Autoantibody Profile in Eosinophilic Granulomatosis and Polyangiitis: Predominance of Anti-Alpha-Enolase Antibodies K

Home , Autoantibody

Autoantibody profile in eosinophilic granulomatosis and polyangiitis: predominance of anti-alpha-enolase antibodies K. Laskari1, B. Hellmich2, G. Adamus3, E. Csernok2

1Rheumatology Unit, 1st Department ABSTRACT Introduction of Propaedeutic Internal Medicine, Objective. To evaluate the autoanti- Eosinophilic granulomatosis and poly- Joint Academic Rheumatology Program, body profile in eosinophilic granuloma- angiitis (EGPA) is a rare systemic ne- Medical School, University of Athens, tosis and polyangiitis (EGPA) patients. crotising pauci-immune vasculitis of Greece; 2Department of Internal Medicine, Methods. 33 EGPA patients were unknown aetiology associated with Rheumatology and Immunology, tested for anti-neutrophil cytoplas- peripheral and tissue eosinophilia, Vasculitis-Centre Tübingen-Kirchheim, mic antibodies (ANCA), antinuclear asthma and/or allergic rhinitis affecting Medius Klinik Kirchheim, University antibodies (ANA), rheumatoid factor small-sized vessels (1). Anti-neutro- of Tübingen, Kirchheim-Teck, Germany; (RF), anti-alpha-enolase antibodies, phil cytoplasmic antibodies (ANCA) 3 Ocular Immunology Laboratory, and anti-eosinophil peroxidase (EPO) are the most common autoantibodies Casey Eye Institute, School of Medicine, antibodies. Granulomatosis with poly- detected in about 30–48% of patients Oregon Health & Science University, Portland, OR, USA. angiitis (GPA), microscopic polyangii- (2-5). In this study, we evaluated the tis (MPA), hypereosinophilic syndrome autoimmune serological profile of Katerina Laskari, MD Bernhard Hellmich, MD (HES), rheumatoid arthritis (RA), pri- EGPA patients and searched for novel Grazyna Adamus, PhD, FARVO mary biliary cirrhosis (PBC) patients autoantibody specificities. Eosinophil Elena Csernok, PhD and healthy subjects were tested as a peroxidase (EPO) of the eosinophil Please address correspondence to: control group. granulocytes as well as the glycolytic Elena Csernok, Results. Anti-alpha-enolase antibodies cytoplasmic enzyme alpha-enolase Department of Internal Medicine, were positive in 82% of EGPA patients have been identified as autoantigens in Rheumatology and Immunology, at high titres. Although a high sensitiv- patients with autoimmune disorders (6, Vasculitis-Centre Tübingen-Kirchheim, ity was shown for an anti-alpha-eno- 7). Besides, previous studies report on Medius Klinik Kirchheim, lase antibody titre above 1/100 (82%), University of Tübingen, ANCA directed either against EPO or Eugenstrasse 3, the specificity for EGPA remained low alpha-enolase in systemic vasculitis (8, 73230 Kirchheim unter Teck, Germany. (44%) (AUC=0.653, p=0.008). Anti- 9). Apparently, the anti-alpha-enolase E-mail: [email protected] alpha-enolase antibodies predominat- response does not appear to be restrict- Received on May 14, 2020; accepted in ed in males with EGPA (p=0.048) and ed to a specific disease, but the fine revised form on August 31, 2020. were associated with skin involvement epitopes have not been examined in the Clin Exp Rheumatol 2021; 39 (Suppl. 129): (p=0.040). Most of the EGPA patients majority of those diseases and could S83-S87. positive for anti-alpha enolase anti- be important in determining the im- © Copyright Clinical and bodies (20 out of 27) had a negative munopathogenicity (10). We speculate Experimental Rheumatology 2021. indirect immunofluorescence test (IFT) that these antibodies may also initiate for ANCA. ANCA were positive in 8 inflammation leading to severe necro- Key words: eosinophilic EGPA patients (24%) with a perinu- sis of the endothelium (11). Since both granulomatosis and polyangiitis clear pattern in all but one patient. The eosinophils and ANCA play a pathoge- (EGPA), anti-alpha-enolase ANCA-target antigen was myeloper- netic role in EGPA (12), our aim was antibodies, anti-neutrophil oxidase (MPO) and/or alpha-enolase. to investigate if anti-alpha-enolase and cytoplasmic antibodies (ANCA) A usually fine-speckled ANA pattern anti-EPO antibodies are present in the was observed in 42% of the EGPA pa- sera of EGPA patients and estimate tients. RF was positive in 1 (6%) of the their diagnostic value. 18 EGPA patients tested. There was no association between the presence and Patients and methods levels of autoantibodies and EGPA This study was performed in a cohort disease activity. None of the patients of patients with EGPA. Sera from 37 and controls was positive for anti-EPO MPA, 23 GPA and 50 RA patients antibodies. matched for age, sex and disease du- Conclusion. Alpha-enolase may be ration with the main group were tested a target of autoimmunity in EGPA for ANA, ANCA and/or anti-alpha- patients and shows usually negative enolase antibodies as a control group. Competing interests: none declared. ANCA IFT results. As control group for anti-EPO anti-

Clinical and Experimental Rheumatology 2021 S-83 Autoantibodies in EGPA / K. Laskari et al. bodies, sera from 12 GPA, 4 MPA, 8 Table I. Clinical characteristics of the EGPA patients. Values are expressed either as abso- HES, 5 PBC patients and 100 healthy lute numbers or as proportions (%) or as median (min-max). subjects were used. Gender 18 M:15 F Serum samples were collected at a ter- Age at serum testing [years] 47 yrs. (23-78) tiary referral centre for vasculitis in Age at diagnosis [years] 44 yrs. (17-75) Germany (Klinikum Bad Bramstedt). Disease duration [years] 2 yrs. (1-21) All patients fulfilling the American Eosinophil count at serum testing [/nl] 212 (4-5120) Active disease, patient no. (%) 24 (73%) College of Rheumatology classifica- Complete remission, patient no. (%) 9 (27%) tion criteria for EGPA (13), followed- BVAS score at serum testing 6 (0-25) up for at least one year, were included Manifestations during f-u, patient no. (%) in the study. Patients with acute or - General symptoms 24 (73%) chronic infections and/or a history of - ENT/eye involvement 28 (85%)/7 (21%) - Lung involvement 29 (88%) malignancy were excluded. EGPA pa- - Kidney involvement 5 (15%) tients were tested for the presence of - Arthralgias/arthritis 15 (45%) ANCA, anti-alpha-enolase antibodies, - Skin involvement 19 (58%) ANA, RF and anti-EPO antibodies. - Heart involvement 16 (48%) The design of the work was in accord- - PNS/CNS involvement 21 (64%)/5 (15%) - GI involvement 13 (39%) ance with the standards of the Helsinki - Thrombosis 2 (6%) Declaration of 1975/83. Prednisolone [mg] at serum testing 8 (0-28) Treatment at serum testing, patient number Autoantibody detection - CYC 7 - ANCA - IFN 6 Indirect IFT was performed on etha- - MTX/AZA/LEF 14/2/1 nol-fixed neutrophils in combination - None 3 with additional tests on formalin-fixed Follow up time [years] 6 yrs. (1-19) neutrophils and HEp-2 cells to better ENT: ear nose throat; PNS: peripheral nervous system; CNS: central nervous system; GI: gastrointes- discriminate between P-ANCA/ atyp- tinal; CYC: cyclophosphamide; IFN: interferon; MTX: methotrexate; AZA: azathioprine; LEF: leflu- ical-ANCA and ANA, as previously nomide. described (14). For the detection of proteinase 3 (PR3)-ANCA, a third gen- and L chain) conjugated to peroxidase Statistical analysis eration anti-PR3 hs Enzyme linked Im- (Zymed, San Francisco, CA). As posi- Statistical analysis was performed us- munosorbent Assay (ELISA) and first tive control, we used a reference hu- ing the SPSS program. The Mann- generation anti-MPO ELISA (Orgen- man serum containing anti-recoverin Whitney U and the χ2 tests were used tec, Mainz, Germany) were used (15). antibodies diluted 1:100. As negative as appropriate. Correlations between To identify the ANCA target antigens, control, we omitted serum and applied continuous variables were determined lactoferrin-, human leucocyte elastase-, only a secondary antibody. The results by the Pearson’s correlation coeffi- bacterial permeability increasing pro- are presented as anti-alpha enolase an- cient. A p-value ≤0.05 was considered tein-, and cathepsin-G-ANCA antigen tibody titres. significant. Receiver operating charac- specific kits from Orgentec were used. Sera of 12 EGPA patients were also teristics (ROC) analysis was applied in tested by Western blotting (16). Retinal order to define specificity and sensitiv- - Anti-EPO Autoantibodies _lpha enolase was separated by SDS- ity of anti-alpha-enolase antibodies. In In house ELISA was used. Purified gel electrophoresis on a 10% gel and order to assess diagnostic performance, EPO (Calbiochem) was coated to micr- transferred to an Immobilon membrane the area under the ROC curve (AUC) otitre plates at a protein concentration (Millipore, Bedford, Massachusetts). was calculated. of 1 microgram/ml. EGPA patients and Individual strips containing alpha eno- control sera were used in a 1:50 dilu- lase were incubated with 1:100 diluted Results tion. serum followed by 1:5000 diluted anti- A total of 33 EGPA patients were evalu- human IgG (H and L chain) conjugated ated. The clinical and laboratory patient - Anti-alpha enolase autoantibodies to alkaline phosphatase (Sigma, St. characteristics and treatment details are Anti-alpha enolase antibodies were Louis, MO). shown in Table I. Twenty-four patients detected using ELISA as previous- had active disease (Birmingham Vascu- ly described (16). Falcon microtitre - Other autoantibodies litis Activity Score, BVAS≥1) (17) and plates were coated with 500 ng/well Indirect IFT on HEp-2 cells (Inova) 9 patients had inactive disease (BVAS: of purified bovine retinal α-enolase. and ELISAs (Varelisa ANA profile 0 and prednisolone dose ≤7.5 mg/day). Two-fold serially diluted patients’ se- Enzyme-Immunoassay) were used to Anti-alpha-enolase antibodies were de- rum was added to each well followed detect ANA and their main target an- tected at high titres in 82% of EGPA by incubation with anti-human IgG (H tigens. patients. The results were confirmed

S-84 Clinical and Experimental Rheumatology 2021 Autoantibodies in EGPA / K. Laskari et al.

Table II. Frequency and titre of autoantibodies (Abs) in EGPA and controls.

Abs EGPA MPA GPA RA n=33 n=37 n=23 n=50 (active/inactive n=24/9) (active/inactive n=23/14) (active/inactive n=21/2)

C-/P-ANCA (%) 8 (24) 23 (62) 18 (78) 2 (4) Titre [Median (Min-Max)] 1/128 (1/128-1024) 1/256 (1/128-2048) 1/512 (1/128-4096) 1/1024 both P-ANCA [target antigen] 7 [2 MPO and a-enolase, 23 [6 MPO, 12 MPO and 3 [2 MPO, 1 PR3 and 2 [1 LF, 1 a-enolase] 5 a-enolase] a-enolase, 1 PR3 and a-enolase] a-enolase, 4 a-enolase] C-ANCA [target antigen] 1 [MPO] - 15 [11 PR3, 3 PR3 and - a-enolase, 1 unknown]

Anti-alpha-enolase Abs* (%) 2‡ (82) 27† (73) 5** (22) 30 (60) Titre [Median (Min-Max)] 1/1600 (1/100-6400) 1/3200 (1/200-25600) 1/400 (1/100-12800) 1/200 (1/200-6400) ≤1/200 6 4 1 16 1/400-800 8 6 3 10 1/1600-3200 2 8 - 3 ≥1/6400 11 9 1 2

ANA (%) 15 (45) 17 (46) 7 (30) 32 (64) Titre [Median (Min-Max)] 128 (64-1024) 128 (64-2048) 128 (64-640) 256 (64-2048)

RF (%) 1/18 (6) nd nd 35 (70)

Anti-EPO Abs 0 0 0 0

MPO: myeloperoxidase; PR-3: proteinase 3; LF: lactoferrin; nd: not done. ANCA (IFT) positive ‡ 7/27; † 17/27; **4/5. *10% of healthy subjects were positive for anti-alpha enolase antibodies. Statistically significant comparisons: ANA frequency in RA vs. EGPA (p=0.04), RA vs. GPA (p=0.02); ANA titre in RA vs. EGPA (p=0.025), RA vs. GPA (p=0.005). ANCA frequency in GPA vs. EGPA (p=0.01), MPA vs. EGPA (p=0.001), EGPA vs. RA (p=0.026); ANCA titre in MPA vs. RA, GPA vs. RA (both p<0.001). Anti-alpha-enolase frequency in EGPA vs. GPA (p<0.001), EGPA vs. RA (p=0.033), MPA vs. GPA (p<0.001), RA vs. GPA (p=0.002). No significant difference between EGPA and MPA (p=0.403); Anti-alpha-enolase Ab titre in EGPA vs. RA (p=0.007), MPA vs. RA (p<0.001). No significant difference between EGPA and MPA (p=0.264) or GPA (p=0.509). in all 12 sera tested by Western blot- negative in IFT ANCA (20/27) in con- sociated with disease activity or other ting (6 positive and 6 negative sera trast to MPA (10/27) and GPA (1/5). clinical characteristics (p>0.05). EGPA for anti-alpha-enolase antibodies in ANCA were detected by IFT in 24% patients were seronegative for ENA. ELISA). Anti-alpha-enolase antibodies of EGPA patients showing a perinu- RF was found positive in 1/18 (6%) of were also found in control patients, but clear pattern in the majority of cases. EGPA patients tested. Also, all patients their frequency was higher in EGPA as The ANCA target antigens were MPO and controls were negative for anti- compared to GPA and RA patients and and/or alpha-enolase (Table II). ANCA EPO antibodies. their titre higher in EGPA than RA pa- were more common in EGPA as com- tients (Table II, Fig. 1A). ROC analysis pared to RA and less common in EGPA Discussion showed a high sensitivity for autoanti- than GPA and MPA patients, but with EGPA has been classified alongside body titres above 1/100 (82%), howev- comparable titres (p>0.05) (Table II). MPA and GPA as an ANCA-associated er, the specificity for EGPA remained There was no association between vasculitis (AAV), despite ANCA be- low (44%) (AUC=0.653, p=0.008). the ANCA presence (p=0.653) or an- ing found in the minority of cases. The There was no association between tibody levels (p=0.543) and disease frequency of ANCA was observed in the presence and levels of anti-alpha- activity and other EGPA patient char- only 24% of our EGPA patients with enolase antibodies and disease activity acteristics (p>0.05). When EGPA and the most common perinuclear pattern, (p>0.40). Anti-alpha-enolase antibod- MPA patients were evaluated together, which is in an agreement with previ- ies were detected more often in males MPO-specificity in the ANCA-ELISA ous studies (2-5). Lyons et al. suggest (p=0.048) and were associated with was associated with lower frequency that EGPA might be comprised of two skin involvement (all 19 EGPA patients (p=0.047) and titre (p=0.09) of anti- distinct diseases defined by ANCA sta- with infarcts, purpura, ulcers and/or alpha-enolase antibodies (Fig. 1B). tus, since they had revealed genetic and gangrene were positive for anti-alpha- ANA, with a usually fine-speckled pat- clinical differences between the EGPA enolase antibodies) (p=0.040). Most tern, occurred at lower titres and less MPO-positive subset and the larger of the EGPA patients were positive for frequently in EGPA (45%) than RA ANCA-negative subset (18). This dis- anti-alpha enolase antibodies but were patients (Table II). They were not as- tinction may suggest an important het-

Clinical and Experimental Rheumatology 2021 S-85 Autoantibodies in EGPA / K. Laskari et al.

erogeneity in the pathogenesis of the clinical syndrome. The present data support an idea that alpha-enolase can be considered as potential target autoantigen for ANCA in EGPA patients. Anti-alpha-enolase antibodies have been previously reported in AAV at a lower frequency as compared to the present data (40% vs. 64%) (9). Our study highlights a high frequency and titre of anti-alpha-enolase antibodies, particularly in EGPA patients (82%). In spite of the high sensitivity, we failed to demonstrate a high specificity of these autoantibodies for EGPA. However, further research on the identification of epitopes for alpha-enolase antibodies could reveal a subset of autoantibodies specific for the disease. Neither the presence nor the titre of anti-alpha enolase antibodies was associated with disease activity in both EGPA patients and controls. Alpha-enolase has been identified as an ANCA autoantigen in the past (9) and has been shown on the surface of neutrophils (11). It is of note that, although MPO-specificity is most commonly detected in ELISA, in some EGPA cases the ANCA target antigen remains unknown (4, 5). The present study supports alpha-enolase to be another autoantigen in cases with unknown ANCA specificity and sometimes- pa tients with negative ANCA IFT. Among our AAV patients, 61% of those positive for anti-alpha-enolase antibodies were also positive for ANCA (IFT). In particular, among EGPA patients, only 26% of those positive for anti-alpha- enolase antibodies had a positive IFT for ANCA. It is of interest that patients without ANCA or without MPO-specificity in ANCA-ELISA had more frequently and at higher titres anti-alpha enolase antibodies. Alpha-enolase is not exclusively expressed in neutrophils but also in other cells (11). Indeed, the detection of alpha-enolase in keratinocytes (19) supports the hypothesis that anti-alpha- enolase antibodies may be a causative factor for the development of skin le- Fig. 1. A: Box plot of the antibody titre in EGPA patients and controls positive for anti-alpha enolase antibodies. B: Box plot of the anti-alpha enolase antibody titre in EGPA/MPA patients without ANCA sions in autoimmune patients. In fact, and patients with ANCA with and without MPO specificity. in the present study all EGPA patients The median value in each category is shown by the horizontal line inside the box. with skin lesions were positive for anti- MPO: myeloperoxidase; Abs: antibodies. alpha-enolase antibodies. Skin involve-

S-86 Clinical and Experimental Rheumatology 2021 Autoantibodies in EGPA / K. Laskari et al. ment is a common complication occur- References 12. NGUYEN Y, GUILLEVIN L: Eosinophilic ring in 40%-70% of EGPA patients 1. MONTI S, BOND M, FELICETTI M et al.: One Granulomatosis with Polyangiitis (Churg- year in review 2019: vasculitis. Clin Exp Strauss). Semin Respir Crit Care Med 2018; (12). The histopathologic spectrum of Rheumatol 2019; 37 (Suppl. 117): S3-19. 39: 471-81. dermal small-vessel vasculitis ranges 2. GUILLEVIN L, COHEN P, GAYRAUD M et al.: 13. MASI AT, HUNDER GG, LIE JT et al.: The from eosinophilic vasculitis with nega- Churg-Strauss syndrome. Clinical study and American College of Rheumatology 1990 tive MPO-ANCA to neutrophilic vas- long-term follow-up of 96 patients. Medicine criteria for the classification of Churg- (Baltimore) 1999; 78: 26-37. Strauss syndrome (allergic granulomatosis culitis with positive MPO-ANCA (20). 3. SINICO RA, DI TOMA L, MAGGIORE U et al.: and angiitis). Arthritis Rheum 1990; 33: Thus, in patients with cutaneous vascu- Prevalence and clinical significance of anti- 1094-100. litic lesions without ANCA, the detec- neutrophil cytoplasmic antibodies in Churg- 14. CSERNOK E, MOOSIG F: Current and emerg- tion of anti-alpha-enolase antibodies Strauss syndrome. Arthritis Rheum 2005; 52: ing techniques for ANCA detection in vascu- 2926-35. litis. Nat Rev Rheumatol 2014; 10: 494-501. may help in the diagnosis of EGPA. 4. COMARMOND C, PAGNOUX C, KHELLAF 15. HELLMICH B, CSERNOK E, FREDENHAGEN Finally, in agreement with our results, M et al.: Eosinophilic granulomatosis with G, GROSS WL: A novel high sensitivity ELI- low levels of circulating RF (21) and polyangiitis (Churg-Strauss): clinical charac- SA for detection of antineutrophil cytoplasm antibodies against proteinase-3. Clin Exp ANA (2) have occasionally been de- teristics and long-term followup of the 383 patients enrolled in the French Vasculitis Rheumatol 2007; 25 (Suppl. 44): S1-5. scribed in EGPA but appeared to be Study Group cohort. Arthritis Rheum 2013; 16. ADAMUS G, APTSIAURI N, GUY J et al.: The nonspecific. Regarding anti-EPO- an 65: 270-81. occurrence of serum autoantibodies against tibodies, our results argue against the 5. MOOSIG F, BREMER JP, HELLMICH B et al.: A enolase in cancer-associated retinopathy. Clin Immunol Immunopathol 1996; 78: 120- presence of these antibodies in health vasculitis centre based management strategy leads to improved outcome in eosinophilic 9. and disease. We also consider that dif- granulomatosis and polyangiitis (Churg- 17. LUQMANI RA, BACON PA, MOOTS BA et ferences in techniques may explain in- Strauss, EGPA): monocentric experiences in al.: Birmingham Vasculitis Activity Score consistency in results when compared 150 patients. Ann Rheum Dis 2013; 72: 1011- (BVAS) in systemic necrotizing vasculitis. Q 7. J Med 1994; 87: 1-8. to previously published studies (8). 6. PANCHOLI V: Multifunctional a-enolase: its 18. LYONS PA, PETERS JE, ALBERICI F et al.: Ge- Summarising, the detection of autoan- role in diseases. Cell Mol Life Sci 2001; 58: nome-wide association study of eosinophilic tibodies in the sera of EGPA patients 902-20. granulomatosis with polyangiitis reveals supports an evidence for the role of 7. TAKIGUCHI J, OHIRA H, RAI T et al.: Anti- genomic loci stratified by ANCA status. Nat eosinophil peroxidase antibodies detected in Commun 2019; 10: 5120. autoimmune mechanisms in the de- patients with primary biliary cirrhosis. Hepa- 19. OLSEN E, RASMUSSEN HH, CELIS JE: Iden- velopment of the disease. Anti-alpha- tol Res 2005; 32: 33-7. tification of proteins that are abnormally enolase antibodies occurred in a high 8. DOLMAN KM, DAMSMA I, TOOL ATJ et al.: A regulated in differentiated cultured human percentage of EGPA patients at high novel specificity of anticytoplasmic autoan- keratinocytes. Electrophoresis 1995; 16: tibodies directed against eosinophil peroxi- 2241-8. titres with usually negative IFT for dase. Clin Exp Immunol 1993; 92: 58-64. 20. SHIIYAMA R, CHEN KR, ISHIBASHI M: A ANCA and were associated with skin 9. MOODIE FD, LEAKER B, CAMBRIDGE G et Case of Cutaneous Arteritis Presenting as involvement. We believe that enolase al.: Alpha-enolase: a novel cytosolic autoan- Infiltrated Erythema in Eosinophilic Granu- lomatosis With Polyangiitis: Features of the could serve as a new antigenic target in tigen in ANCA positive vasculitis. Kidney Int 1993; 43: 675-81. Unique Morphological Evolution of Arteritis the disease pathogenetic mechanisms. 10. ADAMUS G, AMUNDSON D, SEIGEL GM, as a Diagnostic Clue. Am J Dermatopathol Further studies with a greater number MACHNICKI M: Anti-enolase-alpha autoan- 2019; 41: 832-34. of patients will reassure our results and tibodies in cancer-associated retinopathy: 21. LANGHAM JG, ELKON KB, PUSEY CD, HUGHES GR: Systemic vasculitis with asth- test for autoantibodies directed against epitope mapping and cytotoxicity on retinal cells. J Autoimmun 1998; 11: 671-77. ma and eosinophilia: a clinical approach to unique epitopes of alpha-enolase spe- 11. ADAMUS G: Impact of Autoantibodies the Churg-Strauss Syndrome. Medicine (Bal- cific for EGPA with a potential- diag against Glycolytic Enzymes on Pathogenic- timore) 1984; 63: 65-81. nostic value and scientific rationale for ity of Autoimmune Retinopathy and Other Autoimmune Disorders. Front Immunol targeted therapy. 2017; 8: 505.

Clinical and Experimental Rheumatology 2021 S-87