Anti-Fibrillarin Antibody in African American Patients with Systemic Sclerosis: Immunogenetics, Clinical Features, and Survival Analysis

Study population

Between 1985 and 2010, 3033 patients with SSc were enrolled in the following cohorts: (1) the Genetics versus ENvironment In Scleroderma Outcomes Study (GENISOS)^3,5,18; (2) the NIH/NIAMS Scleroderma Family Registry and DNA Repository¹⁹; and (3) the Division of Rheumatology, University of Texas Health Science Center at Houston (UTHSC-H)¹⁰. Patients were included if they met the American College of Rheumatology (formerly American Rheumatism Association) classification criteria for SSc²⁰ or had at least 3 of the 5 CREST features (calcinosis, Raynaud’s phenomenon, esophageal dysmotility, sclerodactyly, telangiectasias)²¹. We included all AA patients from these cohorts (n = 278). Patients enrolled in more than one of the cohorts were identified and duplicate entries were omitted. We enrolled 328 unaffected AA controls to determine any HLA class II allele associations with AFA. The unaffected AA individuals were volunteers with no personal or family history of SSc or other autoimmune disease by screening questionnaire. All study subjects enrolled (SSc patients and unaffected controls) provided written informed consent, and the institutional review board of all participating institutions approved the study.

Autoantibody profile and HLA class II allele genotyping

All autoantibody determinations and HLA class II allele typing were conducted in the Division of Rheumatology at UTHSC-H and the Mitogen Advanced Diagnostics Laboratory, University of Calgary, Calgary, Canada. Antinuclear antibodies (ANA) and anticentromere antibodies were determined using indirect immunofluorescence with HEp-2 cells as substrate (Antibodies Inc., Davis, CA, USA). Passive immunodiffusion gels against calf thymus extract were used to examine sera for antitopoisomerase-I (ATA; Scl-70), anti-Ro/SSA, anti-La/SSB, and anti-U1-RNP autoantibodies (Inova Diagnostics, San Diego, CA, USA). Anti-RNA polymerase III (RNAP III) was detected by ELISA kits (MBL Co. Ltd., Nagoya, Japan) and AFA were determined by a line immunoassay at a serum dilution of 1:1000 using purified recombinant fibrillarin protein (Euroline-WB; Euroimmun, Lubeck, Germany) in patients who had a positive ANA in antinucleolar pattern on the indirect immunofluorescence.

As described^5,22, we genotyped HLA class II alleles (DRB1, DQA1, DQB1, and DPB1) on extracted and purified genomic DNA. Further, we examined the HLA class II allele-binding peptide using the ProPred MHC Class II Binding Peptide Prediction Server²³ in order to predict binding peptides of human fibrillarin protein. This prediction is based on quantitative matrices derived from the literature^23,24.

Clinical manifestation

Age, sex, disease type (categorized as limited or diffuse cutaneous involvement at time of enrollment²¹), disease duration (calculated from the onset of the first non-Raynaud’s phenomenon symptom attributable to SSc), and modified Rodnan skin score (MRSS)²⁵ were recorded.

To assess the severity of individual organ system involvement, the Medsger severity indices^13,26 of 8 organ systems were measured: peripheral vessels, skin, joints/tendons, skeletal muscle, gastrointestinal (GI) tract, lung, heart, and kidney. However, these data were available only for the patients enrolled in the GENISOS cohort (n = 78). The presence of digital ulcers was determined based on the participating rheumatologist’s clinical assessment. Arthritis was defined as presence of joint swelling and tenderness on examination not attributable to osteoarthritis, crystalline arthropathy, or trauma. A decrease in range of motion > 25% in at least one joint axis was defined as joint contracture. Dysphagia, diarrhea attributable to SSc, and history of SSc renal crisis were recorded. Electrocardiography and 2-dimensional echocardiography findings and/or presence of an auscultatory friction rub determined the presence of pericarditis or clinically significant pericardial effusion.

As described¹⁸, pulmonary function tests were obtained at enrollment. Interstitial lung fibrosis, defined as chest radiograph showing fibrosis and/or forced vital capacity (FVC) < 75% of predicted value, was recorded.

For the purpose of our review, PAH was defined if the patient had (1) mean pulmonary artery pressure ≥ 25 mm Hg on right heart catheterization; (2) right ventricular systolic pressure ≥ 40 mm Hg on 2-dimensional echocardiography; or (3) if the ratio of FVC% predicted to diffusion capacity of carbon monoxide (DLCO)% predicted was ≥ 1.6. Serum creatine kinase (CK) levels were recorded and myositis was diagnosed if the patient had proximal muscle weakness with at least one of the following: elevated levels of CK, features of myositis on electromyography, and/or a characteristic muscle biopsy.

Death search

The vital status of patients was determined through the National Death Index (NDI) at the US Centers for Disease Control and Prevention, which provided data up until 2007. We then reviewed the US Social Security Death Index (SSDI) to update our results as of August 2010. SSDI is an online death search tool that provides fatality reports based on death certificates and family confirmation. Patients not found on NDI or SSDI were assumed to be alive.

Statistical analysis

Homozygosity for alleles at each of the tested HLA loci was not suggestive of recessive inheritance, regardless of whether the referent comparison group comprised disease-free controls or AFA-negative cases. There were too few homozygous subjects to distinguish additive from dominant modes of inheritance, regardless of the referent. Therefore, a dominant mode of inheritance approach was used to compare the HLA association with AFA. Heterozygosity and homozygosity for a particular allele were both recoded as “1” in a binary (zero or 1) variable created for each specific HLA gene of interest. In other words, subjects negative for the gene on both their alleles for the particular HLA locus were coded “0” for the gene on the new binary variable. Bonferroni correction for multiple comparisons was performed for HLA allelic analyses.

Age, sex, disease type, and disease duration of AFA-positive and AFA-negative patients were evaluated utilizing chi-square and Student t test accordingly. SSc clinical manifestations might have changed over the disease course; therefore logistic regression was used to adjust for disease duration as a possible confounding factor in clinical features and to examine the independent effect of AFA.

We utilized Cox proportional hazards regression analysis to examine the association of AFA with survival. We investigated the potential association of relevant HLA class II with survival of the AA patients with SSc. Survival analysis was corrected for age at enrollment. Survival was calculated from the date of enrollment.

ATA and AFA are the 2 most common antinuclear antibodies among AA patients with SSc. We also compared the clinical features and survival of AA scleroderma patients with AFA (n = 50) to those with ATA (n = 61) for comparative analysis between more homogeneous groups.

All the statistical analyses were performed with SAS Version 9.2 (SAS Institute Inc., Cary, NC, USA) and Stata 11 (StataCorp., College Station, TX, USA). Hypothesis testing was 2-sided with a p ≤ 0.05 significance level.

Study population, disease, and autoantibody characteristics

All 278 AA scleroderma patients from the 3 cohorts were included in the study. The mean age (± SD) of patients at enrollment was 46.9 (13.9) years, and 237 (85.3%) were female. At enrollment, 171 (61.5%) AA patients with SSc were diagnosed with diffuse cutaneous involvement. Average disease duration (± SD) was 6.0 (6.5) years.

ANA on HEp-2 substrate were detected in 93.1% of AA SSc patients. ATA, RNAP-III, and AFA were present in 21.8%, 15.4%, and 18.5% of patients, respectively (Table 1).

HLA class II allelic frequencies

As illustrated in Table 2, comparison of HLA class II allelic frequencies of AFA-positive patients with 329 ethnically matched unaffected controls revealed the HLA-DRB1*08:04 allele more frequently in AFA-positive patients (47.6% vs 6.4%, respectively; OR 11.52, 95% CI 5.43, 24.40; corrected p < 0.0001). Two other alleles located on the same haplotype, DQA1*04:01 and DQB1*03:01, had similar patterns. However, the increased frequency of DQA1*04:01 was not statistically significant.

The frequency of HLA-DRB1*08:04 in AFA-positive patients also was higher in comparison to AA patients without AFA, even after correction for multiple comparisons (47.6% vs 14.9%; OR 5.21, 95% CI 2.44, 11.09; corrected p = 0.0002). Both HLA-DQA1*04:01 and DQB1*03:01 showed similar trends. However, neither of them remained significant after correction for multiple comparisons.

HLA-DPB1*01:01 was also seen more frequently among AFA-positive AA patients compared to unaffected controls and SSc patients without AFA, whereas HLA-DRB1*11:01 seemed to be protective. HLA-DPB1*01:01 and HLA-DRB1*11:01 are not in linkage disequilibrium with HLA-DRB1*08:04. However, the association of these 2 alleles with AFA did not withstand correction for multiple comparisons. The frequencies of all relevant HLA class II alleles in AA SSc patients and unaffected individuals are illustrated in Table 3.

HLA-DRB1*08:04 binding peptides

Using virtual matrix for HLA-DRB1*08:04, at a threshold of 1% (the percentage best scoring natural peptides), we identified 4 binding peptides (FRSKLAAAI, FRGRGRGGG, IHIKPGAKV, and FVISIKANC) from the human fibrillarin protein that could serve as potential binding sites within the antigen-binding groove.

Clinical features

AA SSc patients with AFA were younger at disease onset (p = 0.004) but sex, disease type, and duration were not significantly different compared to AA SSc patients with AFA. Table 4 illustrates the comparison of clinical manifestations between AFA-positive and AFA-negative AA patients with SSc.

After adjustment for disease duration, AA SSc patients with AFA were 3.31-times more likely to have digital ulcers (p = 0.014). Diarrhea and pericarditis occurred more frequently in AFA-positive AA SSc patients (OR 4.84, p = 0.019; OR 2.45, p = 0.092, respectively) than AA patients without AFA. However, there were no differences between AFA-positive and AFA-negative AA SSc patients in MRSS, dysphagia, PAH, SSc-associated interstitial lung fibrosis, FVC and DLCO predicted values, SSc renal crisis, myositis or muscle weakness, serum CK, joint contracture, or sicca symptoms.

AFA-positive patients had higher Medsger peripheral vascular severity index scores (regression coefficient [b] = 0.79, 95% CI 0.27, 1.30; p = 0.003), indicating more severe peripheral vascular involvement, and lower Medsger lung severity index (b = −0.82, 95% CI −1.50, −0.14; p = 0.019), indicating less severe lung involvement. The other Mesdger severity indices were not significantly different (Table 4).

Survival analysis

At the time of analysis, 30% of AFA-positive AA SSc patients and 29.5% of AFA-negative patients were deceased (Table 4). After correction for age at enrollment, AFA-positive patients did not have different survival compared to AFA-negative patients (hazard ratio = 0.79, p = 0.493). In addition, none of the relevant HLA class II was a predictor of mortality in AA patients with SSc (Table 5).

AFA and ATA among AA patients with SSc

Although age at onset of the first non-Raynaud’s symptom was not statistically different between these 2 groups, the AA scleroderma patients with AFA had higher frequency of digital ulcers, and lower GI tract involvement, pericarditis, and Medsger peripheral vascular severity index scores (Table 6). AFA-positive patients had lower Medsger lung severity index, higher FVC and DLCO predicted values, and fewer cases of PAH. Despite less severe lung disease, after adjusting for age of disease onset, AFA-positive patients did not have better or worse survival compared to ATA-positive (Table 5).