN^WT and OTULIN^G281R using the indicated OTULIN concentrations and visualised on silver‐stained SDS–PAGE. Data are representative of three independent experiments.

5. E

   First derivative of tryptophan fluorescence upon thermal unfolding of recombinant OTULINcat^WT and OTULINcat^G281R (1.0 mg/ml) measured by nanodifferential scanning fluorimetry (nano‐DSF). Apparent melting temperatures (T _m) are indicated (dashed lines). Data are representative to two independent experiments.

Figure 3. LUBAC degradation and accumulation of M1‐linked Ub in OTULIN^{G 281R} fibroblasts

1. A

   Immunoblot analysis of whole‐cell lysates from untreated primary healthy control and patient fibroblasts. Data are representative of three independent experiments.

2. B

   Immunoblot analysis of whole‐cell lysates from primary healthy control and patient fibroblasts either left untreated or treated with the proteasome inhibitor MG132 (10 μM) for 24 h. Data are representative of two independent experiments.

3. C

   Schematic representation of the AQUA‐MS/MS‐based proteomics approach for quantification of cellular Ub linkage composition.

4. D

   AQUA‐MS/MS data from TUBE‐based purification of cellular polyUb conjugates from untreated primary fibroblasts from a healthy control or patient III.2 harbouring the OTULIN^G281R mutation. K27, K29, linkages could not be detected, and K33 could not be accurately quantified in all samples. Data are representative of two independent experiments (see Fig EV2F and G).

5. E, F

   Immunoblot analysis of whole‐cell lysates from untreated primary healthy control or patient fibroblasts. Data are representative of three independent experiments.

6. G

   Immunoblot (left) and densitometry (right) analysis of IκBα levels in primary healthy control and patient fibroblasts treated with cycloheximide (CHX) (50 μg/ml) as indicated. Data are representative of three independent experiments.

Source data are available online for this figure.

Figure EV2. Analysis of protein and mRNA expression in OTULIN^{G 281R} and quantitative proteomics analysis of the cellular Ub linkage composition (related to Fig 3)

1. A

   Analysis of antibody recognition of recombinant OTULIN proteins. 200 ng recombinant protein was loaded in each lane. Membranes were stained with Ponceau S before immunoblot analysis was performed with the two anti‐OTULIN antibodies used in this study (#14127 from Cell Signaling Technology and ab151117 from Abcam). Data are representative of two independent experiments.

2. B

   Relative mRNA levels of OTULIN, HOIP, HOIL‐1 and SHARPIN from primary healthy control and patient fibroblasts measured by quantitative RT–PCR. Bars represent mean ± SEM of three independent experiments each performed in duplicate. Statistical significance was determined using the Mann–Whitney U‐test.

3. C

   Immunoblot analysis of whole‐cell lysates from Cre‐ERT2‐Otulin ^−/flox inducible knock‐out MEFs treated with 4‐hydroxytamoxifen (4‐OHT) (100 nM) or ethanol (EtOH) alone as indicated. Data are representative of two independent experiments.

4. D

   Immunoblot analysis whole‐cell lysates from primary healthy control and patient fibroblasts either left untreated or treated with the autophagosome inhibitor bafilomycin A1 (BafA) (100 nM) for 24 h. Data are representative of two independent experiments.

5. E

   Immunoblot analysis of endogenous Ub conjugates purified by TUBE pull‐down from untreated healthy control and patient fibroblasts shows HOIP ubiquitination in the patient cells. Data are representative of two independent experiments.

6. F

   Biological replicate of the experiment shown in Fig 3D. AQUA‐MS/MS data from TUBE‐based purification of cellular polyUb conjugates from untreated primary fibroblasts from a healthy control or patient III.2 harbouring the OTULIN^G281R mutation. K27, K29, linkages could not be detected and K33 could not be accurately quantified in all samples.

7. G

   Relative change (OTULIN^G281R/control) in the amount of cellular Ub linkages. Bars represent the data presented in Figs 3D and EV2F individually.

Source data are available online for this figure.

Figure 4. TNF‐induced NF‐κB and MAP kinase activation is impaired in OTULIN^{G 281R} fibroblasts due to reduced LUBAC recruitment to the TNF‐RSC

1. A

   Immunoblot analysis of IκBα phosphorylation and degradation as well as p65/RelA phosphorylation in healthy control or OTULIN^G281R fibroblasts primary fibroblasts in response to stimulation with TNF (10 ng/ml). Data are representative of three independent experiments.

2. B

   Immunoblot analysis of phosphorylation of the MAP kinases p38 and JNK in healthy control or OTULIN^G281R primary fibroblasts in response to stimulation with TNF (10 ng/ml). Data are representative of three independent experiments.

3. C

   Densitometry analysis of p65/RelA phosphorylation as presented in (A) from three independent experiments.

4. D

   ELISA analysis of IL‐8 secretion in response to TNF stimulation (10 ng/ml) in healthy control or OTULIN^G281R primary fibroblasts. Bars represent mean ± SEM (n = 4).

5. E

   Immunoblot analysis of the native TNF‐RSC purified by immunoprecipitation from healthy control or OTULIN^G281R primary fibroblasts. Data are representative of three independent experiments.

Source data are available online for this figure.

Figure 5. OTULIN deficiency in human THP‐1 monocytes leads to increased NF‐κB activation and spontaneous TNF secretion

1. A

   Immunoblot analysis of whole‐cell lysate from untreated human THP‐1 monocytes with stable expression of a non‐targeting control shRNA or and shRNA targeting OTULIN. Data are representative of two independent experiments.

2. B

   Immunoblot (left) and densitometry (right) analysis of IκBα levels in shControl and shOTULIN THP‐1 cells treated with cycloheximide (CHX) (50 μg/ml) as indicated. Data are representative of three independent experiments.

3. C

   ELISA analysis of spontaneous TNF secretion over 96 h in shControl and shOTULIN THP‐1 cells. Data represent mean ± SEM (n = 4) and were analysed using the two‐way ANOVA test of statistical significance with Sidak's correction for multiple comparisons.

4. D

   Immunoblot analysis of IκBα phosphorylation and degradation as well as p65/RelA phosphorylation in shControl and shOTULIN THP‐1 cells in response to stimulation with TNF (10 ng/ml). Data are representative of three independent experiments.

5. E

   Densitometry analysis of p65/RelA phosphorylation as presented in (D) from three independent experiments.

Source data are available online for this figure.

Figure 6. OTULIN deficiency sensitises cells to TNF‐induced apoptosis

1. A–D

   Viability of healthy control or OTULIN^G281R primary fibroblasts (A, B) and shControl and shOTULIN THP‐1 cells (C, D) after 24 h and 6 h, respectively, of treatment with TNF (100 ng/ml), CHX (50 μg/ml), Q‐VD‐OPh (10 μM) and Nec‐1 (10 μM) as indicated was analysed using MTT reduction assays. Results were normalised to untreated samples. Each experiment was performed in duplicate. Bars represent mean ± SEM (A, n = 9; B, n = 6; C, n = 8; D, n = 4) and were analysed using the two‐way ANOVA test of statistical significance with Sidak's correction for multiple comparisons.

2. E

   Increased apoptotic cell death in the skin of ORAS patient III.2. Serial sections of normal skin (top panels) and a skin biopsy from patient III.2 taken at an inflammatory flare (bottom panels) were immunostained for cleaved caspase‐3 (centre panels) or analysed by terminal deoxynucleotidyl transferase dUTP nick end labelling (TUNEL) assay (right panels). Arrowheads indicate cleaved caspase‐3‐positive mesenchymal cells. None of these markers were present in the healthy control skin. H&E, haematoxylin and eosin. Boxes in H&E panels indicate the areas magnified in the cleaved caspase‐3 and TUNEL panels. Scale bars, 100 μm. Data are representative of stained sections from three healthy controls and one biopsy from ORAS patient III.2.

Figure EV3. Analysis of viability of healthy control or OTULIN^{G 281R} primary fibroblasts and shControl or shOTULIN THP‐1 cells (related to Fig 6)

1. A

   Viability of healthy control or OTULIN^G281R primary fibroblasts after 24 h of treatment with CHX (50 μg/ml), staurosporine (1 μM), LPS (100 ng/ml) or poly(I:C) (1 μg/ml) as indicated was analysed using MTT reduction assays. Results were normalised to untreated samples. Each experiment was performed in duplicate. Bars represent mean ± SEM (n = 9) and were analysed using the two‐way ANOVA test of statistical significance with Sidak's correction for multiple comparisons.

2. B

   Viability of healthy control or shControl and shOTULIN THP‐1 cells after 6 h of combination treatment with TNF (100 ng/ml) and CHX (50 μg/mL) as indicated was analysed using MTT reduction assays. Results were normalised to untreated samples. Each experiment was performed in duplicate. Bars represent mean ± SEM (n = 8) and were analysed using the two‐way ANOVA test of statistical significance with Sidak's correction for multiple comparisons.

Figure 7. HSCT ameliorates clinical ORAS symptoms in patient III.2

1. A

   Schematic of the timeline of clinical disease in patient III.2 (see Appendix Clinical Description). Arrows indicate therapeutic interventions. d, day; mo, month; yr, year.

2. B

   Analysis of the blood cell chimerism in patient III.2 after HSCT. Short terminal repeat (STR) analysis was used to determine the percentage of donor cells in peripheral blood. Arrow indicates point of relapse.

3. C

   Photograph of patient III.2 at age ˜10 years at an episode of inflammation caused by delayed etanercept administration. Arrows indicate the erythematous subcutaneous nodules (panniculitis).

4. D

   Model of the cellular effects of OTULIN deficiency in myeloid cells and fibroblasts in ORAS. Hyper‐signalling and TNF secretion in myeloid cells as well as TNF‐induced cell death of both haematopoietic cells and cell types with LUBAC downregulation, e.g. dermal fibroblasts, may contribute to the inflammation, pathogenesis and clinical manifestation of ORAS.