Article Text

Download PDFPDF
Concise report
Germline variation of TNFAIP3 in primary Sjögren's syndrome-associated lymphoma
  1. Gaetane Nocturne1,
  2. Jessica Tarn2,
  3. Saida Boudaoud1,
  4. James Locke2,
  5. Corinne Miceli-Richard1,3,
  6. Eric Hachulla4,
  7. Jean-Jacques Dubost5,
  8. Simon Bowman6,
  9. Jacques-Eric Gottenberg7,
  10. Lindsey A Criswell8,
  11. Christopher J Lessard9,
  12. Kathy L Sivils9,
  13. Raphael Carapito10,11,
  14. Siamak Bahram11,
  15. Raphaèle Seror1,
  16. Wan-Fai Ng2,
  17. Xavier Mariette12
  1. 1INSERM UMR1184, CEA-iMETI/Division of Immuno-Virology, Université Paris Sud, Le Kremlin—Bicêtree, France
  2. 2Institute of Cellular Medicine and NIHR Biomedical Research Centre for Ageing and Chronic Diseases, Newcastle, UK
  3. 3Department of Rhumatologie, Hopital Bicetre, Le Kremlin Bicêtre, France
  4. 4Service de Médecine Interne, Lille, France
  5. 5Department of Rhumatologie, Hôpital, Clermont Ferrand, France
  6. 6Department of Rheumatology, University Hospitals Birmingham, Birmingham, UK
  7. 7Department of Rheumatology, University Hospital of Strasbourg, Strasbourg, France
  8. 8Department of Medicine, University of California, San Francisco, San Francisco, California, USA
  9. 9Arthritis and Clinical Immunology Program, Oklahoma Medical Research Foundation, Oklahoma City, Oklahoma, USA
  10. 10U1109, INSERM and University of Strasbourg, Strasbourg, France
  11. 11Immunorhumatologie moléculaire, INSERM UMR S_1109, Centre de Recherche en Immunologie et Hématologie, Faculté de Médecine, Fédération de Médecine Translationnelle de Strasbourg (FMTS), Université de Strasbourg, Strasbourg, France and Service de Rhumatologie, Centre National de Référence pour les Maladies Systémiques Autoimmunes Rares, Hôpitaux Universitaires de Strasbourg, Strasbourg, France
  12. 12Assistance Publique-Hôpitaux de Paris, Hôpitaux Universitaires Paris Sud, Le Kremlin—Bicêtre, France
  1. Correspondence to Dr Xavier Mariette, Assistance Publique-Hôpitaux de Paris, Hôpitaux Universitaires Paris Sud, 78 rue du général Leclerc, Le Kremlin—Bicêtre 94275, France; xavier.mariette{at}


Background and objective A germline and coding polymorphism (rs2230926) of TNFAIP3 (A20), a central gatekeeper of nuclear factor-kappa B (NF-kB) activation, was recently found associated with primary Sjögren's syndrome (pSS)-associated lymphoma in a French cohort. We aimed to replicate this association.

Patients and methods The rs2230926 polymorphism was genotyped in cases and controls of European ancestry from two independent cohorts from UK and France. Case control association tests were performed (Fisher's test) in the two cohorts, followed by a meta-analysis of the two cohorts.

Results The UK cohort included 308 controls and 590 patients with pSS including 31 with a history of lymphoma. The French cohort consisted of 448 controls and 589 patients with pSS including 47 with lymphoma. In both cohorts, the rs2230926 missense polymorphism was not associated with pSS. However, in the UK cohort, the rs2230926G variant was significantly associated with pSS-associated lymphoma (OR=2.74, 95% CI (1.07 to 7.03), p=0.0423, compared with patients with pSS without lymphoma, and OR=3.12, 95% CI (1.16 to 8.41), p=0.0314, compared with healthy controls) as observed in the French cohort. The meta-analysis of the two cohorts confirmed these results (OR=2.48, 95% CI (1.87 to 3.28) p=0.0037 and OR=2.60, 95% CI (1.91 to 3.53) p=0.0031, respectively).

Conclusions This study confirms the role of A20 impairment in pSS-associated lymphoma. Subtle germline abnormalities of genes leading to impaired control of NF-kB activation in B cells continuously stimulated by autoimmunity enhance the risk of lymphoma.

  • Sjøgren's Syndrome
  • B cells
  • Autoimmunity

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.


Primary Sjögren's syndrome (pSS) is the autoimmune disease (AID) with the highest risk of lymphoma with a standardised incidence ratio ranging from 7.1 to 15.57.1 It exemplifies the role of disease activity and especially of chronic antigenic stimulation in the occurrence of B cell malignant proliferation. Several clinical observations support this proposition. First, pSS-associated lymphomas often develop from salivary glands, the target organ of the AID, as mucosa associated lymphoid tissue (MALT) lymphomas. Second, a recent study has demonstrated that high systemic disease activity as assessed by EULAR Sjögren's syndrome disease activity index (ESSDAI) was a predictor of lymphoma.2

Nuclear factor-kappa B (NF-kB) activation is the cornerstone mechanism of inflammation at the cellular level. Thus, mechanisms involved in its control may be critical in autoimmunity and in lymphoma development. We hypothesised that A20, encoded by TNFAIP3, a central gatekeeper of NF-kB activation, might be involved in lymphomagenesis in pSS. At least two lines of evidence support this hypothesis. First, independently of autoimmunity the demonstration of the role of somatic deletion of TNFAIP3 in tumour cells in a number of lymphomas subtypes, and particularly the MALT lymphoma subtype.3–7 Second, the association between polymorphisms located within the TNFAIP3 gene and different AIDs.8 To further test this hypothesis, we have studied germline and somatic abnormalities of TNFAIP3 in pSS-associated lymphomas and have demonstrated that 77% of patients with pSS and MALT lymphoma have germline and/or somatic functional abnormalities of TNFAIP3.9

The role of somatic mutations promoting constitutive activation of NF-kB pathway is well established as a critical pathogenic factor in lymphoma development outside the context of autoimmunity.10 But the novel concept supported by our previous study was that, in the context of continuous stimulation of autoimmune B cells, a subtle germline abnormality of genes controlling NF-kB activation might have specific consequences in B cells. To confirm this innovative hypothesis, we have assessed the association between a functional variant of TNFAIP3, the rs2230926, at the germline level and pSS-associated lymphoma in an independent cohort of European patients with pSS from the UK and performed a meta-analysis with our French cohort.


Study population

This study has included subjects from UK and France. Patients with pSS from UK were included from the UK Primary Sjögren’s Syndrome Registry (UKPSSR). The establishment and recruitment strategy of the UKPSSR have been described previously.11 Two French cohorts of patients with pSS were studied as previously described:9 the prospective ASSESS cohort ‘Assessment of Systemic Signs and Evolution of Sjögren’s Syndrome’ and the cohort of patients with pSS followed in the Department of Rheumatology, Hôpitaux Universitaires Paris-Sud. Twenty additional patients with pSS and lymphoma were studied for extension and replication of results. Controls were selected from a population of French healthy blood donors.

All patients fulfilled the American-European Consensus Group criteria for pSS.12

All subjects were of European ancestry. As previously described, all subjects in the French cohort were also genotyped for 48 ancestry informative markers selected to be highly informative for continental ancestry to identify genetic outliers and adjust for population stratification.9 Controls and patients from the UKPSSR were self-reported Caucasian.


Germline DNA was extracted from peripheral blood mononuclear cells (French cohort) or from whole blood samples collected using PAXgene DNA specimen tubes (UKPSSR).

Genotyping of subjects from the UKPSSR has been performed either with the ImmunoChip method (an Illumina iSelect custom array designed by the ImmunoChip Consortium)13 or with the TaqMan single nucleotide polymorphism (SNP) genotyping assay allelic discrimination method (Life Technologies, Foster City, California, USA). Genotyping of French subjects employed competitive allele-specific PCR system (KASpar genotyping, as previously described.14 Total exon sequencing from germline DNA has been performed in the 20 additional patients with pSS with lymphoma as previously described.

Statistical methods

Case control association tests were performed using Fisher's exact test in the two cohorts (UK and France), and was followed by a meta-analysis using a weighted Z score method where the weights were the sample size.15 Statistical analyses were performed using GraphPad Prism V.5.01 for Windows (GraphPad Software, San Diego, California, USA, To assess the relative role of classical predictors of lymphoma compared with the rs2230926, a multivariate logistic regression analysis with backward selection has been performed in the French cohorts of patients with pSS, SAS V.9.3 statistical software (SAS Institute, Cary, North Carolina, USA).


Patients’ characteristics

The UK cohort included 308 controls and 590 patients with pSS. Among patients with pSS, 31 had a history of lymphoma. The French cohort consists of 448 controls and 589 patients with pSS including 47 with lymphoma. Forty-three of them were previously studied in our first report9 and four are new cases. Among the 78 cases of lymphoma, 73 were non-Hodgkin's B cell lymphoma with marginal zone being the most frequent histological type (61/73, 83.6%). The five other lymphomas were Hodgkin's lymphoma (n=4) and one case of mycosis fungoides.

The rs2230926 SNP is associated with the development of lymphoma in patients with

In the UKPSSR, 22/308 (7.14%) controls harboured rs2230926G compared with 51/590 (8.64%) patients with pSS (table 1) (p=0.52), indicating that rs2230926 was not associated with pSS. However, the rs2230926G variant was significantly associated with the development of lymphoma in patients with pSS (OR=2.74 95% CI (1.07 to 7.03), p=0.0423, compared with patients with pSS without lymphoma, and OR=3.12 95% CI (1.16 to 8.41), p=0.0314, compared with healthy controls, table 1). The results of the extended French cohort were similar to those published previously9 (table 1). The meta-analysis of the two cohorts included 756 controls, 1179 patients with pSS and 78 cases of lymphoma (table 1). rs2230926G was present in 76/756 (10.05%) controls, 127/1179 (10.77%) patients with pSS and 17/78 (21.79%) patients with pSS associated lymphoma. Altogether, these data confirm the specific association between the rs2230926 exonic variant and the occurrence of lymphoma in patients with pSS (OR=2.48 95% CI (1.07 to 7.03), p=0.0037, compared with patients with pSS without lymphoma, and OR=2.60 95% CI (1.91 to 3.53), p=0.0031, compared with healthy controls).

Table 1

Association testing of rs2230926 in the two cohorts and meta-analysis

To assess if the exonic variant was associated with lymphoma independently of classical predictors,1 we performed a multivariate logistic regression analysis on the French cohorts of patients with pSS including 476 patients with pSS without lymphoma compared with 47 cases of lymphoma. In 16 of them, the lymphoma has been diagnosed before recruitment to the study and data on clinical predictors before lymphoma were not available. We therefore focused on the classical biological predictors that are usually stable over time (cryoglobulinaemia, C4 level, monoclonal component) in our regression analysis. Using this approach, we found that the two factors significantly and independently associated with lymphoma were low C4 (OR=5.83 95% CI (2.91 to 11.66), p<0.0001) and rs2230926G (OR=2.28 95% CI (1.11 to 6), p=0.028). Last, we run a second multivariate logistic regression analysis also including classical clinical predictors of lymphoma, especially parotid swelling. Only 31/78 patients had all biological and clinical data available before lymphoma. In this analysis the only factors associated with lymphoma occurrence were history of parotid swelling (OR=9.57, 95% CI (3.77 to 24.31)), low C4 (OR=3.09, 95% CI (1.20 to 7.97)) and presence of cryoglobulinaemia (OR=3.92, 95% CI (1.40 to 10.99)).


Among two large cohorts of European subjects, we studied the association between a germline coding variant (rs2230926) of TNFAIP3 and the development of lymphoma in patients with pSS. We confirmed that this germline missense variant was not associated with pSS in general but with a specific complication of the disease: pSS-associated lymphoma.

The rs2230926 minor (G) allele was more common in French controls compared with UK controls. The same trend was seen among the French patients with pSS compared with the UK patients with pSS, including the subset of patients with lymphoma. However, this exonic variant was associated with pSS-associated lymphoma in both cohorts.

Work by our group and others demonstrates that the rs2230926G allele results in impaired control of the NF-kB activation pathway.9 ,16 Further, complete absence of A20 in mice results in death during the neonatal period due to severe, uncontrolled inflammation leading to cachexia.17 Reasons for the presence of the rs2230926 TNFAIP3 variant in 6–12% of healthy Caucasians remains unresolved. It is possible that the modestly increased level of inflammation induced by this variant might be a selective advantage in certain situations, such as infection; whereas in other contexts, this variant could have deleterious effects, such as an increased risk for AID or lymphoma.

pSS is the prototype of the B cell mediated AID.18 The chronic stimulation by immune complexes at the site of the disease (mucosa salivary and lacrimal gland tissue) of polyclonal B cells with rheumatoid factor (RF) activity could be a key event involved in lymphomagenesis occurring in patients with pSS. Thus, autoreactive RF+B cells are continuously activated in mucosa by autoantigens which favour their proliferation in a clonal manner. A perfect control of NF-kB activation may be necessary to prevent such B cell clonal proliferation, which may explain why relatively subtle abnormalities of A20 of germline origin may have considerable biological consequences in B cells. Relative to other genes, the role of TNFAIP3 (A20) in controlling excess activation of NF-kB could be more important in the epithelium. Indeed, cre-lox mice depleted in A20 in skin and gut develop psoriasis and inflammatory bowel disease, respectively.19 Further, a germline gain of function mutation of CARD11, which is involved in the NF-kB process with consequences only in B cells, has also recently been associated with familial B cell lymphomas.20

A multivariate analysis including only stable biological predictors of lymphoma supports an independent role of the TNFAIP3 variant in the development of pSS-associated lymphoma. However, when including classical clinical predictors also, with the limitation of the lack of power, this variant was not independently associated with lymphoma anymore. Finally, our results do not suggest that this polymorphism should be looked at in all patients with pSS, but it gives some clues to the mechanism of lymphomagenesis in this AID.

In conclusion, this study confirms the role of A20 impairment in pSS-associated lymphoma. Subtle germline abnormalities of TNFAIP3 (A20) leading to impaired control of NF-kB activation in B cells continuously stimulated by autoimmunity enhance the risk of lymphoma.


The UKPSSR also received infrastructure support from the Newcastle NIHR Biomedical Research Centre, Newcastle Clinical Research Facility and the NIHR Comprehensive Local Research Network to the 30 recruiting centres. The authors would like to thank all the patients and healthy volunteers who have participated in the UKPSSR. W-FN and SB are investigators of the UKPSSR. The other UKPSSR members (as of 1 July 2011) include, in alphabetical order of their affiliations: Elalaine C Bacabac, Robert Moots (Aintree University Hospitals); Kuntal Chadravarty, Shamin Lamabadusuriya (Barking, Havering and Redbridge NHS Trust); Michele Bombardieri, Constantino Pitzalis, Nurhan Sutcliffe (Bart and the London NHS Trust); Nagui Gendi, Rashidat Adeniba (Basildon Hospital); John Hamburger, Andrea Richards (Birmingham Dental Hospital); Saaeha Rauz (Birmingham & Midland Eye Centre); Sue Brailsford (Birmingham University Hospital); Joanne Logan, Diarmuid Mulherin (Cannock Chase Hospital); Jacqueline Andrews, Paul Emery, Alison McManus, Colin Pease (Chapel Allerton Hospital, Leeds); Alison Booth, Marian Regan (Derbyshire Royal Infi rmary); Theodoros Dimitroulas, Lucy Kadiki, Daljit Kaur, George Kitas (Dudley Group of Hospitals NHS Foundation Trust); Mark Lloyd, Lisa Moore (Frimley Park Hospital); Esther Gordon, Cathy Lawson (Harrogate District Foundation Trust Hospital); Monica Gupta, John Hunter, Lesley Stirton (Gartnavel General Hospital, Glasgow); Gill Ortiz, Elizabeth Price (Great Western Hospital); Gavin Clunie, Ginny Rose, Sue Cuckow (Ipswich Hospital NHS Trust); Susan Knight, Deborah Symmons, Beverley Jones (Macclesfi eld District General Hospital & Arthritis Research UK Epidemiology Unit, Manchester); Andrew Carr, Suzanne Edgar, Marco Carrozzo, Francisco Figuereido, Heather Foggo, Colin Gillespie, Dennis Lendrem, Iain Macleod, Sheryl Mitchell, Jessica Tarn, Bridget Griffiths (Newcastle upon Tyne Hospitals NHS Foundation Trust and Newcastle University); Adrian Jones, Peter Lanyon, Alice Muir (Nottingham University Hospital); Paula White, Steven Young-Min (Portsmouth Hospitals NHS Trust); Susan Pugmire, Saravanan Vadivelu (Queen's Elizabeth Hospital, Gateshead); Annie Cooper, Marianne Watkins (Royal Hampshire County Hospital); Anne Field, Stephen Kaye, Devesh Mewar, Patricia Medcalf, Pamela Tomlinson, Debbie Whiteside (Royal Liverpool University Hospital); Neil McHugh, John Pauling, Julie James, Nike Olaitan (Royal National Hospital for Rheumatic Diseases); Mohammed Akil, Jayne McDermott, Olivia Godia (Royal Sheffi eld Hospital); David Coady, Elizabeth Kidd, Lynne Palmer (Royal Sunderland Hospital); Bhaskar Dasgupta, Victoria Katsande, Pamela Long (Southend University Hospital); Usha Chandra, Kirsten MacKay (Torbay Hospital); Stefano Fedele, Ada Ferenkeh-Koroma, Ian Giles, David Isenberg, Helena Marconnell, Stephen Porter (University College Hospital & Eastman Dental Institute); Paul Allcoat, John McLaren (Whyteman's Brae Hospital, Kirkaldy). The authors thank the following investigators of the ASSESS cohort (all in France) who recruited the patients and conducted follow-up: A L Fauchais (Limoges), S Rist (Orleans), D Sené (Paris), V Le Guern (Paris), G Hayem (Paris), J Sibilia (Strasbourg), J Morel (Montpellier), A Saraux (Brest), A Perdriger (Rennes), X Puechal (Le Mans) and V Goeb (Rouen). The authors thank Dr J Benessiano and all staff members of the Bichat Hospital Biological Resource Center (Paris) for their help in centralising and managing biological data collection from the French ASSESS (Atteinte Systémique et Evolution des patients atteints de Syndrome de Sjögren primitive) cohort, a prospective cohort of patients with Sjögren's Syndrome.



  • Handling editor Tore K Kvien

  • Contributors XM and W-FN have supervised the project; GN, SB, RC, JL have performed experiments; GN, RS, CJL, KLS and LAC have performed the genetic study and statistical analyses; XM, GN, JT,CM-R, SB, SB, J-JD, EH, RS, JEG and W-FN have participated in the writing of the manuscript; SB, J-JD, EH and J-EG have contributed to patients’ recruitment.

  • Funding Grants were obtained from French ministry of health: PHRC N°2006-AOM06133, from French ministry of research: ANR-2010-BLAN-1133 01, from the Medical Research Council, UK (G0800629) and from the NIH P50 AR060804 (KLS and CJL) and U19 AI082714 (KLS and CJL).

  • Competing interests None declared.

  • Patient consent Obtained.

  • Ethics approval Informed consent was obtained from all subjects, and the study was approved by the local research ethics committees.

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