SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES

Objective

To investigate the abundance of autoantibodies to heterogeneous nuclear RNPs (hnRNPs) in systemic rheumatic diseases.

Methods

Recombinant human hnRNPs A1, B1, C1, E1, F, Gi, H1, I, K, and P2 were prepared. Antibodies to these antigens were determined by Western blotting and by enzyme-linked immunosorbent assay (ELISA) (for hnRNPs B1, E1, F, and H1) in serum samples obtained from patients with chronic fatigue syndrome (control subjects) and from patients with various connective tissue diseases.

Results

Western blotting analysis in 106 control subjects and 298 patients with a connective tissue disease revealed that antibodies to all tested hnRNP antigens, except hnRNP Gi, were significantly more prevalent in patients with Sjögren's syndrome (SS) than in control subjects. The highest reactivity was observed for hnRNPs B1, E1, F, and H1 (reactivity in >45% of patients with SS and in 2.8% of control subjects). Reactivity with hnRNPs B1, E1, F, and H1 was also evaluated by ELISA in 89 control subjects and 228 patients with a connective tissue disease. Reactivity with at least 2 of the 4 tested antigens was observed in 1.1% of control subjects, 16% of patients with systemic lupus erythematosus (SLE), and 18% of patients with SS. Reactivity with at least 3 of the 4 antigens was observed in 0% of the control subjects, 3.2% of patients with SLE, and 15% of patients with SS.

Conclusion

Several hnRNPs are target antigens in SS. The combined presence of antibodies to several hnRNPs was strongly associated with connective tissue disease in general and with SS in particular.

Heterogeneous nuclear RNPs (hnRNPs) are abundant nucleoplasmic pre–messenger RNA (mRNA)–binding proteins that have important roles in the biogenesis of mRNA. They associate with RNA polymerase II transcripts and form large multiprotein–RNA complexes that are substrates for RNA processing (1). Heterogeneous nuclear RNPs participate in various cellular functions, such as DNA repair and telomere elongation, chromatin remodeling and translocation, pre-mRNA 3′-end processing and mRNA stability, mRNA splicing and nuclear-cytoplasmic shuttling, translation, regulation of proteins implicated in mediating cellular growth, and apoptosis (for review, see refs.2 and3). There are ∼30 major hnRNP proteins, termed hnRNPs A1 through U. Heterogeneous nuclear RNPs are present in the nucleoplasm. Some of them, such as hnRNPs A1, D, F/H, and K, also have cytoplasmic expression (4). Most hnRNPs are composed of at least 1 RNA recognition motif (the RNA-binding domain).

Heterogeneous nuclear RNP proteins appear to be an important target of autoimmune responses. A number of hnRNP proteins (e.g., A1, A2, B, C, H, I, and R) have been described as autoantigens, as reviewed by Caporali et al (5). Antibodies to hnRNP A2 and its alternatively spliced variants B1 and B2 (the RA33 complex) are the most studied antibodies to hnRNP antigens. They have been described in rheumatoid arthritis (RA), systemic lupus erythematosus (SLE), and mixed connective tissue disease (MCTD) (for review, see ref.6). Antibodies to hnRNP A1 have been reported in SLE (7). Antibodies to hnRNP L have been identified in association with anti–hnRNP A/B antibodies (8). Anti–hnRNP I antibodies have been reported in systemic sclerosis (SSc) (9). Antibodies to hnRNP C1/C2 and to hnRNP R have been reported in sporadic cases (10, 11). HnRNP D (AUF1) was described as an autoantibody target in patients with systemic rheumatic diseases (mainly SLE and RA) (12). Antibodies to hnRNP G have been described in dogs with SLE (13). Recently, our group described antibodies to hnRNP H1 in patients with Sjögren's syndrome (SS) (14).

In an attempt to further distinguish the multitude of autoantibodies to hnRNPs in systemic rheumatic diseases, we screened for autoantibodies to several members of the hnRNP family.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES

Study population.

Two groups of serum samples were included in the study. The first group of samples was used for Western blot analysis. The samples were obtained from well-defined patients with an autoimmune connective tissue disease (n = 298). This group included patients with SLE (n = 71; male:female ratio 10:61, mean age 48.9 years [range 18–80 years]), primary SS (n = 56; male:female ratio 1:13, mean age 57.9 years [range 23–83 years]), SSc (n = 58; male:female ratio 19:39, mean age 56.7 years [range 21–74 years]), dermatomyositis (DM) (n = 29; male:female ratio 12:17, mean age 52 years [range 25–82 years]), polymyositis (PM) (n = 18; male:female ratio 1:5, mean age 64 years [range 49–83 years]), RA (n = 47; male:female ratio 13:34, mean age 59.6 years [range 35–82 years]), and MCTD (n = 19; male:female ratio 5:14, mean age 48.4 years [range 20–72 years]). Patients with chronic fatigue syndrome (n = 106; male:female ratio 10:43, mean age 40 years [range 18–75 years]) served as controls. Among this group, patients with organic disease (such as connective tissue diseases) were excluded.

The second group of samples were used for enzyme-linked immunosorbent assay (ELISA). This group consisted of samples obtained at the time of diagnosis from patients with SLE (n = 62; male:female ratio 11:51, mean age 42.7 years [range 15–73 years]), primary SS (n = 34; male:female ratio 5:29, mean age 49.9 years [range 21–73 years]), SSc (n = 65; male:female ratio 17:48, mean age 51.4 years [range 17–75 years]), DM (n = 22; male:female ratio 9:13, mean age 48.7 years [range 24–77 years]), PM (n = 11; male:female ratio 2:9, mean age 59.5 years [range 43–78 years]), RA (n = 23; male:female ratio 7:16, mean age 51.6 years [range 26–74 years]), and MCTD (n = 11; male:female ratio 3:8, mean age 40.4 years [range 16–66 years]). Patients with chronic fatigue syndrome (n = 89; male:female ratio 14:75, mean age 40.7 years [range 20–72 years]) served as controls. Among this group, patients with organic disease (such as connective tissue diseases) were excluded. Serum samples from blood donors (n = 82; male:female ratio 42:40, mean age 45 years [range 22–61 years]), patients with spondylarthritis (n = 54; male:female ratio 28:26, mean age 48 years [range 17–73 years]), patients with Crohn's disease (n = 55; male:female ratio 26:29, mean age 44 years [range 21–72 years]), and patients with ulcerative colitis (n = 70; male:female ratio 38:32, mean age 44 years [range 20–83 years]) were also included.

All patients with SS had disease characteristics that conformed with the American-European consensus classification criteria (15). Patients with SLE, patients with SSc, and patients with RA met the respective classification criteria of the American College of Rheumatology (16–18). Patients with PM and those with DM met the criteria of Bohan and Peter (19), and patients with MCTD met the criteria described by Alarcón-Segovia and Cardiel (20). The serum samples that were used for this study were obtained from the serum data bank. Samples were obtained from patients as part of routine screening for autoantibodies in the clinical laboratory. There was no informed consent for this study, but the study was approved by the local ethics committee.

Preparation of recombinant hnRNPs.

Complementary DNA (cDNA) of hnRNPs A1, H1, F, and I (obtained from Douglas Black, PhD, Howard Hughes Medical Institute, Los Angeles, CA), hnRNP P2 (obtained from T. Takumi, MD, PhD, Osaka Bioscience Institute, Osaka, Japan), hnRNPs E1 and K (obtained from Kent Duncan, PhD, University of Berlin, Berlin, Germany), hnRNP G (obtained from S. H. Chiou, MD, PhD, National Chung Hsing University, Taichung, Taiwan), and hnRNP C1 (obtained from Z. F. Chang, PhD, National Taiwan University, Taipei, Taiwan) was used.

Complementary DNA of hnRNP B1 was constructed by isolating RNA from HeLa cells using the RNeasy Mini Kit (Qiagen). RNA was converted to cDNA using the RevertAid H Minus First Strand cDNA Synthesis Kit (Fermentas).

Cloning of the polymerase chain reaction (PCR) products was performed using recombination-based Gateway technology (Invitrogen) according to the supplier's instructions. The PCR product was first cloned in the plasmid pDONR-221 using a BP Clonase reaction and subsequently in the expression plasmid pDEST-17 by a LR Clonase reaction. The gene was extended with a hexahistidine (6× His) coding sequence at its 5′ end upon cloning in the pDEST-17 vector. Thereafter, expression was induced in Escherichia coli BL21-AI with 0.1 mM IPTG and 0.1% L-arabinose for 4 hours at 37°C, and the protein was purified from inclusion bodies using a nickel–nitrilotriacetic acid affinity column (additional information is available from the corresponding author).

Protein identification by matrix-assisted laser desorption ionization–time-of-flight/time-of-flight (MALDI-TOF/TOF) technology.

Gel pieces containing the protein of interest were washed with high-performance liquid chromatography (HPLC)–grade water, dried in a SpeedVac (Savant), and digested overnight at 37°C with 10 μl of 25 mg/liter trypsin (sequence grade) in 200 mmoles/liter ammonium bicarbonate. The resulting peptide mixture was subjected to a ZipTip C18 cleanup and analyzed by MALDI-TOF/TOF (Applied Biosystems 4800 Proteomics Analyzer) in the presence of α-cyano-4-hydrocinnamic acid (HPLC grade).

Gel electrophoresis and Western blotting.

Samples were separated on 12.5% 1D sodium dodecyl sulfate–agarose gels. The proteins were either subjected to Western blotting or stained by Coomassie brilliant blue. For Western blot analysis, PVDF membranes with electrotransferred proteins (Hybond P and NovaBlot apparatus; GE Healthcare) were consecutively treated with 5% (weight/volume) bovine serum albumin, human serum (1:500; overnight), goat anti-human IgG (1:5,000), and horseradish peroxidase–conjugated rabbit anti-goat IgG (1:5,000) with intermittent washings in Tris buffered saline (3 × 10 minutes). Protein–antibody interactions were visualized by use of 0.7 mmoles/liter 3,3′-diaminobenzidine tetrahydrochloride. Blotting signals were scanned with a CanoScan LiDE 25 scanner (Canon) and quantified by ImageJ software (Image Processing and Analysis in Java). We included a positive and a negative control serum sample in each run. The coefficient of variation of the positive control (for all proteins) was <25% over 103 runs.

ELISA.

Immobilizer Amino plates (Nunc) were coated with 1 μg/ml hnRNP B1, E1, F, or H1 (100 μl/well) for 2 hours at room temperature. After washing the plates 4 times with 1× phosphate buffered saline (PBS)–Tween 20 (0.05%), 50 μl of serum sample was added (dilution 1:50 in PBS). The plates were incubated at room temperature for 30 minutes. Thereafter, 100 μl of horseradish peroxidase–conjugated goat anti-human IgG (diluted 1:5,000 in PBS) was added to each well, and the plates were incubated at room temperature for 30 minutes. Between each step, the plates were washed 4 times with 1× PBS–Tween 20 (0.05%). The color reaction was developed using 3,3′,5,5′-tetramethylbenzidine as substrate (100 μl/well).

The wells were incubated in the dark at room temperature for 15 minutes. This enzymatic coloration was stopped by the addition of 100 μl 0.5M H2SO4, and the optical density was measured at 450 nm. A positive sample was used as standard. To this standard (6 bi-fold serial dilutions ranging from 1/50 to 1/1,600), 100 arbitrary units were assigned. A 4-parameter curve was fitted to the 2-fold dilutions of the standard serum, using SoftMax Pro 5.2 software (Molecular Devices) on an EMax microplate reader (Molecular Devices). For each sample, arbitrary values were assigned based on the fitted standard curve. In each run, a positive control and a negative control were included. The optical density of the samples was measured. Arbitrary units were calculated based on a standard curve. The coefficient of variation of the positive control (for all antigens) was <13% over 12 runs.

Statistical analysis.

Linear regression analysis after log transformation (with correction for multiple comparisons), Kruskal-Wallis analysis (with Bonferroni correction), and Fisher's exact test analysis were performed by using Analyse-it for Microsoft Excel (version 2.09). Continuous data were analyzed by Kruskal-Wallis analysis. Categorical data were analyzed using Fisher's exact test, in which, for each antigen, the proportion of patients positive for antibodies was compared with the proportion of controls positive for antibodies. The P values in the tables were corrected for the numbers of proteins tested (multiple comparisons). Logistic regression (method = ENTER) on categorical data was performed using SPSS software (version 17.0).

Clustering analysis was performed by Random Forests analysis in the R program. To identify which hnRNPs best separated the disorders, Random Forests analysis was performed. Random Forests is a method that grows many classification trees or regression trees. Every tree in the forest is different: first, at each node, a best split is chosen from a random subset of the predictors rather than all of them. Second, every tree is built using a bootstrap sample of the observations. The out-of-bag data, approximately one-third of the observations, are then used to estimate the prediction accuracy. Random Forests analysis therefore provides an unbiased estimate of the classification error as the forest is built.

RESULTS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES

We evaluated reactivity with various members of the hnRNP family (A1, B1, C1, E1, F, H1, Gi, I, K, and P2) by Western blotting in serum samples obtained from control subjects (n = 106) and patients with a connective tissue disease (n = 298) (see Patients and Methods). Figure 1 shows the reactivity with 4 hnRNPs in controls, patients with primary SS, and patients with a connective tissue disease other than primary SS. For all antigens, the highest reactivity was observed in samples from patients with primary SS. For all antigens except Gi, reactivity was statistically significantly higher in samples from patients with primary SS compared with samples from controls and patients with connective tissue diseases other than primary SS (P < 0.03 by Kruskal-Wallis test with Bonferroni correction for multiple comparisons, and with linear regression after log transformation).

thumbnail image

Figure 1. Serum samples obtained from control (Contr) patients with chronic fatigue syndrome (n = 106), patients with primary Sjögren's syndrome (SS) (n = 56), and patients with various autoimmune connective tissue diseases (CTDs) (systemic sclerosis [n = 58], dermatomyositis [n = 29], mixed connective tissue disease [n = 19], polymyositis [n=18], systemic lupus erythematosus [n = 71], and rheumatoid arthritis [n = 47]) were tested for reactivity with recombinant human heterogeneous nuclear RNPs B1, E1, F, and H1 by Western blot analysis. The peak (×) area values were measured by densitometry. In addition to individual results, data are shown as box plots, where the boxes represent the 25th to 75th percentiles, the lines within the boxes represent the median, and the notched sections represent the confidence intervals. Whiskers extend to the maximum within ±1.5 interquartile ranges (IQRs) of the first or third quartile. + = observations outside 1.5 IQRs; ∗ = observations outside 3.0 IQRs.

Download figure to PowerPoint

The reactivity in connective tissue diseases other than primary SS was higher than the reactivity in controls for all antigens except hnRNP Gi, hnRNP K, and hnRNP P2 (P < 0.04 by Kruskal-Wallis test with Bonferroni correction for multiple comparisons, and with linear regression after log transformation). Random Forests analysis (cluster analysis) using continuous values revealed that antibodies to hnRNP F, hnRNP H, and hnRNP E contributed most as predictors of disease (revealed by Random Forests automated backward selection).

The prevalence of antibodies to the various hnRNPs was calculated for the different pathologies, using a cutoff that corresponded to a specificity of 97.5% in controls (patients with chronic fatigue syndrome). The data are summarized in Table 1. Analysis using Fisher's exact test (with correction for multiple comparisons) revealed that antibodies to all antigens except hnRNP Gi were significantly more prevalent in SS samples than in controls. Antibodies to hnRNPs B1, E1, F, and H1 were observed in >45% of patients with SS. Antibodies to hnRNPs A1, B1, and P2 were significantly more prevalent in MCTD samples than in controls, and antibodies to hnRNPs A1, B1, and F were significantly more prevalent in DM samples than in controls.

Table 1. Prevalence of antibodies to hnRNPs A1, B1, C1, E1, F, Gi, H1, I, K, and P2, as measured by Western blotting*
 CFS controls (n = 106)Patients with connective tissue diseases
SSc (n = 58)PM (n = 18)DM (n = 29)RA (n = 47)SLE (n = 71)MCTD (n = 19)SS (n = 56)
  • *

    Except where indicated otherwise, values are the number (%). A cutoff corresponding to a specificity of 97.5% in control patients with chronic fatigue syndrome (CFS) was used. hnRNPs = heterogeneous nuclear RNPs; SSc = systemic sclerosis; PM = polymyositis; DM = dermatomyositis; RA = rheumatoid arthritis; SLE = systemic lupus erythematosus; MCTD = mixed connective tissue disease; SS = Sjögren's syndrome.

  • P < 0.05 versus controls, by Fisher's exact test (corrected for multiple comparisons).

  • P < 0.05 versus controls, by Fisher's exact test (corrected for multiple comparisons) and by logistic regression analysis.

  • §

    P < 0.05 versus controls, by logistic regression analysis..

hnRNP A13 (2.8)8 (13.8)2 (11.1)8 (27.6)6 (12.8)10 (14.1)5 (26.3)21 (37.5)
hnRNP B13 (2.8)9 (15.5)2 (11.1)10 (34.5)8 (17.0)8 (11.3)5 (26.3)28 (50.0)
hnRNP C13 (2.8)7 (12.1)1 (5.6)5 (17.2)4 (8.5)8 (11.3)4 (21.1)15 (26.8)
hnRNP E13 (2.8)4 (6.9)1 (5.6)5 (17.2)4 (8.5)9 (12.7)3 (15.8)28 (50.0)
hnRNP F3 (2.8)5 (8.6)2 (11.1)8 (27.6)8 (17.0)§9 (12.7)4 (21.1)28 (50.0)
hnRNP Gi3 (2.8)5 (8.6)1 (5.6)2 (6.9)1 (2.1)4 (5.6)1 (5.3)3 (5.4)
hnRNP H13 (2.8)2 (3.4)2 (11.1)5 (17.2)3 (6.4)5 (7.0)3 (15.8)26 (46.4)
hnRNP I3 (2.8)3 (5.2)2 (11.1)1 (3.4)2 (4.3)2 (2.8)1 (5.3)12 (21.4)
hnRNP K3 (2.8)3 (5.2)1 (5.6)5 (17.2)4 (8.5)5 (7.0)1 (5.3)14 (25.0)
hnRNP P23 (2.8)7 (12.1)2 (11.1)3 (10.3)6 (12.8)12 (16.9)6 (31.6)15 (26.8)

Logistic regression analysis confirmed the following: 1) that antibodies to hnRNPs B1 and F were associated with SS, 2) that anti–hnRNP B1 antibodies were associated with DM, and 3) that anti–hnRNP P2 antibodies were associated with MCTD (Table 1). Random Forests analysis (automated backward selection) using binary values revealed that antibodies to hnRNPs F, B1, E1, and H1 contributed as predictors of disease.

Reactivity with at least 1 of the 10 antigens was observed in 19% of controls and in 45% of DM, 49% of RA, 38% of SLE, 53% of MCTD, and 73% of SS samples. Reactivity with at least 3 antigens was observed in 2% of controls, 14% of SSc, 28% of DM, 14% of SLE, 26% of MCTD, and 48% of SS samples. Reactivity with at least 5 antigens was observed in 0% of controls and in 7% of SSc, 14% of DM, 7% of SLE, 16% of MCTD, and 32% of SS samples (Table 2).

Table 2. Reactivity with antibodies to hnRNPs A1, B1, C1, E1, F, H1, Gi, I, K, and P2, as determined by Western blotting*
 CFS controls (n = 106)Patients with connective tissue diseases
SSc (n = 58)PM (n = 18)DM (n = 29)RA (n = 47)SLE (n = 71)MCTD (n = 19)SS (n = 56)
  • *

    Except where indicated otherwise, values are the number (%). A cutoff corresponding to a specificity of 97.5% in control patients with chronic

  • fatigue syndrome (CFS) was used. hnRNPs = heterogeneous nuclear RNPs; SSc = systemic sclerosis; PM = polymyositis; DM = dermatomyositis;

  • RA = rheumatoid arthritis; SLE = systemic lupus erythematosus; MCTD = mixed connective tissue disease; SS = Sjögren's syndrome.

  • P < 0.05 versus controls, by Fisher's exact test (corrected for multiple comparisons).

Reactivity with ≥1 antigens20 (18.9)15 (25.9)6 (33.3)13 (44.8)23 (48.9)27 (38.0)10 (52.6)41 (73.2)
Reactivity with ≥2 antigens6 (5.7)10 (17.2)3 (16.7)11 (37.9)11 (23.4)15 (21.1)6 (31.6)32 (57.1)
Reactivity with ≥3 antigens2 (1.9)8 (13.8)2 (11.1)8 (27.6)5 (10.6)10 (14.1)5 (26.3)27 (48.2)
Reactivity with ≥4 antigens1 (0.9)6 (10.3)1 (5.6)6 (20.7)4 (8.5)8 (11.3)5 (26.3)23 (41.1)
Reactivity with ≥5 antigens0 (0.0)4 (6.9)1 (5.6)4 (13.8)1 (2.1)5 (7.0)3 (15.8)18 (32.1)
Reactivity with ≥6 antigens0 (0.0)4 (6.9)1 (5.6)3 (10.3)1 (2.1)2 (2.8)2 (10.5)16 (28.6)
Reactivity with ≥7 antigens0 (0.0)1 (1.7)1 (5.6)3 (10.3)0 (0.0)1 (1.4)1 (5.3)13 (23.2)

In a next step, ELISAs were used to evaluate reactivity with hnRNPs B1, E1, F, and H1 in serum samples from controls (patients with chronic fatigue syndrome; n = 89), blood donors (n = 82), and in diagnostic samples from patients with a connective tissue disease (n = 228) (see Patients and Methods). Reactivity was also measured in inflammation control samples (obtained from 54 patients with spondylarthritis, 55 patients with Crohn's disease, and 70 patients with ulcerative colitis). Heterogeneous nuclear RNPs B1, E1, F, and H1 were selected because they displayed high reactivity for SS by Western blotting, and because Random Forests analysis revealed that these 4 antibodies contributed most to disease prediction. Figure 2 shows reactivity with 2 of these antigens (hnRNP B1 and hnRNP F) in different patient groups and controls. Differences in reactivity between controls and patients were evaluated by Kruskal-Wallis analysis with Bonferroni correction and by linear regression analysis after log transformation. For all 4 antigens, significantly higher antibody levels were observed in patients with primary SS, patients with SLE, and patients with RA compared with control patients with chronic fatigue syndrome (P < 0.01). For hnRNPs H1, F, and B1, higher antibody levels were also observed in MCTD compared with controls (P < 0.001). For hnRNP B1, significantly higher antibody levels were also observed in SSc compared with controls (P < 0.0001).

thumbnail image

Figure 2. Serum samples obtained from control patients with chronic fatigue syndrome (CFS; n = 89), blood donors (BD; n = 82), patients with Crohn's disease (CD; n = 55), patients with ulcerative colitis (UC; n = 70), and patients with spondylarthritides (Spond; n = 54), and diagnostic serum samples obtained from patients with various connective tissue diseases (systemic sclerosis [SSc; n = 65], polymyositis [PM; n = 11], dermatomyositis [DM; n = 22], rheumatoid arthritis [RA; n = 23], systemic lupus erythematosus [SLE; n = 62], mixed connective tissue disease [MCTD; n = 11], and Sjögren's syndrome [n = 34]) were tested for reactivity with recombinant human hnRNP B1 and hnRNP F by enzyme-linked immunosorbent assay. In addition to individual results, data are shown as box plots, where the boxes represent the 25th to 75th percentiles, the lines within the boxes represent the median, and the notched sections represent the confidence intervals. Whiskers extend to the maximum within ±1.5 interquartile ranges (IQRs) of the first or third quartile. + = observations outside 1.5 IQRs; ∗ = observations outside 3.0 IQRs. See Figure 1 for other definitions.

Download figure to PowerPoint

Significantly lower antibody levels were observed in samples obtained from blood donors compared with controls (P < 0.0001). Kruskal-Wallis testing with Bonferroni correction and linear regression analysis after log transformation revealed no difference between reactivity in chronic fatigue syndrome controls and that in patients with spondylarthritis (except for reactivity with hnRNP E1, for which there was decreased reactivity in patients with spondylarthritis compared with chronic fatigue syndrome controls). Kruskal-Wallis testing with Bonferroni correction and linear regression analysis after log transformation revealed no difference between reactivity in chronic fatigue syndrome controls and reactivity in patients with Crohn's disease or ulcerative colitis, except for reactivity with hnRNP H1 and hnRNP F, for which there was increased reactivity in patients with inflammatory bowel disease compared with chronic fatigue syndrome controls. Results of Random Forests analysis (automated backward selection) on continuous values confirmed that antibodies to hnRNP B contributed as a predictor of disease.

The prevalence of antibodies to hnRNPs B1, E1, F, and H1 was calculated in serum samples obtained from patients with various connective tissue diseases, using a cutoff that corresponded to a specificity of 97.5% in chronic fatigue syndrome controls. The data are summarized in Table 3. The prevalence of antibodies to hnRNP B1 was significantly higher in primary SS (44%), MCTD (45%), and SLE (37%) compared with controls (3.4%) (both by logistic regression analysis and by Fisher's exact test analysis after correction for multiple comparisons). The prevalence of antibodies to hnRNP F was significantly higher in primary SS (21%) than in chronic fatigue syndrome controls (3.4%) (both by logistic regression analysis and by Fisher's exact analysis after correction for multiple comparisons). There was no statistically significant difference between the reactivity observed in chronic fatigue syndrome controls and the reactivity in patients with spondylarthritis, Crohn's disease, or ulcerative colitis (by Fisher's exact test after correction for multiple comparisons and logistic regression). Random Forests analysis on binary values confirmed that antibodies to hnRNP B and hnRNP F contributed most as predictors of disease.

Table 3. Prevalence of antibodies to hnRNPs B1, E1, F, and H1, as determined by ELISA*
 CFS controls (n = 89)BD (n = 82)CD (n = 55)UC (n = 70)Patients with connective tissue diseases
SpA (n = 54)DM (n = 22)MCTD (n = 11)PM (n = 11)RA (n = 23)SLE (n = 62)SSc (n = 65)SS (n = 34)
  • *

    Except where indicated otherwise, values are the number (%). A cutoff corresponding to a specificity of 97.5% in control patients with chronic fatigue syndrome (CFS) was used. hnRNPs = heterogeneous nuclear RNPs; ELISA = enzyme-linked immunosorbent assay; BD = blood donor; CD = Crohn's disease; UC = ulcerative colitis; SpA = spondylarthritis; DM = dermatomyositis; MCTD = mixed connective tissue disease; PM = polymyositis; RA = rheumatoid arthritis; SLE = systemic lupus erythematosus; SSc = systemic sclerosis; SS = Sjögren's syndrome.

  • P < 0.05 versus controls, by logistic regression analysis.

  • P < 0.05 versus controls, by Fisher's exact test (corrected for multiple comparisons) and by logistic regression analysis.

  • §

    P < 0.05 versus controls, by Fisher's exact test (corrected for multiple comparisons).

hnRNP B13 (3.4)0 (0.0)0 (0.0)0 (0.0)1 (1.9)5 (22.7)5 (45.5)1 (9.1)5 (21.7)23 (37.1)10 (15.4)15 (44.1)
hnRNP E13 (3.4)0 (0.0)0 (0.0)1 (1.4)1 (1.9)1 (4.5)0 (0.0)1 (9.1)1 (4.3)3 (4.8)2 (3.1)7 (20.6)§
hnRNP F3 (3.4)1 (1.2)4 (7.3)2 (2.9)0 (0.0)1 (4.5)1 (9.1)1 (9.1)3 (13.0)9 (14.5)1 (1.5)7 (20.6)
hnRNP H12 (2.2)1 (1.2)3 (5.5)2 (2.9)0 (0.0)0 (0.0)1 (9.1)0 (0.0)0 (0.0)6 (9.7)1 (1.5)4 (11.8)

Two of 62 patients with SLE had secondary SS. One of these patients had no antibodies to any of the 4 hnRNPs tested by ELISA, and the other patient had elevated antibodies to hnRNP B1. One of 11 patients with MCTD had secondary SS; this patient had elevated levels of antibodies to hnRNP F.

Reactivity with at least 1 of the 4 tested antigens was observed in 11% of the controls and in 35% of RA, 47% of SLE, 64% of MCTD, and 56% of primary SS samples (Table 4). Reactivity with at least 2 of the 4 tested antigens was observed in 1.1% of controls and in 16% of SLE samples and 18% of primary SS samples (Table 4). Reactivity with at least 3 of the 4 antigens was observed in 0% of controls and in 15% of patients with SS (Table 4).

Table 4. Reactivity with hnRNPs B1, E1, F, and H1, as determined by ELISA*
 CFS controls (n = 89)BD (n = 82)CD (n = 55)UC (n = 70)Patients with connective tissue diseases
SpA (n = 54)DM (n = 22)MCTD (n = 11)PM (n = 11)RA (n = 23)SLE (n = 62)SSc (n = 65)SS (n = 34)
  • *

    Except where indicated otherwise, values are the number (%). A cutoff corresponding to a specificity of 97.5% in control patients with chronic fatigue syndrome (CFS) was used. hnRNPs = heterogeneous nuclear RNPs; ELISA = enzyme-linked immunosorbent assay; BD = blood donor; CD = Crohn's disease; UC = ulcerative colitis; SpA = spondylarthritis; DM = dermatomyositis; MCTD = mixed connective tissue disease; PM = polymyositis; RA = rheumatoid arthritis; SLE = systemic lupus erythematosus; SSc = systemic sclerosis; SS = Sjögren's syndrome.

  • P < 0.05 versus controls, by Fisher's exact test (corrected for multiple comparisons).

Reactivity with ≥1 antigens10 (11.2)1 (1.2)6 (10.9)4 (5.7)2 (3.7)4 (18.2)7 (63.6)3 (27.3)8 (34.8)29 (46.8)12 (18.5)19 (55.9)
Reactivity with ≥2 antigens1 (1.1)1 (1.2)1 (1.8)1 (1.4)0 (0)1 (4.5)0 (0.0)0 (0.0)1 (4.3)10 (16.1)2 (3.1)6 (17.6)
Reactivity with ≥3 antigens0 (0.0)0 (0.0)0 (0)0 (0)0 (0)1 (4.5)0 (0.0)0 (0.0)0 (0.0)2 (3.2)0 (0.0)5 (14.7)
Reactivity with ≥4 antigens0 (0.0)0 (0.0)0 (0)0 (0)0 (0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)3 (8.8)

DISCUSSION

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES

We aimed to evaluate the many autoantibodies to hnRNPs in patients with systemic rheumatic diseases. Therefore, we tested the reactivity with various members of the hnRNP family. Using Western blotting, we confirmed previous observations that patients with SLE, patients with RA, and patients with MCTD have higher reactivity with hnRNP B1 compared with control subjects (6). We also confirmed our previous observation (14) that reactivity with hnRNP H1 is higher in patients with primary SS compared with controls. Patients with primary SS displayed reactivity with 9 of the 10 hnRNPs tested. The highest reactivity was observed for hnRNPs B1, E1, F, and H1. Cluster analysis (Random Forests analysis) confirmed that antibodies to these antigens contributed most as predictors of disease.

The simultaneous presence of antibodies to several hnRNPs was strongly associated with primary SS and with some other systemic rheumatic diseases. For example, antibodies to at least 5 of the 10 antigens tested were observed in none of the controls but were observed in 32% of primary SS, 16% of MCTD, 7% of SLE, 14% of DM, and 7% of SSc samples. The simultaneous presence of antibodies to at least 7 of the 10 antigens was observed in 23% of SS samples and in 10% of DM samples.

The high reactivity with hnRNPs B1, E1, and F in SS was confirmed by ELISA. However, ELISA analysis did not reveal significantly higher reactivity with hnRNP H1 in SS patients compared with controls. This might be related to differences in conformation of the protein between ELISA analysis (protein in native conformation) and Western blot analysis (denatured protein). Antibodies to hnRNP B1 were observed in serum samples obtained from SS patients and in samples obtained from patients with several other connective tissue diseases, including MCTD and SLE. ELISA analysis further confirmed the notion that the simultaneous presence of antibodies to several hnRNPs was strongly associated with connective tissue disease in general and with primary SS in particular. For example, reactivity with at least 2 of the 4 tested antigens was observed in 1.1% of controls, in 16% of patients with SLE, and in 18% of patients with primary SS. Reactivity with at least 3 of the 4 antigens was observed in none of the controls and in 3.2% of patients with SLE and 15% of patients with primary SS.

ELISAs were performed on diagnostic samples (i.e., samples that were obtained at the time of diagnosis). Preliminary data show that some patients test positive for anti-hnRNP antibodies and negative for antibodies to the classic extractable nuclear antigens (SSA, SSB, Scl-70, CENP-B, Sm, U1 RNP, RNP 70, Jo-1) or double-stranded DNA (dsDNA). For example, of the 34 samples from patients with SS, 19 tested positive for at least 1 of the 4 antigens tested (hnRNPs B1, E1, F, and H1). Two of these patients were negative for anti-SSA or anti-SSB antibodies (Op De Beéck K, et al: unpublished observations). We also identified diagnostic samples from patients with SLE and from those with SSc that were negative for the classic antibodies to extractable nuclear antigens and to dsDNA but positive for antibodies to hnRNPs (Op De Beéck K, et al: unpublished observations).

Among the 10 hnRNPs studied, the highest identity was observed between hnRNPs A1 and B1 and between hnRNPs F and H1. After manual alignment (www.ncbi.nlm.nih.gov), hnRNPs A1 and B1 showed 60.5% identity. Both proteins have 2 moderate homologous RNA recognition motifs and a glycine-rich region. Heterogeneous nuclear RNPs H1 and F have an identity of 78%. Even though the proteins are closely sequence-related, they exhibit differences, especially in their COOH terminus, where hnRNP F is shorter and contains a deletion when compared with hnRNP H1 (21). These differences occur outside the RNA recognition motif regions (21). Matunis et al (22) reported that monoclonal antibodies produced against bacterially expressed hnRNP F were specific for both hnRNP F and hnRNP H1. Therefore, hnRNPs F and H1 are highly related immunologically. Principal components analysis of our ELISA data demonstrated a correlation between anti-hnRNP H1 antibodies and anti-hnRNP F antibodies.

Solid-phase assays are increasingly replacing indirect immunofluorescence testing for antinuclear antibody detection. Heterogeneous nuclear RNPs could be included as antigens in multiparameter diagnostic assays (e.g., multiplexed bead assays, microarrays) in order to increase the sensitivity of solid-phase assays for antinuclear antibodies.

In conclusion, several hnRNPs are target antigens in SS. The combined presence of antibodies to several hnRNPs was strongly associated with connective tissue disease in general and with SS in particular. Detection of antibodies to hnRNPs may become a novel diagnostic approach for connective tissue disease in general and SS in particular.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Bossuyt had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Op De Beéck, Maes, Van den Bergh, Hooijkaas, Bossuyt.

Acquisition of data. Op De Beéck, Maes, Van den Bergh, Derua, Waelkens, Vermeersch, Westhovens, De Vlam, Verschueren, Hooijkaas, Blockmans, Bossuyt.

Analysis and interpretation of data. Op De Beéck, Van den Bergh, Waelkens, Van Steen, Vermeersch, Westhovens, Verschueren, Hooijkaas, Bossuyt.

REFERENCES

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES
  • 1
    Dreyfuss G, Matunis MJ, Pinol-Roma S, Burd CG. hnRNP proteins and the biogenesis of mRNA. Annu Rev Biochem 1993; 62: 289321.
  • 2
    Krecic AM, Swanson MS. hnRNP complexes: composition, structure, and function. Curr Opin Cell Biol 1999; 11: 36371.
  • 3
    Carpenter B, MacKay C, Alnabulsi A, MacKay M, Telfer C, Melvin WT, et al. The roles of heterogeneous nuclear ribonucleoproteins in tumour development and progression. Biochim Biophys Acta 2006; 1765: 85100.
  • 4
    Kamma H, Portman DS, Dreyfuss G. Cell type-specific expression of hnRNP proteins. Exp Cell Res 1995; 221: 18796.
  • 5
    Caporali R, Bugatti S, Bruschi E, Cavagna L, Montecucco C. Autoantibodies to heterogeneous nuclear ribonucleoproteins. Autoimmunity 2005; 38: 2532.
  • 6
    Steiner G, Skriner K, Hassfeld W, Smolen JS. Clinical and immunological aspects of autoantibodies to RA33/hnRNP-A/B proteins: a link between RA, SLE and MCTD [review]. Mol Biol Rep 1996; 23: 16771.
  • 7
    Jensen L, Kuff EL, Wilson SH, Steinberg AD, Klinman DM. Antibodies from patients and mice with autoimmune diseases react with recombinant hnRNP core protein A1. J Autoimmun 1988; 1: 7383.
  • 8
    Siapka S, Patrinou-Georgoula M, Vlachoyiannopoulos PG, Guialis A. Multiple specificities of autoantibodies against hnRNP A/B proteins in systemic rheumatic diseases and hnRNP L as an associated novel autoantigen. Autoimmunity 2007; 40: 22333.
  • 9
    Montecucco C, Caporali R, Cobianchi F, Biamonti G. Identification of autoantibodies to the I protein of the heterogeneous nuclear ribonucleoprotein complex in patients with systemic sclerosis. Arthritis Rheum 1996; 39: 166976.
  • 10
    Heegaard NH, Larsen MR, Muncrief T, Wiik A, Roepstorff P. Heterogeneous nuclear ribonucleoproteins C1/C2 identified as autoantigens by biochemical and mass spectrometric methods. Arthritis Res 2000; 2: 40714.
  • 11
    Hassfeld W, Chan EK, Mathison DA, Portman D, Dreyfuss G, Steiner G, et al. Molecular definition of heterogeneous nuclear ribonucleoprotein R (hnRNP R) using autoimmune antibody: immunological relationship with hnRNP P. Nucleic Acids Res 1998; 26: 43945.
  • 12
    Skriner K, Hueber W, Suleymanoglu E, Hofler E, Krenn V, Smolen J, et al. AUF1, the regulator of tumor necrosis factor α messenger RNA decay, is targeted by autoantibodies of patients with systemic rheumatic diseases. Arthritis Rheum 2008; 58: 51120.
  • 13
    Lin TY, Chan LC, Fan YH, Lin CH, Chow KC, Lin SL, et al. Use of a recombinant protein containing major epitopes of hnRNP G to detect anti-hnRNP G antibodies in dogs with systemic lupus erythematosus. Res Vet Sci 2006; 81: 3359.
  • 14
    Van den Bergh K, Hooijkaas H, Blockmans D, Westhovens R, Op De Beeck K, Verschueren P, et al. Heterogeneous nuclear ribonucleoprotein H1, a novel nuclear autoantigen. Clin Chem 2009; 55: 94654.
  • 15
    Vitali C, Bombardieri S, Jonsson R, Moutsopoulos HM, Alexander EL, Carsons SE, et al, and the European Study Group on Classification Criteria for Sjögren's Syndrome. Classification criteria for Sjögren's syndrome: a revised version of the European criteria proposed by the American-European Consensus Group. Ann Rheum Dis 2002; 61: 5548.
  • 16
    Hochberg MC, for the Diagnostic and Therapeutic Criteria Committee of the American College of Rheumatology. Updating the American College of Rheumatology revised criteria for the classification of systemic lupus erythematosus [letter]. Arthritis Rheum 1997; 40: 1725.
  • 17
    Subcommittee for Scleroderma Criteria of the American Rheumatism Association Diagnostic and Therapeutic Criteria Committee. Preliminary criteria for the classification of systemic sclerosis (scleroderma). Arthritis Rheum 1980; 23: 58190.
  • 18
    American College of Rheumatology Subcommittee on Rheumatoid Arthritis Guidelines. Guidelines for the management of rheumatoid arthritis: 2002 update. Arthritis Rheum 2002; 46: 32846.
  • 19
    Bohan A, Peter JB, Bowman RL, Pearson CM. Computer-assisted analysis of 153 patients with polymyositis and dermatomyositis. Medicine (Baltimore) 1977; 56: 25586.
  • 20
    Alarcon-Segovia D, Cardiel MH. Comparison between 3 diagnostic criteria for mixed connective tissue disease: study of 593 patients. J Rheumatol 1989; 16: 32834.
  • 21
    Honore B, Rasmussen HH, Vorum H, Dejgaard K, Liu X, Gromov P, et al. Heterogeneous nuclear ribonucleoproteins H, H′, and F are members of a ubiquitously expressed subfamily of related but distinct proteins encoded by genes mapping to different chromosomes. J Biol Chem 1995; 270: 287809.
  • 22
    Matunis MJ, Xing J, Dreyfuss G. The hnRNP F protein: unique primary structure, nucleic acid-binding properties, and subcellular localization. Nucleic Acids Res 1994; 22: 105967.