Anti-ribosomal P protein IgG autoantibodies in patients with systemic lupus erythematosus: diagnostic performance and clinical profile

Systemic lupus erythematosus (SLE) is a chronic autoimmune disease characterized by multi-organ involvement and by the production of autoantibodies directed against a variety of nuclear and cytoplasmic antigens [1, 2]. Autoantibodies can be detected in patients’ sera years before the diagnosis of SLE is made [3]. Some antibodies are relevant to diagnosis, whereas others are associated with prognostic features or disease activity status [2, 4].

Antibodies against double-stranded DNA (anti-dsDNA) and Smith antigen (anti-Sm) are considered very specific for SLE diagnosis, and both are part of the immunologic classification criteria for this disease [5]. Furthermore, high levels of anti-dsDNA are associated with higher disease activity in SLE [6].

One subset of SLE-specific autoantibodies is directed against ribosomal P (Rib-P) phosphoproteins [2]. The Rib-P antigen consists of three protein components of the 60S ribosomal subunit designated P0 (38 kDa), P1 (19 kDa), and P2 (17 kDa). A pentameric complex of one copy of P0 and two copies each of P1 and P2 interacts with the 28S rRNA molecule to form a GTPase domain that is active during the elongation step of protein translation [7–12]. The major immunoreactive epitope of these ribosomal antigens has been localized to the 22 amino acid carboxy-terminal domain, which is present in all three proteins, and contains two phosphorylated serine residues proteins [2, 8–14].

Anti-Rib-P antibodies are directed against the three subunits [2, 9, 15], and are able to penetrate certain cells, binding to ribosomal proteins and blocking protein synthesis [15]. Anti-Rib-P antibodies enhance the production of tumor necrosis factor (TNF) and interleukin (IL)-6 by activated monocytes and also upregulate the expression of TNF and IL-6 messenger RNA in activated monocytes, indicating that human peripheral blood monocytes express the ribosomal P epitope upon activation [15].

Ethnic background may influence the likelihood of anti-Rib-P antibodies occurring in patients with SLE, with the frequency ranging from 6% to 46% in different ethnic groups [2, 7, 11, 14–16]. In most ethnic groups, anti-Rib-P antibodies are present in 6 to 20% of patients, whereas 36% of Chinese patients with SLE were reported to be positive [7, 11, 12, 15].

Anti-Rib-P antibodies seem to be highly specific for SLE, and might also be a marker for SLE disease activity [12, 14, 15]. The presence of anti-Rib-P antibodies in patients with SLE has been reported to be associated with younger age at disease onset, multiple organ involvement, and an overall severe disease course [8], including presence of central nervous system involvement [2, 4, 7, 11, 12, 15], nephritis [2, 7, 12, 15], photosensitivity [2], malar rash [2], and hepatic involvement [2, 7, 12]. Moreover, it has become evident that anti-Rib-P antibodies are more prevalent in juvenile-onset than in adult onset SLE [11, 12]. Bonfa et al. first assessed the association of anti-Rib-P antibodies with psychiatric features in patients with psychosis secondary to SLE [17]; however, other studies have not confirmed this association [7, 8].

We hypothesized that anti-Rib-P autoantibodies might be useful for SLE diagnosis. To test this hypothesis, we used a new fluorescence enzyme immunoassay (FEIA) kit to quantify levels of anti-Rib-P in patients with SLE, controls with other rheumatic diseases (rheumatic disease control (RDC) group, which included rheumatoid arthritis (RA), juvenile idiopathic arthritis (JIA), ankylosing spondylitis (AS), and psoriatic arthritis (PsA)), and healthy controls (HC group).

ROC curves were constructed to obtain the most adequate cut-off values for the Portuguese population; these curves are of particular relevance for the new anti-Rib-P test kit. Curves were also constructed for the other tests for coherence analysis. The curves are presented in Figure 1 with the AUC and the corresponding P-values identified. For both the anti-Rib-P and anti-Sm tests, the cut-off values after analysis of the ROC curves were 4.45 U/ml and 3.4 U/ml, respectively. For anti-dsDNA, the cut-off value given by the manufacturer (15 U/ml) was used because it corresponds to the value obtained from the ROC curves. With these adjustment in cut-off values, we identified a higher number of patients with SLE who were positive for either anti-Rib-P or anti-Sm, without incurring more false-positive results in the control groups than we did with the manufacturer cut-off points (data not shown).

We found that the levels of anti-Rib-P autoantibodies were significantly higher in the SLE group (mean concentration of 4.9 ± 20.2 U/ml) than in the HC group (0.07 ± 0.21 U/ml; P = 0.016) or the RDC group (0.6 ± 1.8 U/ml; P = 0.017). In 18 samples of the SLE group (14.2%), anti-Rib-P was above the cut-off value of 4.45 U/ml for positivity (mean concentration of 30.6 ± 46.9 U/ml). Of note, in the RDC group, two patients with RA (0.8%) were positive for anti-Rib-P autoantibodies (18.9 ± 9.8 U/ml), whereas none of the HCs tested positive for this antibody.

The mean concentration of anti-Sm antibodies for the whole SLE group was 2.8 ± 13.8 U/ml, and 12 of these positive samples (9.4%) had a significantly higher mean concentration (31.1 ± 40.8 U/ml) than those of the HC group (0.02 ± 0.11 U/ml; P = 0.028) or the RDC group (0.1 ± 0.3 U/ml; P = 0.035). Positive results (above 3.4 U/ml) for anti-Sm autoantibodies were found only in the SLE group.

Patients with SLE also had a significantly greater mean concentration of anti-dsDNA antibodies (44.6 ± 73.8 U/ml) than found in the HC group (3.5 ± 8.1 U/ml; P < 0.001) or the RDC group (2.6 ± 4.2 U/ml; P < 0.001). Of the 127 patients with SLE, 63 (49.6%) were positive for anti-dsDNA (mean concentration 88.4 ± 88.5 U/ml.), compared with 6 samples (6.0%) in the HC group and 5 samples (2.0%) in the RDC group (1 RA, 1 JIA, and 3 AS samples).

Anti-Rib-P levels were related at (P ≤ 0.20) in univariate analysis with age (β = −0.125), Caucasian ethnicity (β = −0.190), erythrocyte sedimentation rate (ESR; β = 0.175), disease activity (SLEDAI2K; β = 0.154), malar rash (β = 0.142), renal disorder (β = 0.153), hematologic disorder (β = 0.130), and current corticosteroid dosage (β = 0.119). Hence, these variables were included in the multivariate analysis, which showed that Caucasian ethnicity (β = −0.190, P = 0.034) was the only factor independently associated with anti-Rib-P levels in patients with SLE (Table 4). Anti-Rib-P antibodies were not associated with previous neurologic disorder (seizure or psychosis) or with the occurrence of neuropsychiatric lupus features within the subsequent 3 years of follow-up.

The variables potentially associated with anti-Sm levels from the univariate analysis (at P ≤ 0.20) were Caucasian ethnicity (β = −0.060), ESR (β = 0.203), C-reactive protein (CRP) (β = 0.372), SLEDAI2K (β = 0.125), malar rash (β = −0.138), photosensitivity (β = 0.148), serositis (β = 0.277), renal disorder (β = 0.176), anti-RNP antibodies (β = 0.304), current corticosteroid dosage (β = 0.164), and use of immunosuppressants (β = 0.209). Higher CRP levels (β = 0.304, P = 0.003), serositis (β = 0.321; P = 0.002), and previous positivity for anti-RNP antibodies (β = 0.297; P = 0.003) were found to be independently associated with anti-Sm levels in patients with SLE (Table 4).

For anti-dsDNA levels, age (β = −0.237), age at disease onset (β = −0.169), disease duration (β = −0.176), ESR (β = 0.187), SLEDAI2K (β = 0.413), arthritis (β = −0.150), renal (β = 0.287), hematologic (β = 0.259), and immunologic disorders (β = 0.186), and current corticosteroid dosage (β = 0.130) came out as candidate predictors for higher anti-dsDNA levels (at P ≤ 0.20 in univariate analysis).In the multivariate analysis, SLEDAI2K (β = 0.338; P < 0.001), renal disorder (β = 0.252; P = 0.004), and shorter disease duration (β = −0.246; P = 0.005) were found to be independently associated with anti-dsDNA levels (Table 4).

Confirming earlier studies, the current work shows that anti-Rib-P protein autoantibodies are very specific for SLE diagnosis. The presence of antibodies against ribosomal P proteins was found to be very specific for patients with SLE compared with either HCs or with controls who had other rheumatic diseases. Moreover, the test had high levels of specificity and sensitivity. However, the choice of the most reliable test to determine these autoantibodies requires a comparative study between different tests and the study of a larger and multi-ethnic population.

In addition to determining the levels of anti-Rib-P autoantibodies, we used the same FEIA detection method to determine levels anti-Sm and anti-dsDNA autoantibodies in the same study groups. Both anti-Sm and anti-dsDNA antibodies have also been reported to be very specific for patients with SLE [21–23]; however, we found that anti-dsDNA antibodies were present at low levels in 6% of HCs and 2% of RDCs samples.

The commercial kit that we used for the determination of anti-Rib-P protein (EliA test) is an FEIA, designed as a sandwich immunoassay, containing a mixture of the three Rib-P antigens (P0, P1, and P2), which has been described previously as having high sensitivity and specificity [7, 11, 24]. We also used ROC curves to check the accuracy of this kit for the Portuguese population. ROC curves can be used to evaluate the diagnostic performance of a test, adjusting for a particular study population, and to determine the capability of a test to allow discrimination between the positive group and the control group [25, 26]. Based on the ROC curves, we adjusted the cut-off values for both anti-Rib-P and anti-Sm to 4.45 U/ml and 3.4 U/ml, respectively. These values corresponded to the lowest concentration that allowed the highest possible sensitivity without losing specificity, establishing a cut-off value for the SLE group in comparison with the HC and RDC groups. For anti-dsDNA determination, we used the manufacturer’s cut-off value (15 U/ml) in subsequent analyses, as this gave the best combination of sensitivity and specificity. The cut-off confirmation should be performed when using a new kit or when using an existing kit in a different population. The adjusted values might be either higher or lower than those established by the manufacturer, as confirmed by the work of Mahler and colleagues [12].

Our results showed increased levels of all the three autoantibodies in patients with SLE, and a higher percentage of positive samples for at least one of the autoantibodies in the SLE group. Although anti-dsDNA autoantibodies were present in more individuals in the SLE group than in the other two groups, the presence of anti-Rib-P or anti-Sm was more specific for SLE diagnosis.

We reviewed the medical records of the individuals in the HC and RDC groups who had a positive result for either anti-Rib-P or anti-dsDNA antibodies (none was positive for anti-Sm). Both of the anti-Rib-P-positive results were detected in patients with RA, one of these patients had presented with some lupus-like characteristics (skin rash, leucopenia, and aphthous ulcer) at some point in the disease course, and thus this case could be classified as an overlap RA/SLE. Interestingly, a similar case was previously reported, referring to a anti-Rib-P-positive patient with RA, who later developed renal disease, and their condition evolved into full-blown SLE [11]. None of our HC or RDC group who were positive for anti-dsDNA autoantibodies had presented any lupus-like characteristics at any time.

When we used multivariate analysis on our SLE group, the only independent association with anti-Rib-P antibodies we identified was ethnicity: lower anti-Rib-P levels were present in individuals of Caucasian ethnicity. To our knowledge, no previous reports have established this association. However, given the small number of non-Caucasian patients in our group, these findings need to be replicated in larger SLE populations with different ethnic backgrounds.

Many previous studies have reported a relationship between the presence of anti-Rib-P antibodies and that of some clinical features, namely malar rash, renal involvement, and neuropsychiatric events, particularly psychosis [8, 11, 27, 28]. However, there are also reports that corroborate our findings of an absence of such an association between the presence of anti-Rib-P antibodies and clinical features or disease activity [7, 13, 16]. In addition, our analysis differed from previous reports because we also took into consideration anti-Rib-P levels.

We found that Rib-P positivity was not associated with previous neuropsychiatric features classifiable by the ACR criteria [29, 30], and the presence of these autoantibodies did not have a predictive value for the occurrence of neuropsychiatric symptoms in the subsequent 3 years. Thus, these autoantibodies seem to be very specific for SLE, but their value for diagnosis of neuropsychiatric lupus seems to be limited, possibly because both anti-Rib-P positivity and neuropsychiatric symptoms are relatively rare. However, given the high specificity, the inclusion of these autoantibodies as part of the SLE classification criteria might be useful. To confirm this, further studies encompassing a larger and multi-ethnic population are needed. Besides its possible use in established SLE, it will be important to assess the performance of such a test in patients with early-stage disease to confirm whether the inclusion of anti-Rib-P testing can improve diagnostic accuracy for SLE.

We performed a multivariate analysis for anti-Sm and anti-dsDNA levels, which revealed some associations of these antibodies with features of the disease. Serositis and CRP levels were positively associated with higher anti-Sm levels. High CRP levels are usually associated with an ongoing infection in patients with SLE, although they have been also associated with serositis, independently of the existence of an infection [31, 32]. This is in line with our results, as we found that CRP levels were increased in patients with serositis (P = 0.047). However, the association between serositis and anti-Sm antibodies contradicts the results of a previous report from Wang and co-workers [33].

The multivariate analysis also showed anti-RNP levels to be independently associated with anti-Sm levels. Both anti-Sm and anti-RNP antibodies recognize complexes that contain small nuclear RNA species, and the occurrence of anti-Sm antibodies along with anti-RNP antibodies has been reported previously [34].

Our observations regarding anti-dsDNA are in line with other classic findings depicting an association with renal involvement, as well as in relation to lower disease duration and higher disease activity [35].

We also evaluated cross-positivity for the three studied autoantibodies, and verified that 78% of the anti-Rib-P-positive patients were also positive for one or both of the other antibodies determined. Previous studies have also shown that the presence of anti-Rib-P antibodies is often associated with anti-dsDNA antibodies, but the simultaneous presence of anti-Rib-P and anti-Sm is not consensual between studies [2, 8, 11, 13]. However, we found that four patients (3.1%) in our SLE group (22% of the anti-Rib-P-positive patients) were positive for anti-Rib-P autoantibodies only. We reviewed the clinical records of these four patients, and found no particular clinical features in common.

The presence of anti-Rib-P antibodies in patients negative for anti-DNA and anti-Sm indicates that these autoantibodies might be useful for SLE diagnosis, as previously reported by Mahler and colleagues [10]. Based on previous suggestions by other authors, and considering that disease classification criteria are constantly subject to confirmation and re-evaluation studies, as recently published by Petri and colleagues [36], we propose that further studies should be performed to evaluate the relevance of anti-Rib-P antibody testing for SLE diagnosis.