Article Text

Type I interferon signature predicts response to JAK inhibition in haploinsufficiency of A20
  1. Daniella Muallem Schwartz1,
  2. Sarah A Blackstone1,
  3. Natalia Sampaio-Moura2,
  4. Sofia Rosenzweig2,
  5. Aarohan M Burma1,
  6. Deborah Stone2,
  7. Patrycja Hoffmann2,
  8. Anne Jones2,
  9. Tina Romeo2,
  10. Karyl S Barron1,
  11. Meryl A Waldman3,
  12. Ivona Aksentijevich2,
  13. Daniel L Kastner2,
  14. Joshua D Milner1,
  15. Amanda K Ombrello2
  1. 1 NIAID, National Institutes of Health, Bethesda, Maryland, USA
  2. 2 NHGRI, National Institutes of Health, Bethesda, Maryland, USA
  3. 3 NIDDK, National Institutes of Health, Bethesda, Maryland, USA
  1. Correspondence to Dr Daniella Muallem Schwartz, NIAID, National Institutes of Health, Bethesda, Maryland, USA; Daniella.Schwartz{at}

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

The anti-inflammatory protein A20, encoded by TNFAIP3, is a ubiquitin-modifying enzyme that targets proinflammatory molecules, including those upstream of the transcription factor nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB). Patients with heterozygous loss-of-function TNFAIP3 mutations develop haploinsufficiency of A20 (HA20), a systemic autoinflammatory disease that can cause severe end-organ pathology.1–3 No available medication directly targets NF-κB signalling; thus, treatment decisions are based on clinical experience. Mild cases are treated with disease-modifying antirheumatic drugs, whereas severe cases are treated with systemic corticosteroids and biological agents, including tumour necrosis factor (TNF)-α and IL-1 receptor (IL-1R) blockade.1 2 We report that a type I interferon (IFN) signature, or elevation of IFN-stimulated genes (ISGs), correlates with disease activity and predicts response to janus kinase (JAK) inhibition in HA20.

A cohort of 12 patients with HA20 is followed at the NIH Clinical Center. All patients were diagnosed by Sanger sequencing, and increased NF-κB activity was confirmed with luciferase assay.2 Five patients had disease that was treatment-refractory or caused end-organ pathology. P1 was a 15-year-old female with p.T604Rfs*93 and severe gastrointestinal ulcerations refractory to TNF-α and IL-1R inhibition. P2–P4 were members of the same extended family with p.F224Sfs*4, aged 28, 32 and 61 years. All three had incomplete responses to TNF-α and IL-1R blockade; P2 had retinal vasculopathy and neuroinflammation, whereas P3 had membranous nephropathy. P5 was an 8-year-old girl with an unreported p.L626Vfs*45 and autoinflammatory liver disease. Notably, T604Rfs*93 and p.L626Vfs*45 are in the fourth zinc finger domain (ZnF4), which mediates several critical functions of A20.4 Autoinflammatory disease activity index scores ranged from 51 to 117 for P1–P4; P5 was asymptomatic but had severe hepatic inflammation with fibrosis (figure 1A–B).

Figure 1

(A) Autoinflammatory disease activity index (AIDAI) scores for five HA20 patients with severe active disease. P1–P4 had AIDAI scores between 51 and 117. P5 was asymptomatic, with an AIDAI score of 0. (B) Organ pathology in five HA20 patients with active disease. Disease manifestations were heterogeneous and included orogenital ulcerations (P1–P4), neuroinflammatory disease (P2), membranous nephropathy (P3, H&E) and autoinflammatory hepatitis with fibrosis (P5, Masson trichrome). (C) Expression of IFN-stimulated genes (ISGs) in HA20 patients and healthy volunteers. Expression is shown for five HA20 patients with active disease (P1–P5), three HA20 patients with quiescent disease (P6–P8) and four healthy volunteers (HC1–HC4). The dendogram shows unbiased hierarchical clustering of the 12 subjects. (D,E) Clinical response of HA20 patients to tofacitinib treatment. (D) Values assessed before and after treatment initiation are shown for AIDAI scores (P1–P3), mean erythrocyte sedimentation rate (ESR) and mean C reactive protein (CRP) (P1–P3, P5). (E) Mean aspartate aminotransferase (AST) and alanine aminotransferase (ALT) levels are shown before and throughout treatment with tofacitinib (P5). (F) Immunological response of HA20 patients to tofacitinib treatment. Geometric mean expression of interferon-stimulated genes are shown in healthy volunteers, patients with primary interferonopathies, HA20 before treatment (P1–P5), and HA20 after treatment (P1–P5). *P<0.05, **p<0.01. Mann-Whitney analysis (for unpaired analysis of healthy volunteers vs HA20), paired t-test (for paired analysis before and after treatment with tofacitinib). HA20, haploinsufficiency of A20.

NF-κB and other A20-regulated signalling molecules can induce type I IFNs,2 leading us to hypothesise that treatment-refractory HA20 might be characterised by increased ISG expression. Accordingly, we measured expression in the whole blood of all five patients and found elevated ISGs compared with healthy volunteers (figure 1C). As comparators, we also investigated three patients with quiescent HA20; ISGs were not elevated in these subjects (figure 1C). To determine whether NF-κB activation indirectly induced ISGs in HA20 patients, we stimulated healthy and HA20 peripheral blood mononuclear cells in vitro with the NF-κB activating cytokines TNF-α and IL-1β. We were unable to detect IFNA or IFN-α in stimulated or unstimulated cells (data not shown). ISG expression decreased over time in unstimulated HA20 cells; for some ISGs, stimulation prevented this reduction. However, neither TNF-α nor IL-1β significantly induced ISGs in healthy or HA20 cells (online supplementary figure S1A–B). This suggests that A20-mediated regulation of NF-κB may induce some ISGs, but that other mechanisms also promote ISG expression in HA20.

Patients with mutations that enhance IFN signalling have been successfully treated with JAK inhibitors, which target the signalling molecules downstream of type I IFNs.5 This led us to hypothesise that JAK inhibition would be an effective therapeutic strategy for treatment-refractory HA20. Under the IRB-approved protocol 94-HG-0105, we initiated treatment with tofacitinib monotherapy 2.5 mg two times per day for P5, and 5 mg two times per day for P1, P3 and P4. P2 declined tofacitinib. At the time of analysis, treatment duration ranged from 5 to 24 months. Clinical and immunological responses were seen in all four patients (figure 1D–F). Proteinuria in P3 improved from 3+ to 1+ on dipstick analysis, and hepatic transaminases in P5 decreased progressively (figure 1E). Tofacitinib was well-tolerated in all four patients. No opportunistic or severe infections were reported during treatment. Mean haematological parameters and lipid levels remained stable, and there were no cardiovascular or thrombotic events (online supplementary figure S1C–D).

This is the first report that a type I IFN signature correlates with active disease and predicts clinical response in HA20, an autoinflammatory disease that is not a primary interferonopathy. We also expand the spectrum of HA20-associated phenotypes to include severe hepatic inflammation in the absence of systemic features. Together with a recent report of HA20 treatment with the JAK1/2 inhibitor baricitinib, we provide compelling evidence that JAK inhibition is safe and effective for HA20.6 A20 directly targets NF-κB signalling and the NLRP3 inflammasome; although A20 is also described to regulate IFN signalling, the underlying mechanisms are incompletely characterised.4 7 Our data suggest that disease activity correlates with elevations of multiple proinflammatory cytokines and increased ISG expression.2 Future studies will be needed to identify the targets of A20 in various inflammatory cells, and the mechanisms through which A20 constrains type I IFN responses.


We thank the patients and their families for participating in the study. We thank G Ben-Yakov for assistance in managing the patients and D Kleiner for high-resolution histology images.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.


  • Handling editor Josef S Smolen

  • Contributors Conceptualisation: DMS and AKO. Methodology: DMS and AKO. Investigation: DMS, SB, NS-M, AB, SR, DS, PH, TR, AJ, IA and KSB. Formal analysis: DMS and SB. Writing: DMS, SB and AKO. Supervision: DLK, JDM and AKO.

  • Funding This study was supported by the Intramural Research Programs of the National Human Genome Research Institute, the National Institute for Allergy and Infectious Diseases, the National Institute for Arthritis and Musculoskeletal and Skin Diseases, and the NIH Transitional Program for Clinical Research.

  • Competing interests None declared.

  • Patient consent for publication Not required.

  • Ethics approval NIH Institutional Review Board, 94-HG-0105, IRB00000006.

  • Provenance and peer review Not commissioned; externally peer reviewed.