Xiao-wei Yang 1,2,3,4, Ying Tan 1,2,3,4, Feng Yu 1,2,3,4 and Ming-hui Zhao 1,2,3,4. Introduction

Transcription

1 Nephrol Dial Transplant (2012) 27: doi: /ndt/gfs179 Advance Access publication 13 June 2012 Combination of anti-c1q and anti-dsdna antibodies is associated with higher renal disease activity and predicts renal prognosis of patients with lupus nephritis Xiao-wei Yang 1,2,3,4, Ying Tan 1,2,3,4, Feng Yu 1,2,3,4 and Ming-hui Zhao 1,2,3,4 1 Renal Division, Department of Medicine, Peking University First Hospital, Beijing, PR China, 2 Institute of Nephrology, Peking University, Beijing, PR China, 3 Key Laboratory of Renal Disease, Ministry of Health of China, Beijing, PR China and 4 Key Laboratory of Chronic Kidney Disease Prevention and Treatment, Ministry of Education of China, Beijing, PR China Correspondence and offprint requests to: Feng Yu; yufengevert@gmail.com Abstract Background. Although nephritogenic autoantibodies are considered to play a central role in the initiation of lupus nephritis, whether these autoantibodies are associated with renal clinical and pathological activity or renal outcome is still controversial. Here, we investigated the associations of certain serum autoantibodies with renal disease activity and renal outcome in a large cohort of Chinese patients with lupus nephritis. Methods. One hundred and thirty-six Chinese patients with biopsy-proven lupus nephritis and with long-term follow up data were studied. Sera at renal biopsy were tested for a panel of autoantibodies, including anti-nuclear antibodies, anti-double-stranded DNA (anti-dsdna) antibodies, anti-extractable nuclear antigen antibodies, anti-c-reactive protein antibodies, anti-c1q antibodies, anti-cardiolipin antibodies and anti-β2-glycoprotein I antibodies. Associations of these autoantibodies with clinical features, laboratory findings, histopathological data and renal outcomes were further investigated. Results. Among the various autoantibodies, anti-dsdna and anti-c1q antibodies were better than other antibodies to evaluate the renal disease activity. Anti-dsDNA antibodies were correlated with higher incidence of leukocyturia (P < 0.05), total pathological activity index (AI) score (P < 0.05), endocapillary hypercellularity (P < 0.05), subendothelial hyaline deposits (P < 0.05) and leukocyte infiltration (P < 0.05). Anti-C1q antibodies were correlated with leukocyturia (P < 0.01), hematuria (P < 0.003) and the majority of the histopathological AIs including total AI score (P < 0.003), endocapillary hypercellularity (P < 0.003), cellular crescents (P < 0.05), karyorrhexis/fibrinoid necrosis (P < 0.003), subendothelial hyaline deposits (P < 0.003) and leukocyte infiltration (P < 0.01). Patients with both anti-dsdna and anti-c1q antibodies had higher renal disease activity and poorer renal outcome (log-rank test: P = 0.048) compared with those without the two antibodies. In univariate survival analysis of renal prognosis, neither the presence of anti-c1q nor the presence of antidsdna antibodies was a risk factor of renal survival. However, the combination of the two antibodies predicted renal prognosis (hazard ratio 4.40, 95% confidence interval: , P = 0.02). Conclusions. Anti-C1q antibodies are more closely correlated with renal disease activity than the other autoantibodies. The combination of anti-c1q and anti-dsdna autoantibodies indicates higher renal disease activity and predicts poor renal outcome. Keywords: anti-c1q autoantibody; anti-dsdna autoantibody; autoantibodies; lupus nephritis Introduction Systemic lupus erythematosus (SLE) is the prototype of autoimmune disease and is characterized by the production of a variety of autoantibodies. Of the various major organ manifestations of lupus, lupus nephritis continues to be a principal cause of morbidity and mortality [1]. After decades of scrutiny, it was proposed that the development of glomerulonephritis in patients with SLE was correlated with the presence of some specific nephritogenic autoantibodies, such as anti-double-stranded DNA (anti-dsdna) antibodies [2 4], anti-c1q antibodies [5 8], anti-ribosomal P proteins antibodies [9], anti-sm antibodies [10], and anti-c-reactive protein (CRP) antibodies [11, 12], whereas more than 100 autoantibodies were reported in SLE. Although there is a consensus stating that these nephritogenic autoantibodies may play a central role in the initiation of lupus nephritis, whether these autoantibodies are associated with renal clinical and pathological activity or renal outcome is still controversial [13, 14]. To identify predictive markers of renal disease activity, which allows earlier diagnosis of onset or relapse and earlier application of appropriate treatment, is the major focus of clinical research in lupus nephritis, and recent studies proposed that the combination of the classical biomarkers might show a better specificity or sensitivity [15, 16]. The Author Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved. For Permissions, please journals.permissions@oup.com

2 Anti-C1q and anti-dsdna antibodies in lupus nephritis 3553 This study investigated the associations of clinical manifestations, laboratory data, pathological features and outcome of lupus nephritis with a panel of autoantibodies, including anti-nuclear antibodies (ANA), anti-extractable nuclear antigen (ENA), anti-dsdna, anti-c1q, anti-crp, anti-cardiolipin and anti-β2-glycoprotein I (anti-β2-gpi) antibodies in a large cohort of Chinese patients with lupus nephritis. We found that the presence of antidsdna or anti-c1q antibodies at renal biopsy was correlated with renal disease activity, and the combination of anti-dsdna and anti-c1q antibodies indicated higher renal disease activity and predicted renal outcome. Materials and methods Patients One hundred and thirty-six Chinese patients with renal biopsy-proven lupus nephritis, diagnosed from 2000 to 2006 in Peking University First Hospital, were enrolled in this study. All the patients fulfilled the 1997 American College of Rheumatology revised criteria for SLE [17]. Clinical evaluation The following clinical data were collected and analyzed: gender, fever, malar rash, photosensitivity, oral ulcer, alopecia, arthritis, serositis, neurologic disorder, anemia, leukocytopenia, thrombocytopenia, hematuria and leukocyturia. The clinical disease activity was assessed by the Systemic Lupus Erythematosus Disease Activity Index (SLEDAI) [18] and European Consensus Lupus Activity Measurement (ECLAM) [19, 20]. The criteria for clinical remission were the same as we reported previously [21]. The remission of lupus nephritis includes complete remission and partial remission. Complete remission was defined as a urinary protein excretion of <0.3 g/day, with normal urinary sediments (red blood cell < 3/high power (HP), white blood cell < 5/HP), serum albumin and renal function. Partial remission was defined as the presence of any one of the following features: a decrease in serum creatinine to below 130 μmol/l for patients with a baseline serum creatinine of 130 μmol/ L, but 260 μmol/l; a decrease in serum creatinine by >50% for patients with a baseline serum creatinine of >260 μmol/l; a decrease in urinary protein excretion by >50%, and below 3.0 g/24 h, with a serum albumin of 30 g/l and stable renal function. The patients were followed up in our outpatient clinic specified for lupus nephritis. The primary endpoint was defined as death and the secondary endpoints were defined as end-stage renal disease (ESRD) or doubling of serum creatinine. Renal histopathology The renal biopsy specimens were examined by light microscopy and direct immunofluorescence. Lupus nephritis was re-classified according to the International Society of Nephrology/Renal Pathology Society (ISN/RPS) 2003 classification system [22]. Light microscopy examination Renal biopsy specimens were fixed in 4.5% buffered formaldehyde for light microscopy. Consecutive serial 3 μm-thick sections were used for histological staining. Stains employed included hematoxylin and eosin (H&E), periodic acid-schiff, silver methenamine (Meth) and Masson s trichrome. Pathological parameters such as activity indices (AIs) and chronicity indices were approached by renal pathologists using a previously reported system involving semi-quantitative scoring of specific biopsy features with mild modification [23, 24]. The AI include endocapillary hypercellularity, cellular crescents, karyorrhexis/fibrinoid necrosis, subendothelial hyaline deposits, interstitial inflammation and leukocyte infiltration, whereas the chronicity indices include glomerular sclerosis, fibrous crescents, tubular atrophy and interstitial fibrosis. Direct immunofluorescence examination The intensity of fluorescence of direct immunofluorescence for immunoglobulin G (IgG), immunoglobulin A, immunoglobulin M, C3, C1q and fibrin deposits was semi-quantitatively graded from 0 to 4, respectively. Serum samples Sera from patients with lupus nephritis were obtained from peripheral blood at the time of renal biopsy. All the sera were stored at 20 C until use. Informed consent was obtained for blood sampling and renal biopsy from each patient. The research was in compliance of the Declaration of Helsinki. The design of this work was approved by the local ethical committees. Assay of autoantibodies Serum ANA were detected using indirect immunofluorescence assay (EUROIMMUN, Lübeck, Germany) and anti-dsdna antibodies were detected using Crithidia luciliae indirect immunofluorescence test (EUROIMMUN). Anti-ENA antibodies, including anti-sm, anti-ssa (Sjögren s syndrome A antigen), anti-ssb (Sjögren s syndrome B antigen) and anti-ribonucleoprotein (RNP) antibodies, were detected using immunodotting assay (EUROIMMUN). Anti-cardiolipin and antiβ2-gpi antibodies were detected using enzyme-linked immunosorbent assay (ELISA) (EUROIMMUN). Anti-C1q antibodies were detected by a method established before [8]. In brief, purified normal human C1q (Sigma), diluted at 5 μg/ml in 0.05 M bicarbonate buffer (ph 9.6), was coated onto the wells of onehalf of a polystyrene microtitre plate (Costar, Mankato, MN). The wells in the other half were coated with the same bicarbonate buffer alone to act as antigen-free wells to exclude non-specific binding. The volumes of each well for this step and for subsequent steps were 100 μl. All incubations were carried out at 37 C for 1 h and the plates were washed three times with 0.01 M phosphate-buffered saline containing 0.1% Tween-20 (PBST). Then, the plates were blocked with PBST containing 0.01% gelatin. The sera diluted at 1:200 with a high-salt buffer, including PBST containing 1% bovine serum albumin (PBSA), and 0.5 M NaCl were added in duplicate to both antigen-coated and antigen-free wells. Each plate contained a blank control, negative control and a known positive control. After incubation and washing, the wells were then incubated with 1:5000 diluted horseradish peroxidase-conjugated goat anti-human IgG (Zhongshan Biotech, Beijing, China) to detect anti-c1q antibodies. The reaction was developed with a 0.1 M citrate phosphate buffer (ph 5.0) containing 0.04% O-phenylenediamine and 0.1% H 2 O 2, and then the reaction was stopped with 1 M H 2 SO 4. The results were recorded as the net optical absorbance (the average value of antigen wells minus the average value of antigen-free wells) at 490 nm in an ELISA reader (Bio- Rad 550, Tokyo, Japan). The antibody levels were expressed as percentage (unit) of a known positive control (the binding of the positive control was regarded as 100 units). Samples were considered positive if they exceeded the mean plus 2 SD from the 63 normal bloods. Anti-CRP antibodies were detected by a method established before [11]. The procedures were similar to the assay for the detection of anti- C1q antibodies, except for the purified normal human CRP (Sigma, St Louis, MO) was diluted at 5.6 μg/ml in 0.05 M bicarbonate buffer (ph 9.6) and the sera diluted at 1:50 with PBSA. Statistical analysis Statistical software SPSS 16.0 (SPSS, Chicago, IL) was employed for statistical analysis. Quantitative data were expressed as mean ± SD, median with range (minimum, maximum) or number (%). For comparison of clinical and pathological features of patients, t-tests, the Mann Whitney U-test, one-way ANOVA, the Kruskal Wallis test and χ 2 test were used with the Bonferroni correction or Dunnett s test for multiple comparisons. The Cox regression model was applied to identify prognostic factors associated with renal outcome. Only variables with P < 0.05 in the univariate Cox regression analysis were used in the multiple Cox regression. Results were expressed as hazard ratio (HR) with 95% confidence intervals (CI). Patients survival curves were derived from the Kaplan Meier analysis and were compared by log-rank test. Statistical significance was considered as P < 0.05.

3 3554 X. Yang et al. Results Baseline data of patients with lupus nephritis General clinical and renal histopathological profiles of the patients at renal biopsy are listed in Tables 1 and 2. Of the 136 patients, 115 (84.6%) were females and 21 (15.4%) were males. The age of the patients ranged from 14 to 70 with a median age of 32. According to the 2003 classification of lupus nephritis, 8 patients was classified as Class II (5.9%), 20 as Class III (14.7%, including 5 as Class III + V), 73 as Class IV [53.7%, 7 as Class IV-segmental (IV-S) and 66 as Class IV-global (IV-G), including 2 as Class IV + V] and 35 as Class V (25.7%). None were Classes I and VI in this study. Classes III and IV were further subdivided into active (A), active/chronic (A/C) and chronic (C) groups. Table 1. General clinical profiles of patients with lupus nephritis at renal biopsy Number of patients 136 Gender (male/female) 21/115 Age (median, range) (years) 32 (14 70) The time between presentation of lupus nephritis 27.0 ± 47.6 (0 240) and biopsy (mean ± SD and range) (months) Number of fever (non-infectious) (%) 71 (52.2) Number of malar rash (%) 78 (57.4) Number of photosensitivity (%) 33 (24.3) Number of oral ulcer (%) 42 (30.9) Number of alopecia (%) 55 (40.4) Number of arthralgia (%) 82 (60.3) Number of serositis (%) 29 (21.3) Number of neurologic disorder (%) 13 (9.6) Number of anemia (%) 93 (68.4) Number of leukocytopenia (%) 26 (19.1) Number of thrombocytopenia (%) 16 (11.8) Number of hematuria (%) 106 (77.9) Number of leukocyturia (non-infection) (%) 69 (50.7) Number of acute renal failure (%) 35 (25.7) Hemoglobin (g/l) ± Urine protein (g/24 h) 4.97 ± 3.58 Serum creatinine (μmol/l) ± C3 (g/l) 0.46 ± 0.22 SLEDAI ± 5.80 ECLAM 4.83 ± 1.58 SLEDAI, SLE Disease Activity Index; ECLAM, European Consensus Lupus Activity Measurement. Within Class III, the number of III (A) was 18, III (A/C) was 2 and III (C) was 0. Within Class IV-S, the number of IV-S (A) was 6, IV-S (A/C) was 1 and IV-S (C) was 0. Within Class IV-G, the number of IV-G (A) was 49, IV-G (A/C) was 17 and IV-G (C) was 0. All of the patients received oral prednisone therapy. The majority of patients completed treatment with oral cyclophosphamide (21 of 136) or monthly intravenous cyclophosphamide ( mg/month) (59 of 136). The other patients received mycophenolate mofetil (10 of 136), leflunomide (38 of 136) and azathioprine (3 of 136). Five patients received prednisone alone. Most patients achieved clinical remission, 36 with complete remission and 78 with partial remission. Twenty-two patients presented as treatment failure. During the follow-up, with an average of 57.9 ± 59.0 months (6 360 months), 1 patient died of chronic heart failure, 13 patients reached secondary endpoints including 7 with a doubling of serum creatinine and 6 with ESRD. Prevalence of autoantibodies and association with renal disease activity Ninety-six of the 136 (70.6%) patients were anti-dsdna antibody-positive, and the presence of anti-dsdna antibodies was not correlated with SLEDAI or ECLAM, but was associated with a lower serum C3 level (P < 0.01; Tables 3 and 4). Regarding renal disease activity, patients with anti-dsdna antibodies had higher indices reflecting active renal disease, especially certain histopathological indices, including leukocyturia (P < 0.05), total AI score (P < 0.05), endocapillary hypercellularity (P < 0.05), subendothelial hyaline deposits (P < 0.05) and leukocyte infiltration (P < 0.05). However, the presence of anti-dsdna antibodies was not associated with the levels of proteinuria, serum creatinine and histopathological parameters of chronicity indices. Seventy-four of the 136 (54.4%) patients were anti-c1q antibody-positive. The presence of anti-c1q antibodies was correlated with higher SLEDAI, ECLAM and a lower serum C3 level (P < 0.003, <0.003 and <0.003, respectively). Patients with anti-c1q antibodies had higher acute renal damage indices, including leukocyturia Table 3. Autoantibodies profiles of patients with lupus nephritis at renal biopsy Table 2. Renal histopathological profiles of patients with lupus nephritis at renal biopsy Number of biopsies 136 AI score 8.03 ± 5.12 Endocapillary hypercellularity 2.09 ± 1.11 Cellular crescents 1.49 ± 1.90 Karyorrhexis/fibrinoid necrosis 1.07 ± 1.00 Subendothelial hyaline deposits 1.29 ± 1.22 Interstitial inflammatory cell infiltration 1.24 ± 0.85 Glomerular leukocyte infiltration 0.79 ± 0.90 Chronicity index score 2.84 ± 2.17 Glomerular sclerosis 0.51 ± 0.73 Fibrous crescents 0.18 ± 0.45 Tubular atrophy 1.16 ± 0.80 Interstitial fibrosis 1.00 ± 0.82 Positive autoantibodies % of cases ANA 97.8 (133/136) Anti-dsDNA antibody 70.6 (96/136) Anti-SSA antibody 38.2 (52/136) Anti-SSB antibody 8.1 (11/136) Anti-Sm antibody 15.4 (21/136) Anti-RNP antibody 27.9 (38/136) Anti-C1q antibody 54.4 (74/136) Anti-CRP antibody 35.3 (48/126) Anti-cardiolipin antibody 8.0 (9/113) Anti-β2-GPI antibody 8.7 (10/115) ANA, anti-nuclear antibody; SSA, Sjögren s syndrome A antigen; SSB, Sjögren s syndrome B antigen; RNP, ribonucleoprotein; CRP, C-reactive protein; β2-gpi, β2-glycoprotein I.

4 Table 4. Associations between autoantibodies and clinical and histopathological parameters of patients with lupus nephritis Anti-dsDNA /+ Anti-SSA /+ Anti-SSB /+ Anti-Sm /+ Anti-RNP /+ Anti-C1q /+ Anti-CRP /+ Clinical and laboratory data SLEDAI NS NS NS NS NS ± 4.12/20.26 ± 5.78*** ± 4.12/18.71 ± 6.05* ECLAM NS NS NS NS NS 4.29 ± 1.53/5.28 ± 1.49*** NS C ± 0.26/0.42 ± 0.19** NS NS NS NS 0.57 ± 0.25/0.36 ± 0.13*** NS (g/l) Proteinuria NS NS NS NS NS NS NS (g/day) Hematuria (%) NS NS NS NS 79 (80.6%)/24 (63.2%)* 37 (59.7%)/66 (89.2%)*** NS Leukocyturia (%) 13 (32.5%)/53 (55.2%)* NS NS NS 55 (56.1%)/11 (28.9%)** 21 (33.9%)/45 (60.8%)** NS Serum creatinine NS NS NS NS NS NS NS (μmol/l) Histopathological data AI score 6.60 ± 4.97/8.96 ± 4.92* NS NS NS 8.83 ± 4.93/6.82 ± 5.07* 6.32 ± 4.58/9.89 ± 4.84*** NS Endocapillary 1.83 ± 1.20/2.29 ± 0.96* NS NS NS 2.29 ± 1.02/1.82 ± 1.09* 1.84 ± 1.16/2.42 ± 0.88*** NS hypercellularity Cellular crescents NS NS NS NS NS 1.13 ± 1.43/1.85 ± 2.16* NS Karyorrhexis/ NS NS NS NS NS 0.61 ± 0.93/1.51 ± 0.86*** NS fibrinoid necrosis Subendothelial 0.88 ± 1.14/1.53 ± 1.20* NS NS NS NS 0.94 ± 1.17/1.68 ± 1.16*** NS hyaline deposits Interstitial NS NS NS NS NS NS NS inflammation Leukocyte infiltration 0.55 ± 0.81/0.90 ± 0.92* 0.96 ± 1.00/0.58 ± 0.67** NS NS NS 0.58 ± 0.78/1.01 ± 0.96** NS Chronicity index NS NS NS 3.11 ± 2.24/2.05 ± 1.43** NS 3.45 ± 2.57/2.51 ± 1.64* NS score Glomerular sclerosis NS NS 0.57 ± 0.74/0.18 ± 0.40* NS NS 0.74 ± 0.87/0.36 ± 0.54*** NS Fibrous crescents NS NS 0.19 ± 0.47/0.00 ± 0.00*** NS NS 0.27 ± 0.58/0.09 ± 0.29*** NS Tubular atrophy NS NS NS 1.27 ± 0.82/0.86 ± 0.57** NS 1.29 ± 0.89/1.14 ± 0.71** NS Interstitial fibrosis NS NS NS 1.10 ± 0.84/0.67 ± 0.58* NS 1.16 ± 0.91/0.93 ± 0.73* NS SLEDAI, SLE Disease Activity Index; ECLAM, European Consensus Lupus Activity Measurement; SSA, Sjögren s syndrome A antigen; SSB, Sjögren s syndrome B antigen; RNP, ribonucleoprotein; CRP, C-reactive protein; NS, no significance. Note: *P < 0.05, **P < 0.01 and ***P < 0.003; P values shown were before the Bonferroni corrections. Statistical significance was considered as P < (0.05/19 comparisons) after the Bonferroni test. Anti-C1q and anti-dsdna antibodies in lupus nephritis 3555

5 3556 X. Yang et al. (P < 0.01) and hematuria (P < 0.003), and most of histopathological parameters of AIs including AI score (P < 0.003), endocapillary hypercellularity (P < 0.003), cellular crescents (P < 0.05), karyorrhexis/fibrinoid necrosis (P < 0.003), subendothelial hyaline deposits (P < 0.003) and leukocyte infiltration (P < 0.01). It was correlated with lower histopathological parameters of chronicity indices, including chronicity index score, glomerular sclerosis, fibrous crescents, tubular atrophy and interstitial fibrosis (P < 0.05, <0.003, <0.003, <0.01 and <0.05, respectively). Anti-ENA autoantibodies were correlated with some renal clinical or histopathological parameters. The presence of anti-ssa antibodies was correlated with milder renal leukocyte infiltration (P < 0.01). The presence of anti-ssb antibodies was correlated with milder glomerular sclerosis and fibrous crescents (P < 0.05 and <0.003, respectively); patients with anti-sm antibodies had lower scores of some chronic histopathological parameters of renal injury, including total chronicity index score, tubular atrophy and interstitial fibrosis (P < 0.01, <0.01 and <0.05, respectively). The presence of anti-rnp antibodies was correlated with lower hematuria, leukocyturia, AI score and endocapillary hypercellularity (P < 0.05, <0.01, <0.05 and <0.05, respectively). In this cohort, the presence of anti-crp antibodies was associated with higher SLEDAI score (P < 0.05), without any correlation with parameters of renal disease activity. The prevalence of anti-cardiolipin or anti-β2-gpi antibodies was relatively low. We did not find any correlation between the two autoantibodies and clinical or histological disease activity. The P-values shown above were not corrected for multiple tests, and statistical significance was considered as P < (0.05/19 comparisons) after the Bonferroni correction. The detailed data were presented in Supplementary Table S1. Association of renal disease activity and the presence of anti-dsdna and anti-c1q antibodies As both anti-dsdna and anti-c1q antibodies were associated (either borderline or significantly) with renal disease activity based on the above results, we stratified the patients into four groups according to the presence of anti-dsdna and anti-c1q antibodies. In this cohort, 39.7% patients were double positive with the two autoantibodies, whereas 14.7% patients had neither of them. The clinical and histological parameters of renal disease activity among the four groups were further analyzed (details in Table 5). There existed significant differences in most of the renal parameters, including total AI score, endocapillary hypercellularity, karyorrhexis/fibrinoid necrosis, subendothelial hyaline deposits, leukocyte Table 5. Association of renal disease activity and the presence of anti-dsdna and anti-c1q antibodies Anti-dsDNA / anti-c1q Anti-dsDNA+ only Anti-C1q+ only Anti-dsDNA+/ anti-c1q+ P-value a Demographic data Number of patients (%) 20 (14.7%) 42 (30.9%) 20 (14.7%) 54 (39.7%) Gender (male/female) 1/19 6/36 4/16 10/ Age (years) ± ± ± ± Clinical and laboratory data SLEDAI ± 4.10*** ± 3.86*** ± ± 5.71 <0.001 ECLAM 4.00 ± 1.41** 4.43 ± 1.58* 5.35 ± ± C3 (g/l) 0.68 ± 0.29*** 0.52 ± 0.22*** 0.41 ± ± 0.13 <0.001 Proteinuria (g/day) 5.66 ± ± ± ± Hematuria (%) 14 (70.0%) 26 (61.9%) 18 (90.0%) 48 (88.9%) Leukocyturia (%) 5 (25%) 18 (42.9%) 10 (50.0%) 36 (66.7%) Serum creatinine (μmol/l) ± ± ± ± Histopathological data Histopathological type Class II (%) 1 (5.0%) 3 (7.1%) 3 (15.0%) 1 (1.9%) Class III + IV (%) 10 (50.0%) 23 (54.8%) 14 (70.0%) 46 (85.2%) Class V (%) 9 (45.0%) 16 (38.1%) 3 (15.0%) 7 (13.0%) AI score 4.90 ± 3.93*** 7.00 ± 4.75*** 8.30 ± ± 4.53 <0.001 Endocapillary hypercellularity 1.65 ± 1.31*** 1.92 ± 1.09** 2.00 ± 1.08* 2.57 ± Cellular crescents 1.00 ± ± ± ± Karyorrhexis/fibrinoid necrosis 0.40 ± 0.82*** 0.71 ± 0.97*** 1.30 ± ± 0.81 <0.001 Subendothelial hyaline deposits 0.55 ± 0.94*** 1.11 ± 1.23** 1.20 ± 1.24* 1.85 ± 1.09 <0.001 Interstitial inflammation 0.95 ± ± ± ± Leukocyte infiltration 0.35 ± 0.59*** 0.69 ± ± ± Chronicity index score 3.25 ± ± ± ± Glomerular sclerosis 0.60 ± ± 0.83* 0.40 ± ± Fibrous crescents 0.20 ± ± ± ± Tubular atrophy 1.25 ± ± ± ± Interstitial fibrosis 1.20 ± ± ± ± SLEDAI, SLE Disease Activity Index; ECLAM, European Consensus Lupus Activity Measurement. a Note: *P < 0.05, **P < 0.01 and ***P < 0.001, compared with the anti-dsdna+/anti-c1q+ group with Dunnett s test for multiple comparison correction.

6 Anti-C1q and anti-dsdna antibodies in lupus nephritis 3557 infiltration and glomerular sclerosis (P < 0.001, 0.001, <0.001, <0.001, and 0.015, respectively). Among the four groups, patients who were anti-dsdna +/anti-c1q+ had the highest scores of these parameters, especially those histopathological indices, with significant difference compared with patients who were anti-dsdna /anti-c1q or anti-dsdna+ only. Renal outcome and the presence of anti-dsdna and anti- C1q antibodies As shown in Table 6, in univariate survival analysis of renal prognosis of our patients with lupus nephritis, we found that sex (male), ECLAM, serum creatinine, anti- SSB antibodies, total AI score, cellular crescents, karyorrhexis/fibrinoid necrosis, interstitial inflammation, tubular atrophy, interstitial fibrosis and total chronicity index score were the risk factors for long-term renal outcome in lupus nephritis. Neither the presence of anti-c1q nor the presence of anti-dsdna antibodies at renal biopsy was a risk factor for renal outcome. However, the combination of the two antibodies could predict renal prognosis (HR 4.40, 95% CI: , P = 0.02). In the Kaplan Meier analysis, patients with both anti-dsdna and anti- C1q antibodies at renal biopsy had significantly poorer renal outcome in comparison with those without the two autoantibodies both positive (P = 0.048; Figure 1). In a further multivariate Cox hazard analysis of the risk factors, only tubular atrophy could independently indicate renal prognosis (Table 7). Discussion The kidney is one of the most commonly involved organs in SLE, and the development of nephritis is a major determinant of disease morbidity and mortality. The identification of non-invasive biomarkers in lupus nephritis is important for predicting renal involvement, reflecting clinical and pathological disease activity, guiding Table 6. Univariate survival analysis of renal prognosis of patients with lupus nephritis Univariate Hazard ratio 95% CI P-value a Demographic data Age , Sex , Clinical and laboratory data SLEDAI , ECLAM , C , Proteinuria , Leukocyturia , Hematuria , Serum creatinine , <0.001 ANA , Anti-dsDNA antibody , Anti-Sm antibody , Anti-SSA antibody , Anti-SSB antibody , Anti-RNP antibody , Anti-C1q antibody , Anti-CRP antibody , Anti-cardiolipin antibody , Anti-β2-GPI antibody , Anti-C1q antibody+/anti-dsdna antibody , Renal pathological data AI score , Endocapillary hypercellularity , Cellular crescents , Karyorrhexis/fibrinoid necrosis , Subendothelial hyaline deposits , Interstitial inflammation , <0.001 Leukocyte infiltration , Chronicity index score , Glomerular sclerosis , Fibrous crescents , Tubular atrophy , <0.001 Interstitial fibrosis , SLEDAI, SLE Disease Activity Index; ECLAM, European Consensus Lupus Activity Measurement; ANA, anti-nuclear antibody; SSA, Sjögren s syndrome A antigen; SSB, Sjögren s syndrome B antigen; RNP, ribonucleoprotein; CRP, C-reactive protein; β2-gpi, β2-glycoprotein I; CI, confidence interval. a Only variables with P < 0.05 in the univariate Cox regression analysis were used in the multiple Cox regression.

7 3558 X. Yang et al. treatment, monitoring relapse and judging prognosis. In recent years, several autoantibodies have been suggested as specific and sensitive biomarkers for the development of lupus nephritis, but the results are still controversial among different studies. Using a large lupus nephritis cohort in our referral center, we found that anti-c1q antibody was more closely associated with renal activity than the other autoantibodies, and the combination of antidsdna and anti-c1q antibodies could evaluate the renal disease activity best. Anti-dsDNA antibodies are diagnostic for SLE [17]. In our study, the prevalence of anti-dsdna antibodies was much higher than other autoantibodies except ANA, and we found that the presence of anti-dsdna antibodies were correlated with higher indices reflecting active renal disease, especially certain histopathological indices, including endocapillary hypercellularity, subendothelial hyaline deposits, leukocyte infiltration and total AI score. However, most of the P-values were not small enough to Fig. 1. The Kaplan Meier analysis of renal survival between patients with both anti-dsdna and anti-c1q antibodies and those without the two autoantibodies at renal biopsy. Group 1: patients with both antidsdna and anti-c1q antibodies; Group 2: patients without anti-dsdna and anti-c1q antibodies simultaneously positive. pass through the Bonferroni corrections for multiple comparisons. On the other hand, the presence of anti-dsdna antibodies at renal biopsy was not a risk factor of the renal survival by the log-rank test, and nearly 30% of our patients with active renal pathological features did not have serum anti-dsdna antibodies at renal biopsy. Some previous studies even showed that the association between anti-dsdna antibodies and SLE disease activity was limited in a few patients [25, 26]. So, it is important to find better surrogates of lupus disease activity. C1q, the first component of the classical pathways of complement, played an ambivalent role in the pathogenesis of SLE. On the one hand, the over-activation of it could lead to organ damage during inflammatory process, while on the other hand, the deficiency of it was a risk factor to develop SLE due to its role in the clearance of apoptotic bodies [27]. A number of reports have described the association of anti-c1q antibodies with the presence and the disease activity of renal disease in SLE, even slightly better than the other tests, including anti-dsdna antibodies, to indicate the clinical activity of lupus nephritis [16]. The incidence of anti-c1q antibodies in lupus nephritis reported in previous studies varied widely [6, 8]. The discrepancies might be due to the different pathological types, different assays and different epitopes [28]. Our current study also showed that the prevalence of anti-c1q antibodies was 54.4%, and the presence of anti-c1q antibodies was closely correlated with acute injury indices of renal histopathology, with a better correlation compared with anti-dsdna antibodies, such as endocapillary hypercellularity, karyorrhexis/fibrinoid necrosis, subendothelial hyaline deposits, leukocyte infiltration and total AI scores in lupus nephritis. However, similar to anti-dsdna antibodies, the presence of the autoantibodies was not a risk factor of the renal survival. A recent study highlighted the predictive value for renal flare by simultaneous detection of anti-c1q and antidsdna antibodies in lupus nephritis [15]. So, we further combined the two autoantibodies together to evaluate renal disease activity and renal outcome. The patients with the two autoantibodies simultaneously positive presented with the most severe renal histopathological disease Table 7. Multivariate survival analysis of renal prognosis of patients with lupus nephritis Multivariate Hazard ratio 95% CI P-value Sex , ECLAM , Serum creatinine , Anti-SSB antibody , Anti-C1q antibody+/anti-dsdna antibody , AI score , Cellular crescents , Karyorrhexis/fibrinoid necrosis , Interstitial inflammation , Chronicity index score , Tubular atrophy , Interstitial fibrosis , ECLAM, European Consensus Lupus Activity Measurement; SSB, Sjögren s syndrome B antigen; CI, confidence interval.

8 Anti-C1q and anti-dsdna antibodies in lupus nephritis 3559 activity, such as endocapillary hypercellularity, karyorrhexis/fibrinoid necrosis, subendothelial hyaline deposits, leukocyte infiltration and total AI score. More importantly, using univariate survival analysis and renal survival curve comparison, we found that the combination of the two autoantibodies could predict poor renal outcome. In conclusion, anti-c1q antibodies are more closely correlated with renal disease activity than other autoantibodies. The combination of anti-c1q and anti-dsdna autoantibodies was not only an indicator of higher renal disease activity but also a predictor of poor renal prognosis of patients with lupus nephritis. Supplementary data Supplementary data are available online at oxfordjournals.org. Acknowledgements. This work was supported by grants of Chinese 973 project (No. 2012CB517702), National Natural Science Foundation of China to Innovation Research Group (No ), National Natural Science Foundation of China (No ) and Beijing Natural Science Foundation (No ). Conflict of interest statement. None declared. References 1. Cervera R, Khamashta MA, Font J et al. Morbidity and mortality in systemic lupus erythematosus during a 10-year period: a comparison of early and late manifestations in a cohort of 1,000 patients. Medicine (Baltimore) 2003; 82: Swaak T, Smeenk R. Clinical significance of antibodies to double stranded DNA (dsdna) for systemic lupus erythematosus (SLE). Clin Rheumatol 1987; 6(Suppl 1): Okamura M, Kanayama Y, Amastu K et al. Significance of enzyme linked immunosorbent assay (ELISA) for antibodies to double stranded and single stranded DNA in patients with lupus nephritis: correlation with severity of renal histology. Ann Rheum Dis 1993; 52: Linnik MD, Hu JZ, Heilbrunn KR et al. Relationship between antidouble-stranded DNA antibodies and exacerbation of renal disease in patients with systemic lupus erythematosus. Arthritis Rheum 2005; 52: Sinico RA, Radice A, Ikehata M et al. Anti-C1q autoantibodies in lupus nephritis: prevalence and clinical significance. Ann N Y Acad Sci 2005; 1050: Marto N, Bertolaccini ML, Calabuig E et al. Anti-C1q antibodies in nephritis: correlation between titres and renal disease activity and positive predictive value in systemic lupus erythematosus. Ann Rheum Dis 2005; 64: Fang QY, Yu F, Tan Y et al. Anti-C1q antibodies and IgG subclass distribution in sera from Chinese patients with lupus nephritis. Nephrol Dial Transplant 2009; 24: Trendelenburg M, Lopez-Trascasa M, Potlukova E et al. High prevalence of anti-c1q antibodies in biopsy-proven active lupus nephritis. Nephrol Dial Transplant 2006; 21: Do NA, Viana VS, Testagrossa LA et al. Antibodies to ribosomal P proteins: a potential serologic marker for lupus membranous glomerulonephritis. Arthritis Rheum 2006; 54: Alba P, Bento L, Cuadrado MJ et al. Anti-dsDNA, anti-sm antibodies, and the lupus anticoagulant: significant factors associated with lupus nephritis. Ann Rheum Dis 2003; 62: Tan Y, Yu F, Yang H et al. Autoantibodies against monomeric C-reactive protein in sera from patients with lupus nephritis are associated with disease activity and renal tubulointerstitial lesions. Hum Immunol 2008; 69: Sjowall C, Zickert A, Skogh T et al. Serum levels of autoantibodies against C-reactive protein correlate with renal disease activity and response to therapy in lupus nephritis. Arthritis Res Ther 2009; 11: R Katsumata Y, Miyake K, Kawaguchi Y et al. Anti-C1q antibodies are associated with systemic lupus erythematosus global activity but not specifically with nephritis: a controlled study of 126 consecutive patients. Arthritis Rheum 2011; 63: Trendelenburg M. Anti-C1q antibodies as a diagnostic marker of proliferative lupus nephritis: comment on the article by Katsumata et al. Arthritis Rheum 2012; 64: , Matrat A, Veysseyre-Balter C, Trolliet P et al. Simultaneous detection of anti-c1q and anti-double stranded DNA autoantibodies in lupus nephritis: predictive value for renal flares. Lupus 2011; 20: Moroni G, Radice A, Giammarresi G et al. Are laboratory tests useful for monitoring the activity of lupus nephritis? A 6-year prospective study in a cohort of 228 patients with lupus nephritis. Ann Rheum Dis 2009; 68: Hochberg MC. Updating the American College of Rheumatology revised criteria for the classification of systemic lupus erythematosus. Arthritis Rheum 1997; 40: Bombardier C, Gladman DD, Urowitz MB et al. Derivation of the SLEDAI. A disease activity index for lupus patients. The Committee on Prognosis Studies in SLE. Arthritis Rheum 1992; 35: Bombardieri S, Vitali C, Caponi L et al. Activity criteria in systemic lupus erythematosus. Clin Exp Rheumatol 1994; 12(Suppl 11): S45 S Vitali C, Bencivelli W, Isenberg DA et al. Disease activity in systemic lupus erythematosus: report of the Consensus Study Group of the European Workshop for Rheumatology Research. II. Identification of the variables indicative of disease activity and their use in the development of an activity score. The European Consensus Study Group for Disease Activity in SLE. Clin Exp Rheumatol 1992; 10: Yu F, Tan Y, Liu G et al. Clinicopathological characteristics and outcomes of patients with crescentic lupus nephritis. Kidney Int 2009; 76: Weening JJ, D Agati VD, Schwartz MM et al. The classification of glomerulonephritis in systemic lupus erythematosus revisited. Kidney Int. 2004; 65: Austin HR, Boumpas DT, Vaughan EM et al. Predicting renal outcomes in severe lupus nephritis: contributions of clinical and histologic data. Kidney Int 1994; 45: Austin HR, Muenz LR, Joyce KM et al. Diffuse proliferative lupus nephritis: identification of specific pathologic features affecting renal outcome. Kidney Int 1984; 25: Petri M, Genovese M, Engle E et al. Definition, incidence, and clinical description of flare in systemic lupus erythematosus. A prospective cohort study. Arthritis Rheum 1991; 34: Esdaile JM, Abrahamowicz M, Joseph L et al. Laboratory tests as predictors of disease exacerbations in systemic lupus erythematosus. Why some tests fail. Arthritis Rheum 1996; 39: Pickering MC, Botto M, Taylor PR et al. Systemic lupus erythematosus, complement deficiency, and apoptosis. Adv Immunol 2000; 76: Tan Y, Zhou W, Yu F et al. Detection of anti-c1q antibodies and anti-c1q globular head domain antibodies in sera from Chinese patients with lupus nephritis. Mol Immunol 2009; 46: Received for publication: ; Accepted in revised form: