Anti-carbamylated protein antibodies as a new biomarker of erosive joint damage in systemic lupus erythematosus

Moreover, the presence of rheumatoid factor (RF), ACPA, and anti-CarP antibodies was investigated. All autoantibody assays were carried out in triplicate, commercial ELISA kits were used and the results were evaluated according to the manufacturers’ instructions. ACPA titers were obtained using a commercial ELISA kit (DELTA BIOLOGICALS, Rome, Italy). Values above 25 U/mL were considered positive. A solid-phase ELISA kit was used in order to determine RF (Diamedix, Miami, USA). Diluted samples are incubated with purified RF antigen (human IgG) bound to the microtiter well. Any RF-IgM antibody present in the sample binds to the immobilized human IgG to form antigen-antibody complexes. Unbound antibody is washed from the wells and horseradish peroxidase-conjugated anti-human IgM is added. Values above 10 U/mL were considered positive. Anti-CarP antibodies were detected by a “home-made” ELISA, using carbamylated fetal calf serum (Ca-FCS) and non-modified FCS as antigens. Ca-FCS was obtained using the method described by Shi et al. [25]. In brief, Nunc Maxisorp polystyrene plates (Thermo Scientific, Waltham, MA, USA) were coated overnight at 4 °C with non-modified FCS and Ca-FCS (12 μg/ml in 0.05 M Na₂CO₃/NaHCO₃ buffer, pH 9.6). After washing, plates were blocked with phosphate-buffered saline (PBS), containing 1% bovine serum albumin (BSA) (Sigma, Saint Louis, Missouri, USA) for 6 h at 4 °C. Subsequently, the wells were incubated with patients’ serum diluted 1:50 in PBS/0.05% Tween20 (PBS-T)/BSA 1% overnight at 4 °C. Alkaline phosphatase-conjugated anti-human IgG antibodies (Sigma, Saint Louis, MO, USA) were diluted in PBS-T/BSA 1% (1:1000), and incubated for 2 h at room temperature. Paranitrophenyl-phosphate was used as a substrate and the optical density (OD) was measured at 405 nm wavelength. All assays were performed in triplicate and the absorbance of control wells (non-modified FCS) was subtracted to account for non-specific binding. A titration curve of two positive serum samples with medium/high immunoreactivity for Ca-FCS was performed, to assess the performance of the tests and to transform the absorbance of Ca-FCS in arbitrary units per milliliter (aU/mL). The cutoff was established as the mean OD + 3 standard deviations (SD) of 56 age-matched and sex-matched healthy subjects (blood donors) and then the obtained value was converted into arbitrary units per milliliter (corresponding to 340 aU/mL).

The statistical analyses were performed using version 5.0 of the GraphPad statistical package. Normally distributed variables were summarized using the mean ± standard deviation (SD), and non-normally distributed variables by the median and interquartile range (IQR). Frequencies were expressed by percentage. Wilcoxon’s matched pairs test and the paired t test were performed accordingly. Univariate comparisons between nominal variables were calculated using the chi-square test or Fisher’s exact test, where appropriate. Kruskal-Wallis one-way analysis of variance was applied to evaluate the comparisons between multiple groups. Spearman’s test was used to test correlation. Two-tailed P values were reported; P values less than 0.05 were considered significant.

Multivariate analysis was performed using binary logistic regression. The results are presented as standard error (SE) with the 95% confidence interval (CI). In order to perform the multivariate analysis, we used a step-forward model including, progressively, those variables with P < 0.1 (to also include those with a trend toward an association) to have a stronger model. Area under the receiver operating characteristic curve (ROC-AUC) was analyzed to evaluate the sensitivity and specificity of anti-CarP and ACPA in identifying the presence of erosive arthritis. Interobserver reproducibility was determined using κ statistics, and κ values were evaluated according to the method of Landis and Koch [28].