The zinc finger (ZnF) protein A20 is a ubiquitin-editing enzyme encoded by the TNFAIP3 (tumor necrosis factor a-induced protein 3) gene that contains an amino-terminal ovarian tumor (OTU) domain followed by seven ZnF domains. Due to its E3 ligase activity (ZnF4), its deubiquitinase function (OTU) and its ubiquitin-binding domain (ZnF7), A20 regulates the fate of several substrates such as RIPK1 (receptor-interacting serine/threonine-protein kinase 1), NEMO (NF-kappa-B essential modulator), and TRAF6 (TNF receptor-associated factor 6), all converging to the canonical NF-κB pathway (Boone et al., 2004; De et al., 2014; Tokunaga et al., 2012; Wertz et al., 2004). Owing to these various functions, A20 plays a pivotal role in the regulation of inflammation and immunity by inhibiting the canonical NF-κB signaling pathway and by preventing cell death (Lork et al., 2017; Martens and van Loo, 2020). Abnormal expression or function of A20 therefore contributes to the onset of several autoimmune or inflammatory disorders such as Crohn’s disease, systemic lupus erythematosus, rheumatoid arthritis, type 1 diabetes mellitus, psoriasis, and atherosclerosis as shown by genome-wide association studies (reviewed in Malynn and Ma, 2019). Furthermore, germline variations in TNFAIP3 lead to an autoinflammatory disease of autosomal dominant inheritance referred to as A20 haploinsufficiency (HA20; MIM#616744), where insufficient suppression of NF-κB activity is observed (Zhou et al., 2016). HA20 is a Behçet-like autoinflammatory syndrome of early onset associated with variable clinical signs. The hallmark features of HA20 are recurrent fever, painful oral, genital and/or gastrointestinal ulcers with diarrhea, arthralgia, or polyarthritis (Aeschlimann and Laxer, 2019). Since the first description of HA20 (Zhou et al., 2016), more than 25 TNFAIP3 truncating variations have been reported (reviewed in Yu et al., 2020) whereas the pathogenic significance has been confirmed for only a part of the 12 reported missense variations (Aslani et al., 2022; Chen et al., 2020a; Dong et al., 2019; Jiang et al., 2022; Kadowaki et al., 2021; Mulhern et al., 2019; Niwano et al., 2022; Shigemura et al., 2016; Tian et al., 2022). Indeed, a body of genetic and experimental evidence is necessary to determine non-ambiguously the deleterious character of missense variations.

In this study, we report a non-previously described TNFAIP3 missense variation, c.707T>C, p.(Leu236Pro), identified in three patients from the same family. We provide a phenotypic description of all affected patients, as well as cytokine and A20 level quantification in patient-derived samples. Furthermore, we assessed the pathogenicity of this variant, as well as that of previously reported TNFAIP3 missense variations, through in vitro molecular and cellular assays. Notably, by studying the location and the intra-molecular amino acid interactions of Leu236 with the other amino acids involved in pathogenic TNFAIP3 missense variations, we identified a critical region within the OTU domain of A20 where the pathogenic TNFAIP3 missense variations are in close proximity and where the affected residues share common amino acid interactions on the 3D protein structure.

The diagnosis of HA20 is highly challenging due to its heterogeneous clinical presentation and the lack of pathognomonic symptoms. The non-ambiguous pathogenic significance of TNFAIP3 variations is necessary for the confirmation of HA20 diagnosis and adaptation of the clinical care. While in the case of TNFAIP3 nonsense or frameshift variations the deleterious impact is most often established, the pathogenic effect of missense variations is difficult to determine.

Herein, we studied a family in which three patients had a disease phenotype compatible with the diagnosis of HA20 and in whom we identified a novel TNFAIP3 missense variation, the c.707T>C, p.(Leu236Pro) variation. In the current work, we provide solid proof for the pathogenicity of this variation through an in-depth study relying on the analysis of the clinical and biological features of the patients, the intrafamilial segregation of the variation with the disease, the absence of occurrence of the identified variation in the general population, the conservation of the affected residue, and most importantly, on the functional evaluation of the variant through a combination of different assays: expression and stability of the variant protein, study of its ability to inhibit the NF-κB pathway and of its deubiquitination activity. By assessing, through dedicated in silico and in vitro assays, the pathogenicity of other missense variations located in the OTU domain of A20 and that fulfill the same criteria, we identified a critical region in the OTU domain in which residues involved in pathogenic missense variations establish common interactions.

The clinical features and the cytokine profile of the patients presented in our study are in line with those exhibited by previously reported HA20 patients. Indeed, elevated levels of pro-inflammatory cytokines in the plasma and supernatants of LPS-stimulated PBMCs of the patients presented in our study resemble that of previously reported patients harboring TNFAIP3 pathogenic truncating variations (Rajamäki et al., 2018; Zhou et al., 2016). Accordingly, the patients reported herein exhibit typical HA20 phenotypic features: early-onset, recurrent fever, articular symptoms, recurrent oral, genital and/or gastrointestinal manifestations. Nevertheless, HA20 clinical manifestations can vary considerably among patients even among those who carry the same variation (Aeschlimann and Laxer, 2019). Depending on disease severity, colchicine treatment, corticosteroids, or cytokine antagonists can be required for HA20 treatment. Colchicine treatment was satisfactory for the control of the symptoms of individuals II.1 and II.2, suggesting a less severe disease presentation, whereas corticosteroids and azathioprine were required for individual I.1. The difference in disease severity observed in these patients cannot be attributed to the missense nature of the variation. In fact, other HA20 patients carrying pathogenic missense variations were shown to be resistant to colchicine and required steroid therapy (Dong et al., 2019; Kadowaki et al., 2021; Shigemura et al., 2016). Disease management of the patients reported in our study did not require cytokine antagonists. Nevertheless, in light of the elevated IL6 levels in PBMCs supernatant prior to LPS stimulation, the use of IL6 antagonist for the control of the patients’ symptoms could be of therapeutic interest in case the patients develop resistance to their current treatments, as previously reported for the treatment of two HA20 patients (Lawless et al., 2018; Zhou et al., 2016).

Further findings supporting the pathogenic significance of the p.(Leu236Pro) variation include enhanced UPS-dependent degradation of the A20_Leu236Pro protein, which is most likely responsible for the HA20 observed in the patients’ PBMCs. We show that this abnormally enhanced degradation is also observed for another missense variant (p.(Leu275Pro), Aslani et al., 2022) for which functional assays have not been reported yet. Interestingly, the in silico tools predicted high destabilizing effects for both A20_Leu236Pro and A20_Leu275Pro. Altogether, the flow cytometry-based functional assay, in combination with the in silico approach, identifies enhanced recognition by the protein quality control machinery and thus increased degradation as one of the mechanisms explaining the pathogenicity of TNFAIP3 missense variants. Our study thus provides a valuable approach that can be used for assessing newly identified missense variations.

Moreover, our experiments indicate that, in addition to reduced levels of the variant protein, the function of A20_Leu236Pro is altered. Indeed, in an attempt to bypass the decreased protein expression in vitro, we doubled the amounts of mutant A20 plasmid used for transfection in the NF-κB luciferase assay. Although this allowed for an increased inhibition of the canonical NF-κB pathway, the inhibitory function of A20_Leu236Pro was not restored up to the levels of the WT protein. Interestingly, such functional deregulation has also been observed for the pathogenic p.(Cys243Tyr) and p.(Glu192Lys) variants (Kadowaki et al., 2021; Shigemura et al., 2016): a 10-fold protein expression increase was required to compensate the loss-of-function of the p.(Cys243Tyr) variant (Shigemura et al., 2016). As for the p.(Glu192Lys) variant, it does not exhibit reduced expression compared to the WT protein but it loses the ability to sufficiently inhibit the NF-κB pathway (Kadowaki et al., 2021), in line with its location at the putative ubiquitin-binding site of A20 (Lin et al., 2008).

The consequences on the native structure of A20 of the three pathogenic missense variations of the OTU domain (p.(Leu236Pro), p.(Glu192Lys), and p.(Cys243Tyr)) converge. Although Glu192 and Cys243 are distant, Leu236 bridges the interactions between the two regions they belong to on the tertiary structure of A20. Moreover, p.(Leu236Pro) and p.(Cys243Tyr) have common consequences on interactions established by surrounding amino acids such as Asn201, Pro235, Trp238, and Tyr306, suggesting a putative importance of these residues for proper protein expression and/or function.

The OTU domain of A20 harbors the active site essential for the DUB activity of A20, which encompasses the catalytic Cys103, His256, and a third residue for which Asp70 or Thr97 are possible candidates (Komander and Barford, 2008; Lin et al., 2008). Study of the crystal structure of the A20 OTU domain shows that the ubiquitin-binding site of A20 may require the surface formed by α-helices (α4, α7, and α8) and β sheets (β2, β3, β4, and β5) and the loops between them; Glu192 is among the surface residues identified in the ubiquitin-binding region, whereas Leu236 and Cys243 are in the loop between β3 and β4 (Lin et al., 2008). Ubiquitin binding requires interactions between hydrophobic residues of ubiquitin and hydrophobic A20 amino acid, making Leu236 a very suitable candidate for surface amino acids in the ubiquitin-binding domain. Overall, the HA20-causing missense variations in the OTU domain, which alter the proper function and/or expression of the protein, are located in a critical region outside the DUB active site but close to the ubiquitin-binding domain.

Figure 3 - Source Data 1 to 5 contain the original western blots used for figure 3Figure 4 - Source Data 1 to 3 contain the original flow cytometry original data used for figure 4.

1. 1. Aeschlimann FA
   2. Laxer RM

   (2019) A20 Haploinsufficiency (Ha20): clinical phenotypes and disease course of patients with a newly recognised NF-kB-mediated Autoinflammatory disease

   Annals of the Rheumatic Diseases 78:728–735.

   https://doi.org/10.1136/annrheumdis-2018-213359

   • PubMed
   • Google Scholar
2. 1. Aslani N
   2. Asnaashari K
   3. Parvaneh N
   4. Shahrooei M
   5. Sotoudeh-Anvari M
   6. Shahram F
   7. Ziaee V

   (2022) Tnfaip3 Mutation causing Haploinsufficiency of A20 with a Hemophagocytic Lymphohistiocytosis phenotype: a report of two cases

   Pediatric Rheumatology Online Journal 20:78.

   https://doi.org/10.1186/s12969-022-00735-1

   • PubMed
   • Google Scholar
3. 1. Boone DL
   2. Turer EE
   3. Lee EG
   4. Ahmad RC
   5. Wheeler MT
   6. Tsui C
   7. Hurley P
   8. Chien M
   9. Chai S
   10. Hitotsumatsu O
   11. McNally E
   12. Pickart C
   13. Ma A

   (2004) The Ubiquitin-modifying enzyme A20 is required for termination of toll-like receptor responses

   Nature Immunology 5:1052–1060.

   https://doi.org/10.1038/ni1110

   • PubMed
   • Google Scholar
4. 1. Chen Yu
   2. Huang H
   3. He Y
   4. Chen M
   5. Seidler U
   6. Tian D
   7. Xiao F

   (2020a) A20 Haploinsufficiency in a Chinese patient with intestinal Behcet’s disease-like symptoms: A case report

   Frontiers in Immunology 11:1414.

   https://doi.org/10.3389/fimmu.2020.01414

   • Google Scholar
5. 1. Chen Y
   2. Lu H
   3. Zhang N
   4. Zhu Z
   5. Wang S
   6. Li M
   7. Keskin O

   (2020b) Premps: predicting the impact of Missense mutations on protein stability

   PLOS Computational Biology 16:e1008543.

   https://doi.org/10.1371/journal.pcbi.1008543

   • Google Scholar
6. 1. De A
   2. Dainichi T
   3. Rathinam CV
   4. Ghosh S

   (2014) The Deubiquitinase activity of A20 is Dispensable for NF-ΚB signaling

   EMBO Reports 15:775–783.

   https://doi.org/10.15252/embr.201338305

   • PubMed
   • Google Scholar
7. 1. Dong X
   2. Liu L
   3. Wang Y
   4. Yang X
   5. Wang W
   6. Lin L
   7. Sun B
   8. Hou J
   9. Ying W
   10. Hui X
   11. Zhou Q
   12. Liu D
   13. Yao H
   14. Sun J
   15. Wang X

   (2019) Novel heterogeneous Mutation of Tnfaip3 in a Chinese patient with Behçet-like phenotype and persistent EBV Viremia

   Journal of Clinical Immunology 39:188–194.

   https://doi.org/10.1007/s10875-019-00604-9

   • PubMed
   • Google Scholar
8. 1. El Khouri E
   2. Ghoumid J
   3. Haye D
   4. Giuliano F
   5. Drevillon L
   6. Briand-Suleau A
   7. De La Grange P
   8. Nau V
   9. Gaillon T
   10. Bienvenu T
   11. Jacquemin-Sablon H
   12. Goossens M
   13. Amselem S
   14. Giurgea I

   (2021) Wnt/Β-Catenin pathway and cell adhesion deregulation in Csde1-related intellectual disability and autism spectrum disorders. Mol psychiatry

   Molecular Psychiatry 26:3572–3585.

   https://doi.org/10.1038/s41380-021-01072-7

   • Google Scholar
9. 1. Jiang W
   2. Deng M
   3. Gan C
   4. Wang L
   5. Mao H
   6. Li Q

   (2022) A novel Missense Mutation in Tnfaip3 causes Haploinsufficiency of A20

   Cellular Immunology 371:104453.

   https://doi.org/10.1016/j.cellimm.2021.104453

   • PubMed
   • Google Scholar
10. 1. Kadowaki S
    2. Hashimoto K
    3. Nishimura T
    4. Kashimada K
    5. Kadowaki T
    6. Kawamoto N
    7. Imai K
    8. Okada S
    9. Kanegane H
    10. Ohnishi H

    (2021) Functional analysis of novel A20 variants in patients with atypical inflammatory diseases

    Arthritis Research & Therapy 23:52.

    https://doi.org/10.1186/s13075-021-02434-w

    • PubMed
    • Google Scholar
11. 1. Komander D
    2. Barford D

    (2008) Structure of the A20 OTU domain and mechanistic insights into Deubiquitination

    The Biochemical Journal 409:77–85.

    https://doi.org/10.1042/BJ20071399

    • PubMed
    • Google Scholar
12. 1. Kulathu Y
    2. Garcia FJ
    3. Mevissen TET
    4. Busch M
    5. Arnaudo N
    6. Carroll KS
    7. Barford D
    8. Komander D

    (2013) Regulation of A20 and other OTU Deubiquitinases by reversible oxidation

    Nature Communications 4:1569.

    https://doi.org/10.1038/ncomms2567

    • PubMed
    • Google Scholar
13. 1. Lawless D
    2. Pathak S
    3. Scambler TE
    4. Ouboussad L
    5. Anwar R
    6. Savic S

    (2018) A case of adult-onset still’s disease caused by a novel splicing Mutation in Tnfaip3 successfully treated with Tocilizumab

    Frontiers in Immunology 9:1527.

    https://doi.org/10.3389/fimmu.2018.01527

    • PubMed
    • Google Scholar
14. 1. Lin S-C
    2. Chung JY
    3. Lamothe B
    4. Rajashankar K
    5. Lu M
    6. Lo Y-C
    7. Lam AY
    8. Darnay BG
    9. Wu H

    (2008) Molecular basis for the unique Deubiquitinating activity of the NF-kappaB inhibitor A20

    Journal of Molecular Biology 376:526–540.

    https://doi.org/10.1016/j.jmb.2007.11.092

    • PubMed
    • Google Scholar
15. 1. Lork M
    2. Verhelst K
    3. Beyaert R

    (2017) CYLD, A20 and OTULIN Deubiquitinases in NF-ΚB signaling and cell death: so similar, yet so different

    Cell Death and Differentiation 24:1172–1183.

    https://doi.org/10.1038/cdd.2017.46

    • PubMed
    • Google Scholar
16. 1. Malynn BA
    2. Ma A

    (2019) A20: A Multifunctional tool for regulating immunity and preventing disease

    Cellular Immunology 340:103914.

    https://doi.org/10.1016/j.cellimm.2019.04.002

    • PubMed
    • Google Scholar
17. 1. Martens A
    2. van Loo G

    (2020) A20 at the crossroads of cell death, inflammation, and Autoimmunity

    Cold Spring Harbor Perspectives in Biology 12:a036418.

    https://doi.org/10.1101/cshperspect.a036418

    • PubMed
    • Google Scholar
18. 1. Morgan AA
    2. Rubenstein E

    (2013) Proline: the distribution, frequency, positioning, and common functional roles of Proline and Polyproline sequences in the human Proteome

    PLOS ONE 8:e53785.

    https://doi.org/10.1371/journal.pone.0053785

    • PubMed
    • Google Scholar
19. 1. Mulhern CM
    2. Hong Y
    3. Omoyinmi E
    4. Jacques TS
    5. D’Arco F
    6. Hemingway C
    7. Brogan PA
    8. Eleftheriou D

    (2019) Janus kinase 1/2 inhibition for the treatment of Autoinflammation associated with heterozygous Tnfaip3 Mutation

    The Journal of Allergy and Clinical Immunology 144:863–866.

    https://doi.org/10.1016/j.jaci.2019.05.026

    • PubMed
    • Google Scholar
20. 1. Niwano T
    2. Hosoya T
    3. Kadowaki S
    4. Toyofuku E
    5. Naruto T
    6. Shimizu M
    7. Ohnishi H
    8. Koike R
    9. Morio T
    10. Imai K
    11. Yoshida M
    12. Yasuda S

    (2022) An adult case of suspected A20 Haploinsufficiency mimicking Polyarteritis Nodosa

    Rheumatology 61:e337–e340.

    https://doi.org/10.1093/rheumatology/keac308

    • PubMed
    • Google Scholar
21. 1. Rajamäki K
    2. Keskitalo S
    3. Seppänen M
    4. Kuismin O
    5. Vähäsalo P
    6. Trotta L
    7. Väänänen A
    8. Glumoff V
    9. Keskitalo P
    10. Kaarteenaho R
    11. Jartti A
    12. Hautala N
    13. Jackson P
    14. Nordström DC
    15. Saarela J
    16. Hautala T
    17. Eklund KK
    18. Varjosalo M

    (2018) Haploinsufficiency of A20 impairs protein-protein Interactome and leads into Caspase-8-dependent enhancement of Nlrp3 Inflammasome activation

    RMD Open 4:e000740.

    https://doi.org/10.1136/rmdopen-2018-000740

    • Google Scholar
22. 1. Richards S
    2. Aziz N
    3. Bale S
    4. Bick D
    5. Das S
    6. Gastier-Foster J
    7. Grody WW
    8. Hegde M
    9. Lyon E
    10. Spector E
    11. Voelkerding K
    12. Rehm HL
    13. ACMG Laboratory Quality Assurance Committee

    (2015) Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American college of medical Genetics and Genomics and the Association for molecular pathology

    Genetics in Medicine 17:405–424.

    https://doi.org/10.1038/gim.2015.30

    • PubMed
    • Google Scholar
23. 1. Shigemura T
    2. Kaneko N
    3. Kobayashi N
    4. Kobayashi K
    5. Takeuchi Y
    6. Nakano N
    7. Masumoto J
    8. Agematsu K

    (2016) Novel heterozygous C243Y A20/Tnfaip3 Gene Mutation is responsible for chronic inflammation in Autosomal-dominant Behçet’s disease

    RMD Open 2:e000223.

    https://doi.org/10.1136/rmdopen-2015-000223

    • PubMed
    • Google Scholar
24. 1. Tian Y
    2. Wu B
    3. Peng L
    4. Wang J
    5. Shen M

    (2022) Three Chinese Pedigrees of A20 Haploinsufficiency: clinical, cytokine and molecular characterization

    Frontiers in Immunology 13:955079.

    https://doi.org/10.3389/fimmu.2022.955079

    • PubMed
    • Google Scholar
25. 1. Tokunaga F
    2. Nishimasu H
    3. Ishitani R
    4. Goto E
    5. Noguchi T
    6. Mio K
    7. Kamei K
    8. Ma A
    9. Iwai K
    10. Nureki O

    (2012) Specific recognition of linear Polyubiquitin by A20 zinc finger 7 is involved in NF-ΚB regulation

    The EMBO Journal 31:3856–3870.

    https://doi.org/10.1038/emboj.2012.241

    • PubMed
    • Google Scholar
26. 1. Wertz IE
    2. O’Rourke KM
    3. Zhou H
    4. Eby M
    5. Aravind L
    6. Seshagiri S
    7. Wu P
    8. Wiesmann C
    9. Baker R
    10. Boone DL
    11. Ma A
    12. Koonin EV
    13. Dixit VM

    (2004) De-Ubiquitination and Ubiquitin Ligase domains of A20 Downregulate NF-kappaB signalling

    Nature 430:694–699.

    https://doi.org/10.1038/nature02794

    • PubMed
    • Google Scholar
27. 1. Yu MP
    2. Xu XS
    3. Zhou Q
    4. Deuitch N
    5. Lu MP

    (2020) Haploinsufficiency of A20 (Ha20): updates on the Genetics, phenotype, pathogenesis and treatment

    World Journal of Pediatrics 16:575–584.

    https://doi.org/10.1007/s12519-019-00288-6

    • PubMed
    • Google Scholar
28. 1. Zhou Q
    2. Wang H
    3. Schwartz DM
    4. Stoffels M
    5. Park YH
    6. Zhang Y
    7. Yang D
    8. Demirkaya E
    9. Takeuchi M
    10. Tsai WL
    11. Lyons JJ
    12. Yu X
    13. Ouyang C
    14. Chen C
    15. Chin DT
    16. Zaal K
    17. Chandrasekharappa SC
    18. Hanson EP
    19. Yu Z
    20. Mullikin JC
    21. Hasni SA
    22. Wertz IE
    23. Ombrello AK
    24. Stone DL
    25. Hoffmann P
    26. Jones A
    27. Barham BK
    28. Leavis HL
    29. van Royen-Kerkof A
    30. Sibley C
    31. Batu ED
    32. Gül A
    33. Siegel RM
    34. Boehm M
    35. Milner JD
    36. Ozen S
    37. Gadina M
    38. Chae J
    39. Laxer RM
    40. Kastner DL
    41. Aksentijevich I

    (2016) Loss-of-function mutations in Tnfaip3 leading to A20 Haploinsufficiency cause an early-onset Autoinflammatory disease

    Nature Genetics 48:67–73.

    https://doi.org/10.1038/ng.3459

    • PubMed
    • Google Scholar

Decision letter after peer review:

Thank you for submitting your article "Identification of an A20 critical region harboring missense variations that lead to autoinflammation" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Carla Rothlin as the Senior Editor. The following individual involved in review of your submission has agreed to reveal their identity: Daniella Schwartz (Reviewer #1).

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.

Essential revisions:

The three reviewers and the guest reviewing editor agreed that the authors need to do additional experiments to support their conclusions on the effect of the novel p.L236P and other HA20-associated missense variants on protein function:

1. The authors should demonstrate using in vitro Ub-binding experiments in transfected cells that the mutant L236P protein fails to bind Ub proteins. Alternatively, the DUB activity of the mutant protein could be tested both in vitro and in the patient's primary cells.

2. The assays showing the reduced stability of p.L236P A20 compared to WT A20 are convincing. The authors should perform this assay on at least 1-2 other pathogenic missense mutations and one common benign missense variant in the OTU domain.

3. Additional in silico evaluation is recommended for a few other missense mutations, including recently reported L83F, V377I, W356R, and T602S. These variants should be added to Table 1.

Reviewer #1 (Recommendations for the authors):

1) Effects of mutation on protein folding.

The authors present some very nice in silico data; it would be helpful if this data were supported by biochemical folding assays using protein carrying this mutation (p.Leu236Pro) and others in the OTU domain (pathogenic and nonpathogenic) to show that the pathogenic mutations specifically affect protein folding

2) Generalizability of destabilizing mutations to other OTU mutations

The assays showing the reduced stability of p.Leu236Pro A20 and WT A20 (half life, cyclohexamide and MG132 assay) are convincing and well done. To further support the generalizability of this mechanism, it would be helpful to perform this assay on at least 1-2 other pathogenic missense mutations in the OTU domain as well as a benign variant.

3) Missed pathogenic variants.

On my review of the literature (pubmed and infevers.com) I identified several pathogenic or likely pathogenic mutations that were not evaluated by the group. These included:

The following missense mutations in the OTU domain

– L83F

– W356R

And missense outsider the OTU domain

– T602S

These mutations should be in Table 2 and should also be included in the in silico analysis.

Reviewer #2 (Recommendations for the authors):

1. This is a consanguineous family, please show the homozygous gene list, to confirm that no other additional homozygous mutation contributes to the phenotype.

2. Figure 1A, show the sanger sequencing results of patients I.1 and II.1.

3. Figure 2, are the cytokine levels of the patients during flares or during remission?

4. The mutant Leu236Pro A20 has decreased stability, is it possible to explain the mechanisms why the mutation resulted in enhanced UPS dependent degradation?

5. In addition to luciferase assay, please also perform more functional studies to demonstrate the loss-of-function of the variant, such as reduced de-ubiquitination in the mutant protein.

Reviewer #3 (Recommendations for the authors):

I think it may be suitable for a specific clinical journal due to the lack of new findings.

Essential revisions:

The three reviewers and the guest reviewing editor agreed that the authors need to do additional experiments to support their conclusions on the effect of the novel p.L236P and other HA20-associated missense variants on protein function:

1. The authors should demonstrate using in vitro Ub-binding experiments in transfected cells that the mutant L236P protein fails to bind Ub proteins. Alternatively, the DUB activity of the mutant protein could be tested both in vitro and in the patient's primary cells.

As requested, we have conducted additional experiments to assess the deubiquitinase (DUB) function of our mutant. As we no longer have protein samples derived from patients’ PBMCs, we have co-expressed A20 WT or A20 L236P constructs together with the A20 target TRAF6 and K63-linked ubiquitin in HEK293T cells and evaluated the ubiquitination state of TRAF6. This co-immunoprecipitation experiment revealed an increased intensity of the high molecular weight ubiquitin-aggregates smear of TRAF6 in the presence of A20_Leu236Pro compared to the WT condition, thus indicating an altered DUB activity of the A20_Leu236Pro. This experiment has now been described in the Results section (lines 168 to 174 in the Results section) and is represented in a new figure (Figure 3F).

2. The assays showing the reduced stability of p.L236P A20 compared to WT A20 are convincing. The authors should perform this assay on at least 1-2 other pathogenic missense mutations and one common benign missense variant in the OTU domain.

In order to test whether other A20 missense variations than the p.(Leu236Pro) variant also exhibit enhanced proteasomal degradation, we have quantified the levels of A20 proteins upon MG132-induced proteasome inhibition in HEK293T cells. In order to test whether TNFAIP3 missense variations other than the p.(Leu236Pro) variant exhibit enhanced proteasomal degradation, we sought to quantify the levels of several A20 proteins, each protein carrying different reported missense variations upon MG132-induced proteasome inhibition in HEK293T cells. For this purpose, we took advantage of the GFP tag present in the TNFAIP3 constructs to sort the cells by flow cytometry and quantify the intensity of the GFP fluorescent signal in cells treated with MG132 or DMSO as negative control. For this analysis, in addition to A20_WT and A20_Leu236Pro, we included the likely pathogenic Leu275Pro variant, described in Aslani et al. (Aslani et al., 2022) because the in silico PremPS tool predicted a high destabilizing effect of the variation (ΔΔG = 1.80 kcal mol-1 – Table 2) and Leu275Pro protein expression has not been assessed to date. As a positive control for the experiment, we used the construct coding for A20_Leu277* truncated protein (Zhou et al., 2016) which should be recognized by the protein quality control as a misfolded protein. We also chose to include the p.(Asn102Ser) variant, previously suggested to lead to HA20 (Yu Chen et al., 2020). This variant is, however, frequent in the control populations (allele frequency of 0.01236 if considering all populations and of 0.06389 in the Latino population according with gnomAD) and has a low destabilizing predicted effect (ΔΔG = 0.78 kcal mol-1 – Table 2). As a negative control, we chose the polymorphic variant p.(Phe127Cys) (allele frequency = 0.0617 in gnomAD). Fluorescence intensity increased for A20_Leu236Pro, A20_Leu275Pro and the positive control A20_Leu277*, indicating that like A20_Leu236Pro, the A20_Leu275Pro variant exhibits enhanced proteasomal degradation (curve shifted to the right in Figure 4 upper panel and increased fluorescence intensity in Figure 4 lower panel). This observation is specific to pathogenic variants as no fluorescence intensity shift was observed for the non-pathogenic A20_Phe127Cys variant nor for the A20_Asn102Ser variant. Indeed, these results demonstrated the absence of enhanced proteosomal degradation of the A20_Asn102Ser protein allowing to classify this variant as non-pathogenic (Figure 4).

These results have now been added to Results section (lines 217 to 240). They are described in a new figure (Figure 4) and mentioned in the Discussion section (lines 321 to 328).

3. Additional in silico evaluation is recommended for a few other missense mutations, including recently reported L83F, V377I, W356R, and T602S. These variants should be added to Table 1.

We have analyzed additional missense variations and added them to table 2 (table 1 corresponds to the phenotypic features of our patients). These include the T602S variation mentioned by the reviewer as well as the V377M variation (we indeed found the report of a variation involving V377, but it was V377M and not V377I). We have also added all other missense variations that were recently reported in HA20 patients (i.e. T129M, L275Pro and A547T). As for the L83F and the W356R variations, despite their listing in the database Infevers, we did not find the corresponding original data; they were therefore not included in the analysis. We have included the PremPS in silico stability prediction whenever it was possible (presence of the residue in the OTU domain and presence of the residue on the available PDB crystal #3ZJD that we used for the analysis). The manuscript has been modified accordingly (mentioned in blue throughout lines 176 to 215).

Reviewer #1 (Recommendations for the authors):

1) Effects of mutation on protein folding.

The authors present some very nice in silico data; it would be helpful if this data were supported by biochemical folding assays using protein carrying this mutation (p.Leu236Pro) and others in the OTU domain (pathogenic and nonpathogenic) to show that the pathogenic mutations specifically affect protein folding

We do agree that biochemical folding assays would provide an additional element to the set of evidence we provide to the destabilizing effect of the L236P variation. In our study, the folding defect is not directly assessed but strongly supported by the body of evidence we provide: we first observed reduced steady-state protein levels upon transient expression in HEK293T cells which led us to refer to in silico stability prediction tools which revealed a highly destabilizing effect of the Leu236Pro variant. Together, these results called for further investigation of the stability and the degradation of the mutant protein. Therefore, we assessed the stability of the protein by determining its half-life upon protein synthesis inhibition with cycloheximide and we evaluated its degradation by the ubiquitin-proteasome system upon proteasome inhibition with MG132. Overall, these experiments converge towards the targeting of the mutant protein to degradation by the protein quality control system due to its inability to achieve the native state. We hope reviewer #1 finds them convincing as we were unfortunately not able to conduct direct proteins folding assays as they require biophysical expertise or radiolabeling experiments that we do not have access to in our facilities.

As requested for other OTU domain mutants, we have now included the stability prediction for all OTU-domain variants (Table 2) as well as proteasome-degradation assessing for 3 other OTU-domain variants. These results have now been added to the Results section (lines 217 to 240). They are described in a new figure (Figure 4) and mentioned in the Discussion section (lines 321 to 328).

2) Generalizability of destabilizing mutations to other OTU mutations

The assays showing the reduced stability of p.Leu236Pro A20 and WT A20 (half life, cyclohexamide and MG132 assay) are convincing and well done. To further support the generalizability of this mechanism, it would be helpful to perform this assay on at least 1-2 other pathogenic missense mutations in the OTU domain as well as a benign variant.

In order to test whether other A20 missense variations than the A20_Leu236pro variant also exhibit enhanced proteasomal degradation, we have quantified the levels of A20 proteins upon MG132-induced proteasome inhibition in HEK293T cells. In order to test whether TNFAIP3 missense variations other than the Leu236Pro variant exhibit enhanced proteasomal degradation, we sought to quantify the levels of several A20 proteins each proteins carrying different reported missense variations upon MG132-induced proteasome inhibition in HEK293T cells. For this purpose, we took advantage of the GFP tag present in the TNFAIP3 constructs to sort the cells by flow cytometry and quantify the intensity of the GFP fluorescent signal in cells treated with MG132 or DMSO as negative control. For this analysis, in addition to A20 WT and A20 Leu236Pro, we included the likely pathogenic Leu275Pro variant, described in Aslani et al. because the in silico PremPS tool predicted a high destabilizing effect of the variation (ΔΔG = 1.80 kcal mol-1 – Table 2) and Leu275Pro protein expression has not been assessed to date. As a positive control for the experiment, we used the construct coding for A20_Leu277* truncated protein (Zhou et al., 2016) which should be recognized by the protein quality control as a misfolded protein. We also chose to include the p.(Asn102Ser) variant, previously suggested to lead to HA20 (Yu Chen et al., 2020), but is frequent in the control populations (allele frequency of 0.01236 if considering all populations and of 0.06389 in the Latino population according with gnomAD) and with low destabilizing predicted effect (ΔΔG = 0.78 kcal mol-1 – Table 2). As a negative control, we chose the polymorphic variant p.(Phe127Cys) (allele frequency = 0.0617 in gnomAD). Fluorescence intensity increased for A20_Leu236Pro, A20_Leu275Pro and the positive control A20_Leu277*, indicating that like Leu236Pro, the Leu275Pro variant exhibits enhanced proteasomal degradation (curve shifted to the right in Figure 4 upper panel and increased fluorescence intensity in Figure 4 lower panel). No fluorescence intensity shift was observed for the non-pathogenic Phe127Cys variant nor for the Asn102Ser variant. Indeed, these results demonstrated the absence of enhanced proteosomal degradation of the A20_Asn102Ser protein allowing to classify this variant as non-pathogenic (Figure 4).

These results have now been added to Results section (lines 217 to 240). They are described in a new figure (Figure 4) and mentioned in the Discussion section (lines 321 to 328).

3) Missed pathogenic variants.

On my review of the literature (pubmed and infevers.com) I identified several pathogenic or likely pathogenic mutations that were not evaluated by the group. These included:

The following missense mutations in the OTU domain

– L83F

– W356R

And missense outsider the OTU domain

– T602S

These mutations should be in Table 2 and should also be included in the in silico analysis.

We have analyzed additional missense variations and added them to table 2 (table 1 corresponds to the phenotypic features of our patients). These include the T602S variation mentioned by the reviewer. We have also added all other missense variations that were recently reported in HA20 patients (i.e. T129M, L275Pro, V377M and A547T). As for the L83F and the W356R variations, despite their listing in the database Infevers, we did not find the corresponding original data; they were therefore not included in the analysis. We have included the PremPS in silico stability prediction whenever it was possible (presence of the residue in the OTU domain and presence of the residue on the available PDB crystal #3ZJD that we used for the analysis). The manuscript has been modified accordingly (mentioned in blue throughout lines 176 to 215).

Reviewer #2 (Recommendations for the authors):

1. This is a consanguineous family, please show the homozygous gene list, to confirm that no other additional homozygous mutation contributes to the phenotype.

We have now included the list of genes that we screen in our custom sequence capture approach and that could be responsible for autosomal recessive auto-inflammatory diseases (ADA2, IL10, IL10RA, IL10RB, IL1RN, IL36RN, IL6ST, LACC1, LPIN2, MEFV, MVK, NLRP1, OTULIN, TRNT1). We found no mutations in these genes. This information has been added to the Results section (lines 100 to 102).

2. Figure 1A, show the sanger sequencing results of patients I.1 and II.1.

Sanger sequencing of all family members has now been added in supplementary figure S1.

3. Figure 2, are the cytokine levels of the patients during flares or during remission?

At the time of sample collection, Individual I.1 suffered from arthralgia and presented ulcers while individuals II.1 and II.2 were in remission. This information has now been added to the Figure Legend.

4. The mutant Leu236Pro A20 has decreased stability, is it possible to explain the mechanisms why the mutation resulted in enhanced UPS dependent degradation?

Although the identity of the elements specifically involved in A20 protein quality control is, to the best of our knowledge, unknown, the overall mechanisms leading to enhanced UPS-dependent degradation of proteins exhibiting decreased stability rely on the well-established protein quality control system. Indeed, proteins that are unable to achieve the native state – due to a missense variation for instance – are recognized as misfolded and subsequently targeted to degradation. This protein quality control system is composed of molecular chaperones and the ubiquitin proteasome system (UPS). For proteins transiting through the endoplasmic reticulum (like A20), several chaperones recognize misfolded proteins and help their retention in the endoplasmic reticulum, allowing only correctly folded proteins to reach the cytosol. Misfolded proteins are instead ubiquitinated by E3 ligases and targeted to degradation by the proteasome.

5. In addition to luciferase assay, please also perform more functional studies to demonstrate the loss-of-function of the variant, such as reduced de-ubiquitination in the mutant protein.

As requested, we have conducted additional experiments to assess the deubiquitinase (DUB) function of our mutant. As we no longer have protein samples derived from patients’ PBMCs, we have co-expressed A20 WT or A20 L236P constructs together with the A20 target TRAF6 and K63-linked ubiquitin in HEK293T cells and evaluated the ubiquitination state of TRAF6. This co-immunoprecipitation experiment revealed an increased intensity of the high molecular weight ubiquitin-aggregates smear of TRAF6 in the presence of A20_Leu236Pro compared to the WT condition, thus indicating an altered DUB activity of the A20_Leu236Pro. This experiment has now been described in the Results section (lines 168 to 174 in the Results section) and is represented in a new figure (Figure 3F).

Reviewer #3 (Recommendations for the authors):

I think it may be suitable for a specific clinical journal due to the lack of new findings.

For the current version of the paper, we have conducted additional experiments to further prove the pathogenicity of the newly identified variant. Furthermore, we have assessed the proteasomal degradation of a series of variants in addition to the one we report in our study. For this purpose, we have generated several additional constructs and developed a new flow cytometry assay (new Figure 4). This functional assay, in combination with a dedicated in silico approach, identifies the enhanced recognition of the mutated proteins by the protein quality control machinery and thus the increased degradation as one of the mechanisms explaining the pathogenicity of TNFAIP3 missense variants. With all the additional experiments we have performed in response to the editor and the reviewers, we hope that Reviewer # 3 is now convinced that the current version of the manuscript indeed provides new findings to the scientific community.

Author details

1. Elma El Khouri

   Sorbonne Université, Institut National de la Santé et de la Recherche Médicale (INSERM), "Maladies génétiques d’expression pédiatrique", Paris, France

   Contribution

   Conceptualization, Data curation, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing – original draft

   Competing interests

   No competing interests declared

2. Farah Diab

   Sorbonne Université, Institut National de la Santé et de la Recherche Médicale (INSERM), "Maladies génétiques d’expression pédiatrique", Paris, France

   Contribution

   Conceptualization, Formal analysis, Investigation, Writing – original draft

   Competing interests

   No competing interests declared

3. Camille Louvrier

   1. Sorbonne Université, Institut National de la Santé et de la Recherche Médicale (INSERM), "Maladies génétiques d’expression pédiatrique", Paris, France
   2. Département de Génétique Médicale, Hôpital Armand Trousseau, Assistance Publique-Hôpitaux de Paris, Paris, France

   Contribution

   Conceptualization, Formal analysis, Methodology, Writing – review and editing

   Competing interests

   No competing interests declared

4. Eman Assrawi

   Sorbonne Université, Institut National de la Santé et de la Recherche Médicale (INSERM), "Maladies génétiques d’expression pédiatrique", Paris, France

   Contribution

   Conceptualization, Formal analysis, Methodology, Writing – review and editing

   Competing interests

   No competing interests declared

5. Aphrodite Daskalopoulou

   Sorbonne Université, Institut National de la Santé et de la Recherche Médicale (INSERM), "Maladies génétiques d’expression pédiatrique", Paris, France

   Contribution

   Investigation

   Competing interests

   No competing interests declared

6. Alexandre Nguyen

   Département de Médecine Interne et Immunologie Clinique, Normandie Univ, UNICAEN, UR4650 PSIR, CHU de Caen Normandie, Caen, France

   Contribution

   Resources

   Competing interests

   No competing interests declared

7. William Piterboth

   Département de Génétique Médicale, Hôpital Armand Trousseau, Assistance Publique-Hôpitaux de Paris, Paris, France

   Contribution

   Formal analysis, Investigation

   Competing interests

   No competing interests declared

8. Samuel Deshayes

   Département de Médecine Interne et Immunologie Clinique, Normandie Univ, UNICAEN, UR4650 PSIR, CHU de Caen Normandie, Caen, France

   Contribution

   Resources, Writing – review and editing

   Competing interests

   No competing interests declared

9. Alexandra Desdoits

   Service de chirurgie pédiatrique, CHU de Caen Normandie, Caen, France

   Contribution

   Resources

   Competing interests

   No competing interests declared

10. Bruno Copin

    Département de Génétique Médicale, Hôpital Armand Trousseau, Assistance Publique-Hôpitaux de Paris, Paris, France

    Contribution

    Software, Methodology

    Competing interests

    No competing interests declared

11. Florence Dastot Le Moal

    Département de Génétique Médicale, Hôpital Armand Trousseau, Assistance Publique-Hôpitaux de Paris, Paris, France

    Contribution

    Investigation, Methodology

    Competing interests

    No competing interests declared

12. Sonia Athina Karabina

    Sorbonne Université, Institut National de la Santé et de la Recherche Médicale (INSERM), "Maladies génétiques d’expression pédiatrique", Paris, France

    Contribution

    Conceptualization, Funding acquisition, Methodology, Writing – review and editing

    Competing interests

    No competing interests declared

13. Serge Amselem

    1. Sorbonne Université, Institut National de la Santé et de la Recherche Médicale (INSERM), "Maladies génétiques d’expression pédiatrique", Paris, France
    2. Département de Génétique Médicale, Hôpital Armand Trousseau, Assistance Publique-Hôpitaux de Paris, Paris, France

    Contribution

    Conceptualization, Resources, Data curation, Supervision, Funding acquisition, Methodology, Project administration, Writing – review and editing

    Competing interests

    No competing interests declared

14. Achille Aouba

    Département de Médecine Interne et Immunologie Clinique, Normandie Univ, UNICAEN, UR4650 PSIR, CHU de Caen Normandie, Caen, France

    Contribution

    Conceptualization, Resources, Writing – review and editing

    Contributed equally with

    Irina Giurgea

    Competing interests

    No competing interests declared

15. Irina Giurgea

    1. Sorbonne Université, Institut National de la Santé et de la Recherche Médicale (INSERM), "Maladies génétiques d’expression pédiatrique", Paris, France
    2. Département de Génétique Médicale, Hôpital Armand Trousseau, Assistance Publique-Hôpitaux de Paris, Paris, France

    Contribution

    Conceptualization, Resources, Data curation, Supervision, Funding acquisition, Validation, Methodology, Project administration, Writing – review and editing

    Contributed equally with

    Achille Aouba

    For correspondence

    irina.giurgea@inserm.fr

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0002-5035-2958

• 277

  Page views

• 46

  Downloads

• 0

  Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.