Diagnostic value and clinical laboratory associations of antibodies against recombinant ribosomal P0, P1 and P2 proteins and their native heterocomplex in a Caucasian cohort with systemic lupus erythematosus

Introduction In this study, we sought to determine the diagnostic value and clinical laboratory associations of autoantibodies against recombinant ribosomal P0, P1 and P2 proteins and their native heterocomplex in systemic lupus erythematosus (SLE). Methods Autoantibodies against recombinant ribosomal P proteins (aRibPR0, aRibPR1 and aRibPR2) and antibodies against native ribosomal P heterocomplex (aRibPNH) were determined in sera from patients with SLE (n = 163), systemic sclerosis (n = 66), Sjögren's syndrome (n = 54), rheumatoid arthritis (n = 90) and healthy donors (n = 100) using enzyme-linked immunosorbent assay. Test results were correlated to medical records, including the American College of Rheumatology criteria, the Systemic Lupus Erythematosus Disease Activity Index 2000, laboratory data and medications of all SLE patients. Results Sensitivities of 22.0% for aRibPR0, 14.9% for aRibPR2, 14.3% for aRibPNH and 10.7% for aRibPR1 were obtained at a specificity of 99%. The assay for aRibPR0 detection demonstrated the best performance in receiver-operating characteristics analysis, with aRibPR0 detectable in 10% of anti-Smith antibody and anti-double-stranded DNA-negative sera at a specificity of 100%. ARibPR0 positivity was associated with lymphocytopenia. ARibPR1+ patients had significantly higher γ-glutamyl transpeptidase (GGT) levels than their aRibPR1- counterparts. No specific damage occurred in aRibP+ lupus patients compared with a group of age-, sex- and nephritis-matched aRibP- lupus patients within 3 years. Conclusions The determination of antibodies against ribosomal P proteins improves the diagnosis of SLE and should therefore be implemented in upcoming criteria for the diagnosis or classification of SLE. High titers of aRibPR0 can be associated with lymphocytopenia, and high titers of aRibPR1 can be associated with elevated GGT levels. So far, there is no evidence for a prognostic value of aRibPs for damage.


Introduction
Systemic lupus erythematosus (SLE) is a chronic, multifaceted rheumatic disease which is characterised by the generation of autoantibodies predominantly directed against nuclear proteins and nucleic acids [1,2]. However, antibodies against cytoplasmatic antigens such as those binding to ribosomal P proteins (aRibPs) have been reported to be specific for SLE as well [2,3]. In contrast to anti-Smith (anti-Sm) and anti-double-stranded DNA (anti-dsDNA) antibodies, anti-ribosomal P protein antibodies are not included in the current American College of Rheumatology (ACR) classification criteria for SLE [4,5].
The human ribosomal phosphoproteins P0 (38 kDa), P1 (19 kDa) and P2 (17 kDa) are located within the 60S ribosomal subunit, forming a pentameric complex consisting of a P0 anchor and two P1/P2 heterodimers [3]. The subunits of that pentamer have a common immunodominant epitope at the carboxyl terminus [6], which can lead to cross-reactions of anti-ribosomal P antibodies with P0, P1 and P2 units. P proteins can also exist as ribosome-free P0, P1 and P2 forms in the cytoplasm [6,7]. Notably, the P0-like protein is also detectable in the plasma membranes of hepatocytes, lymphocytes and other cells [8][9][10][11].
A comparative investigation of the clinical laboratory associations of antibodies against recombinant ribosomal P0, P1 and P2 proteins (aRibP R 0, aRibP R 1 and aRibP R 2) has never been conducted. Thus, the purpose of the present work was to determine the diagnostic value of antibodies against native ribosomal P heterocomplex (aRibP N H), aRibP R 0, aRibP R 1 and aRibP R 2 for SLE and to analyse their associations with disease features and future damage.
Disease activity of SLE patients was defined based on the Systemic Lupus Erythematosus Disease Activity Index 2000 (SLEDAI 2000) [36][37][38] in 101 patients: 6 of them had no activity (SLEDAI score 0), 35 were mildly active (0 < SLEDAI ≤ 5), 41 had moderate disease activity (5 < SLEDAI ≤ 10), 14 were highly active (10 < SLE-DAI ≤ 20), and 5 had very high activity (SLEDAI > 20). Juvenile onset was diagnosed when the age at diagnosis was 18 years or younger according to the Pediatric Rheumatology International Trials Organization [39]. Twenty-four (14.7%) patients with juvenile onset SLE and 139 (85.3%) patients with adult onset SLE were studied. Disease damage was measured according to the criteria of the Systemic Lupus International Collaborative Clinics (SLICC) [40,41] and the weighted damage score (WDS) [40]. All patients were recruited from the outpatient and inpatient facilities of the Department of Rheumatology and Clinical Immunology, Charité University Hospital, Berlin, Germany. The Ethics Committee of the Medical Faculty of Charité University Hospital approved the study, and written informed consent was obtained from all subjects. Sera from healthy donors were used in cooperation with University of Lübeck, Germany. Written informed consent was obtained from all healthy subjects.

Measurement of antibodies
Microtiter plates (Nunc, Roskilde, Denmark) were coated with 1 μg/ml full-length recombinant ribosomal protein P0, P1 or P2 expressed in insect cells (DIA-RECT, Freiburg, Germany). Sera diluted 1:201 in phosphate-buffered saline (PBS) and 0.1% (wt/vol) casein were added and allowed to react for 30 minutes, followed by three washing cycles with PBS 0.05% (vol/vol) and Tween 20. For detection of bound antibodies, the plates were incubated with antihuman immunoglobulin (IgG) peroxidase conjugate (EUROIMMUN, Lübeck, Germany) for 30 minutes, washed three times and allowed to react with tetramethylbenzidine (EUROIM-MUN) for 15 minutes. After addition of acidic stopping solution (EUROIMMUN), the optical density (OD) was read at 450 nm using an automated spectrophotometer (Spectra Mini; Tecan, Crailsheim, Germany). All steps were performed at room temperature. A highly positive index patient serum was used to generate a standard curve consisting of three calibrators (2, 20 and 200 relative units (RU)/ml). Relative units per milliliter were calculated for all samples using this three-point standard curve. The analytical reproducibility of all aRibP assays was evaluated by repeated testing of two serum samples (10 determinations each) in the same run, giving intraassay coefficients of variation (CV) of 2.4% (aRibP R 0), 2.1% (aRibP R 1) and 2.7% (aRibP R 2), respectively. Relationships between sensitivity and specificity at different cutoff values were examined for all assays by receiveroperating characteristics (ROC) curve analyses, allowing also for the determination of test characteristics at predefined specificities.

Statistical analysis
Statistical analyses were performed using GraphPad Prism 5 software (GraphPad Software, La Jolla, CA, USA). The diagnostic significance of antiribosomal proteins N, P0, P1 and P2 antibodies was assessed and areas under the curve (AUCs) were created using ROC analysis. To determine associations, the Mann-Whitney U test (for comparing medians between groups; MWT), Fisher's exact test (FET) and Spearman's rank test (SRT) were used. Two-tailed t-tests were used throughout with an α set at 0.05.

Results
Reactivity and diagnostic significance of antiribosomal proteins N, P0, P1 and P2 antibodies Antibodies against ribosomal P N H, P R 0, P R 1 and P R 2 proteins (Figure 1), Sm and dsDNA (ELISA and Farr assays) were measured in sera from 163 SLE patients, 210 disease controls and 100 healthy donors to define and compare the sensitivity and specificity in ROC curve analysis ( Table  1). For aRibP N H, a sensitivity of 5.5% and a specificity of 100% were calculated using the manufacturer's cutoff (20 RU/ml). At a predefined specificity of 98% among 210 patients with other rheumatic diseases (SSc, pSS and RA), only five (2.4%), four (1.9%), four (1.9%) and four (1.9%) had elevated aRibP N H, aRibP R 0, aRibP R 1 and aRibP R 2 titers, respectively. At the same specificity among 100 healthy donors, only zero (0%), one (1.0%), two (2.0%) and two (2.0%) patients had high titers of aRibP N H, aRibP R 0, aRibP R 1 and aRibP R 2. Among antiribosomal P protein antibodies, aRibP R 0 had the highest performance with regard to criteria such as AUC and maximum sum of sensitivity and specificity, followed by aRibP N H ( Table 1). All test criteria of aRibP R 0 were inferior to those of the anti-dsDNA ELISA or the Farr assay, but were almost equal to those of the anti-Sm ELISA.
Patients negative for aRibP N H but positive for aRibP R P0-2 Although the native heterocomplex of ribosomal P contains all immunological domains of the subunits P0, P1 and P2, there were considerable differences in the cutoffs and in sensitivities for the detection of aRibP N H, aRibP R 0, aRibP R 1 and aRibP R 2 (Table 1), with outstanding results for aRibP R 0. Thus, we further investigated whether there were patients negative for aRibP N H but positive for aRibP R 0, aRibP R 1 or aRibP R 2 ( Figure 2). Sera fulfilling these criteria would point out that there are some epitopes of ribosomal P proteins that are not accessible to autoantibodies because of the spatial conformation of the native heterocomplex.   At 99% specificity, among 141 aRibP N Hpatients there were 19 (13.5%) positive for aRibP R 0, six (4.3%) positive for aRibP R 1 and 11 (7.8%) positive for aRibP R 2. Some of those sera were further exclusively positive for one of the recombinant aRibPs and showed an increased titer up to twofold of the corresponding cutoff (Figure 2b).

Diagnostic value of anti-ribosomal P protein antibodies in SLE
To investigate the auxiliary diagnostic value of antiribosomal P protein antibodies in SLE, we searched for patients who were negative for antibodies against dsDNA and Sm, but positive for aRibP N H, aRibP R 0, aRibP R 1 or aRibP R 2 at a specificity of 100% (Figure 3). This analysis was performed twice, taking either the results of the anti-dsDNA ELISA (Figure 3a) or those of the Farr assay (Figure 3b).
Among 163 SLE patients, there were 11 (6.7%) individuals who could be diagnosed only by detection of aRibPs, while 63 (38.7%) patients were regularly diagnosed by the presence of anti-dsDNA or anti-Sm antibodies. Considering the excellent Farr assay, these  The cutoffs of all test systems were set to ensure an absolute specificity of 100% (see Table 1 for cutoffs). Anti-dsDNA and anti-Sm antibodies were selected because they are highly specific for systemic lupus erythematosus (SLE) (instead of, for example, anti-phospholipid antibodies) and are part of American College of Rheumatology classification criteria for SLE [4].  Comparison of disease features in aRibP + vs. aRibP -SLE patients To determine the special characteristics of lupus patients with elevated aRibPs, we compared medical records, including ACR criteria, SLEDAI 2000 items and laboratory parameters, including autoantibodies, immunosuppressants and antimalarials, with those of aRibPlupus patients. All clinical laboratory results and detailed demographic information about the study cohort are shown in Table 2.
ARibP N H + patients fulfilled significantly more ACR criteria and more often had photosensitivity. Moreover, the frequency of patients with decreased C3 levels was higher among aRibP N H + patients. Lymphocytopenia was associated with the presence of aRibP R 0, and a higher γ-glutamyl transpeptidase (GGT) level was found in aRibP R 1 + patients. Anti-Sm, anti-dsDNA and anti-U 1ribonucleoprotein (anti-U 1 -RNP) antibodies were much more frequent in all aRibP + patients.

Comparison of disease damage in aRibP + vs. aRibP -SLE patients
To study the prognostic role of ribosomal P protein antibodies, SLICC scores and WDS were assessed in aRibP + patients and in an age-, sex-and nephritismatched group of aRibPpatients at the time of blood sampling and 3 years later. Changes in damage scores (ΔSLICC, ΔWDS) were calculated, and both groups were separately compared. Damage scores from 41 of all 58 aRibP + patients were completely assessable at the time of blood sampling and 3 years later. There were 22 aRibP N H + , 27 aRibP R 0 + , 18 aRibP R 1 + and 23 aRibP R 2 + patients. SLICC and WDS correlated significantly with disease duration and the ages of patients, but not with ACR scores or with anti-dsDNA, anti-Sm or any antiribosomal P protein antibodies. Total disease damage and damage to every organ system separately was not significantly higher in aRibP + patients than in their aRibPcounterparts within these 3 years. Thus, we found no prognostic role for aRibP.

Discussion
In this study, the diagnostic potential, clinical laboratory associations and correlations with disease damage of antibodies directed against the native ribosomal heterocomplex and its recombinantly produced constituents P0, P1 and P2 were investigated. ARibP R 0 revealed the best diagnostic performance among all aRibP types and offered the most diagnostic benefit among sera negative for anti-dsDNA and anti-Sm antibodies. Furthermore, aRibP R 0 + lupus patients had significantly lower lymphocyte counts than their aRibP R 0counterparts. Finally, no prognostic relevance was found for any of the aRibPs during a 3-year period.
Our results concerning the prevalence and high specificity of aRibPs for SLE agree with data described before [3,42]. We further found sensitivities of P R 0 > P N H > P R 2 > P R 1 at specificities of 98% to 99% and P N H > P R 0 = P R 2 > P R 1 at a specificity of 100% in a cohort of 163 lupus patients. This is in contrast to another study where sensitivities of P R 2 = P R 1 = P R 0 were determined at a specificity of 100% in a cohort of 50 SLE patients [13]. Different detection systems and patient cohorts might have contributed to these divergent observations. Since all three subunits of aRibPs share the carboxylterminal epitope, it is of interest to note that an ELISA (referred as anti-C22 ELISA) detecting antibodies against this shared epitope reached the same sensitivity of 22% at a specificity of nearly 99% as aRib P R 0 in our Berlin patient cohort [43].
We have additionally demonstrated that negativity of aRibP N H does not automatically imply negativity of antibodies against its subunits, especially those against ribosomal P0. This could be due to immunologically relevant epitopes that are freely accessible using RibP R 0 alone, but not within the spatial conformation of the native heterocomplex. A biological reason for the higher frequency of aRibP R 0 might be the disposability of ribosomal P0-like protein in the cell membranes of many cells, which could contribute to an increased immunogenicity [8][9][10][11].
Among the vast quantity of antibodies that are detectable in SLE, antibodies against dsDNA and Sm are highly specific and therefore most useful for the verification of the diagnosis. However, aRibPs are also discussed as a diagnostic criterion. Therefore, we asked whether aRibPs provide additional diagnostic benefit in direct comparison to anti-dsDNA and anti-Sm antibodies. Exactly 10% of sera negative in the anti-Sm and anti-dsDNA ELISAs were positive for aRibP R 0 at a specificity of 100%. Even the comparison including the Farr assay revealed that 5.4% of all anti-Sm ELISA and anti-dsDNA RIA (Farr assay) negative sera were positive for aRibP R 0 at 100% specificity. Thus, laboratories using less sensitive assays seem to benefit more from testing for aRibP in suspected cases of SLE. However, to be sure, all patients with suspected diagnosis of SLE should be tested for aRibP. Finally, we conclude that the determination of antibodies against ribosomal P proteins, especially those against P0, would improve the classification and diagnosis of SLE.
By comparing disease features of lupus patients with elevated aRibPs to their seronegative counterparts, we could not confirm an association of aRibP positivity  with lupus nephritis, short disease duration, high disease activity or juvenile onset. These results might be influenced by the Caucasian ethnicity of the study cohort and differences of the test systems. Cases of neuropsychiatric lupus [28,29] and subtypes of lupus nephritis were not recorded in our study. The most striking association of aRibPs with disease features was that aRibP R 0 + lupus patients had significantly lower lymphocytes than aRibP R 0lupus patients. Interestingly, a P0-like protein is also detectable in the plasma membranes of different cells, including lymphocytes [11]. Further, the aRibPs are able to bind and penetrate T-cell lines [44,45], and especially aRibP R 0 can induce apoptosis in Jurkat T-cells [46]. In that context, our data confirm the thesis of Sun et al. [46] that aRibP R 0 contributes in a clinically relevant manner to lymphocytopenia in SLE. Thus, clinicians should keep aRibP R 0 in mind as one differential diagnosis for lymphocytopenia in SLE, along with viral status, drug side effects, hematologic malignancies and other factors.
Another remarkable, significant clinical laboratory association was that aRibP R 1 + patients had an elevated GGT value. The participation of aRibPs in liver pathology of SLE was previously reported in cell cultures [9,11,46] and in case reports [19][20][21]. However, aRibP R 0 were most frequently in focus because of their membrane-bound isoform [8][9][10][11]. As such, in a study of 61 Japanese patients [22], no significant association was found between aRibP R 0 and liver enzymes alanine aminotransferase or aspartate aminotransferase, but the GGT level was not assessed. The correlation shown here between GGT and aRibP R 1 indicates a possible association of aRibP with lupus hepatitis. However, we do not have a clear definition of lupus hepatitis, and it is hard to rule out other causes, such as nutrition, drugs and other autoimmune hepatitis forms. Longitudinal analysis of aRibPs with liver function tests, including GGT in parallel, might reveal this association best.
Up to now, accepted prognostic factors in SLE have only been lupus nephritis and neuropsychiatric SLE. No prognostic laboratory parameter is known. In this study, we investigated whether aRibP + lupus patients would develop more or specific disease damage measured by SLICC or WDS after 3 years than their aRibPcounterparts. However, no significant correlations with any of the antiribosomal P protein antibodies could be found over a 3-year period. Conclusively, we first show that aRibPs are not a prognostic parameter for damage in SLE. Further study with more patients and over longer observation time frames could strengthen this result.

Conclusions
In summary, antiribosomal P protein antibodies are very specific for SLE, can also be found in patients with negative anti-dsDNA and anti-Sm antibodies and therefore have to be discussed in the upcoming classification and diagnostic criteria for SLE. Among all four investigated aRibPs, aRibP R 0 was the most abundant and should be used for the diagnosis of SLE. High aRibP R 0 titers can be associated with lymphocytopenia, and high aRibP R 1 titers can be associated with an elevated GGT level. A prognostic role of antiribosomal P protein antibodies is unlikely.