Does serum B lymphocyte chemoattractant protein13 (CXCL13) level correlate with parameters of disease activity and severity in rheumatoid arthritis? Clinical and musculoskeletal ultrasonographic assessment

Background: Rheumatoid arthritis (RA) affects about 1 – 3% of the population making it one of the most popular autoimmune diseases. Chemokines through switching on their receptors exert a climacteric role in RA pathogenesis. The purpose of this cross-sectional study was to quantify the serum levels of serum B lymphocyte chemoattractant protein13 (CXCL13) chemokine in recent onset RA patients and to correlate its levels with clinical, laboratory, and musculoskeletal ultrasonographic parameters (MSUS) of disease activity and severity. Results: The mean serum CXCL13 value showed a significant increase in the RA patients (388.86 ± 283.63 pg/ml) than in the controls (62.94 ± 31.62 pg/ml) ( P < 0.001). Highly active RA patients had significantly the highest mean of CXCL13 (mean ± SD 819.13 ± 191.05) compared with the moderately active RA patients (mean ± SD 284.95 ± 137.93) ( P < 0.001) and the RA patients with low disease activity (mean ± SD 129.5 ± 21.27) ( P < 0.001) and its levels were positively related with clinical disease activity and musculoskeletal ultrasonographic severity parameters. Conclusion: Serum CXCL13 is correlated with clinical disease activity and MSUS disease severity that encourages its use for monitoring the activity and severity of synovitis in recent onset RA patients. Future studies to detect the effect of disease activity control by medications on CXCL13 levels and the effect of the CXCL13 antagonist on controlling RA disease activity and severity are recommended.


R E T R A C T E D A R T I C L E
Background Rheumatoid arthritis (RA) affects about 1-3% of the world's population, and it is considered one of the most popular autoimmune diseases. Hands and feet's small synovial joints are primarily involved [1]. It is characterized by synovitis extending to the adjacent articular cartilage leading to erosion and further tissue damage. It is not just an articular disease, extra-articular organs could be affected thereby worsening disease prognosis and leading to lifelong disability and discomfort [2].
Etiopathogenesis for RA is obscure, it is a prototypic inflammatory disease that results from the interplay between genetic and environmental factors which alter the immunological homeostasis, wherein immunological stimulation and unwanted inflammation predominate. Inflammatory changes involving the synovium of the inflamed joints leading to pain, warmth, swelling, and stiffness that usually experienced early in the morning or after prolonged inactivity that lasts more than 30 min [3].
Chemotactic cytokines and their receptors exert a climacteric role in the pathogenesis of RA. They synchronize and induce the immune cells migration that direct not just the ordinary turn of events and homeostasis of the immune system but also responsible for diverse immune system intolerance reactions leading to inflammatory and destructive immune responses in many rheumatic disorders [4]. Chemokines super family are grouped according to the gathering of cysteine molecule in their two amino acid terminals. Their names remain on the position of the two cysteine buildups differentiating them into 4 groups: CC, CXC, CX3C, and XC. No other amino acid isolates the two cysteine deposits in the CC chemokines while at least an amino acid isolates the cysteine buildups in the CXC chemokines [5].
Contribution of the CXCL13 in the pathogenesis of synovitis in RA has been recently revealed. At the point when it is created locally, it is related with extra nodal lymphoid aggregates that house ectopic B lymphocyte responses with other cell subsets and intercede development of the immunoglobulin genes. It follows up on various cells having CXCR5 receptors like dendritic cells and bone cells and aids chemotaxis of CD4+ follicular T partner cells. CXCL13 has been thus proposed as a proxy marker of synovitis [1].
Initial outcomes recommended that CXCL13 serum levels and its mRNA expression in the synovium are associated, proposing that the inflamed synovium is a significant wellspring of the circling CXCL13 and clinically, this makes it conceivable and simple to identify and measure this chemokine not just from the synovial membrane, anyway additionally in the synovial liquid and circulation [6].
Early and strict control of the inflammation and the possibility to predict severity of synovial inflammation is vital to set proper algorithm to treat the patients with RA [7].
Remission that is detected clinically does not necessarily mean actual concealment of the inflammation, as remittent patients may proceed to have subclinical active synovitis that can be identified and scored specifically by power doppler (PD) musculoskeletal ultrasound (MSUS) mode. PD recognized signals in the synovium can anticipate joint harm and short-term relapse in treated patients who are considered to be clinically remittent [8,9]. In this manner, identification of surrogate markers dependent on understanding RA pathogenesis and to correlate with the seriousness of synovitis at the clinical and sub-clinical level, is required [10]. We planned to quantify levels of CXCL13 chemokine in the sera of recent onset RA patients and to associate its levels with clinical, laboratory, and musculoskeletal ultrasonographic findings of disease activity and severity.

Methods
This cross-sectional study included 50 recent onset RA patients who fulfilled the updated American College of Rheumatology/European League Against Rheumatism criteria for the classification of RA [11].
They had disease duration ≤ 12 months attending the outpatients′ clinic and the in-patients' units of our department. Exclusion criteria included patients with liver or endocrine disorders, metabolic diseases, recent infections, trauma and, malignancy or any other rheumatic disease.
RA patients had intra-articular steroid injection in the examined joints since 3 months of the study were also excluded.
Forty apparently healthy subjects were filled in as a benchmark group. The study was endorsed in concordance to the 1983 Helsinki Declaration Statement. All members gave an oral consent before partaking in the study.
Full history was taken from the patients including the age, sex, disease duration, morning stiffness duration, patient global health assessment based on visual analog scale from 0 to 10 (pVAS) and medical treatments.
Thorough clinical examination was done, disease activity 28-joint (DAS-28) score was used to assess RA disease activity. Remission was considered if DAS-28 is < 2.6, low activity was considered if DAS-28 was ≥ 2.6 and < 3.2, moderate activity was considered if DAS-28 was ≥ 3.2 and < 5.1 and high activity was considered if DAS-28 was ≥ 5. 1 [12].

Radiological examination
Plain X-ray of hands and feet were assessed and graded according to Larsen et al. [13]. The Larsen list score was applied to 20 joints reciprocally including metacarpophalangeal joints (MCPs), wrists, the second to the fifth metatarsophalangeal joints (MTPs), each joint was graded from (0-5) with a total score range from 0 to 100.
Musculoskeletal ultrasonography (MSUS)/power doppler assessment (PD): RA patients were scanned by MSUS that was done by a rheumatologist experienced in MSUS who didn't know the clinical information of the patients, using a GE Logiqe 9 scanner (General Electric Medica Systems, USA) with a linear transducer (8)(9)(10)(11)(12)(13). Gray-scale (GS) and power doppler (PD) examination were done in both longitudinal and transverse scan for 6 joints in every RA patient; the wrist (dorsal radiocarpal and midcarpal joints), 2nd MCP and supra patellar recess of the knee joint bilaterally [14]. The gain and depth were adjusted according to examined joint.
Each joint was assessed by GS mode for the presence of synovial hypertrophy (SH) and erosions. According to European League against rheumatism-Outcome Measures in Rheumatology (EULAR-OMERACT), synovitis was diagnosed by the presence of a hypoechoic synovial hypertrophy (SH) regardless of the presence of effusion or any grade of PD signal [15]. US-SH was defined as presence of abnormal hypoechoic synovial tissue within the capsule that was not displaceable and poorly compressible and it may exhibit PD signals. US-SH was graded from 0 to 3; grade 0 (normal): no SH regardless the presence of effusion, grade 1: minimal hypoechoic SH up to the level of the horizontal line joining bone surfaces, grade 2: moderate hypoechoic SH extending out the joint line but with the upper surface concave, grade 3: severe hypoechoic SH with or without effusion extending above the joint line but with the upper surface convex.
US detected erosions were defined as intra-and/or extra-articular discontinuity of bone surface that should be seen in two perpendicular planes [15].US synovitis activity was assessed by PD mode for detecting US-PD signals in the SH of each joint that was graded from 0 to 3; grade 0: no PD signal, grade 1: from one to three single spots or one confluent spot and two single spots or up to two confluent spots, grade 2: > grade 1 but < 50% of doppler signals in the background, grade 3: more than grade 2; > 50% of the total gray-scale back ground [15]. For each patient the total US-SH and US-PD score ranged from 0 to 18. Total musculoskeletal score (0 to 36) is the sum of US-SH and US-PD grades of the 6 joints for the patient [14].

Quantitative detection of serum CXCL13 level
Serum samples from all enrolled subjects were gathered and were put away at − 20°C until investigation. Levels were estimated by enzyme-linked immunosorbent assay (ELISA) (Boster Biological Technology Co.) recorded in pg/ml [16]. Serum CXCL13 levels were evaluated by the manufacturer's guidelines utilizing a commercially accessible ELISA unit (Quantikine human CXCL13/BCL/ BCA-1, #DCX130 R&D systems, USA) that was previously validated [17]. All tests were diluted 1:2 in Calibrator Diluent RD6-41 enhanced with mouse and bovine immunoglobulin G to guarantee preaggregation of heterophilic antibodies. Tests were examined in duplicates and the cutoff limit was determined as two standard deviations of the blanks.

Statistical methods
Data were presented as means ± standard deviation (SD), median and interquartile range (IQR), or numbers and percentage. The significance of difference was tested using Student's t test to compare between the mean of two groups of parametric data. For continuous nonparametric data, Mann-Whitney U test was used. Chisquare test was used for categorical parameters. Spearman's relationship coefficient was considered. Regression analysis was used to identify the independent parameters that could predict the US-SH activity being the dependent factor. The receiver operating characteristics (ROC) curve was done to evaluate the diagnostic value of serum CXCL13 levels in recent onset RA. A statistical significance is considered when a P value was < 0.05 and a P value < 0.001 was considered highly significant. All data were tabulated, coded, and analyzed using STATA/ SE version 11.2 for Windows.

Results
This study included 50 RA patients; 37 (74%) females and 13 (26%) males with ages extended in the range between 19 and 63 years and a mean of 36.82 ± 13.66 years. They had disease duration ranged between 3 months and 12 months with a mean of 8.9 ± 3.09 months. Forty apparently healthy volunteers; 28 (70%) female and 12 (30%) male with ages ranging between 21 and 63 years with a mean of 39.3 ± 13.07 years were included as a control group. Both patients and controls were age (P = 0.8) and gender matched  Table 1 and Fig. 1 showed the patients' clinical, laboratory, and radiographic characteristics. DAS-28 ranged between 2.9 and 6.2 with a mean of 4.25 ± 0.99; 7 patients had low disease activity (13%), 31 (62 %) patients were moderately active, and 12 (24%) patients had high disease activity. Eighteen (36%) of patients had extra-articular manifestations in the form of pleurisy, subcutaneous nodules and Sjogren' syndrome.
RA patients were on conventional synthetic disease modifying anti-rheumatic drugs (csDMARD) and/or biologic (b) DMARDS ± oral corticosteroid ranged between 5 and 15 mg/day. Twenty-four RA patients were treated with methotrexate (MTX) alone, MTX and Sulfasalazine (3 patients), Leflunomide (5 patients), combination of MTX, Sulfasalazine, and Leflunomide (3 patients). Fifteen RA patients were receiving bDMARDs: 9 patients were receiving Adalimumab and 6 patients were on Etanercept ± MTX. Periarticular erosions were detected in the hand and feet X-ray in 17 patients (34%) while erosions were detected by MSUS examination in 23 (46%) patients.
No significant differences in the mean serum CXCL13 levels were found according to the gender (P = 0.26), extra-articular organ involvement (P = 0.75), seropositivity for RF (P = 0.23) or Anti-CCP Abs (P = 0.83) in RA patients' group but its levels were statistically significantly increased in the erosive RA patients detected by X-ray or MSUS (553.18 ± 291.85) than patients without erosions (201.07 ± 99.84) (P < 0.001) ( Table 2).
Serum CXCL13 had an AUC of 0.99 at a cutoff point of 111 pg/ml with sensitivity of 96.67% and specificity of 95.0%, positive predictive value (PPV) of 96.67%, and negative predictive value of (NPV) 95.0% as a diagnostic test of RA (Fig. 3).

Discussion
Previous articles elucidate how much chemokines and their receptors are important in RA pathogenesis. CXCL13 chemokine directs B cell chemotaxis, and it is elevated in several autoimmune diseases [18]. This study was planned to quantify the CXCL13 chemokine serum levels in recent onset RA patients and to correlate its levels with clinical and laboratory parameters of disease activity and severity, our study was extended to relate serum CXCL13 levels to more objective scores of disease activity and severity detected by GS and PD MSUS examination.
It has been previously shown that serum levels of CXCL13 levels were significantly raised in recent onset RA patients than in the healthy subjects [16,[19][20][21][22]. This is consistent with this study which showed significant increment of serum CXCL13 levels in the RA patients in contrast to the control subject. Furthermore, we have elucidated that serum CXCL13 levels were significantly elevated in active RA patients, and its levels were correlated significantly with the total DAS scores   and individual disease activity parameters as morning stiffness (MS) duration, the number of swollen and tender joints, p VAS scores for global patient assessment added to ESR values. Collectively, these results supported other investigators who found that serum CXCL13 levels were significantly associated with various measures of disease activity, such as the SJC, the disease activity global assessment, and ESR values [22][23][24].
Our results are consistent with Greisen et al. who detected raised serum CXCL13 levels in untreated early RA active patients and found that their disease activity had been reduced in those patients after treatments that supports the hypothesis that CXCL13 is a surrogate Fig. 2 a-c Dorsal longitudinal MSUS dorsal scan of radiocarpal joint in two different RA patients. a Gray-scale dorsal longitudinal musculoskeletal ultrasonographic scan of a 2nd metacarpophalangeal joint (m: metacarpal bone, p: proximal phalanx) shows synovial hypertrophy (black arrow) grade 2 and an erosion (e) in RA patient with increased serum CXCL13 level of 120 pg/ml. b Grey-scale dorsal longitudinal midline musculoskeletal ultrasonographic scan of the wrist joint (radiocarpal and midcarpal joints) of the same patient shows an erosion (arrow) in the lunate bone (r: radius bone, L:lunate bone, C: capitate bone). c Dorsal PDUS longitudinal midline musculoskeletal ultrasonographic scan of a wrist joint (radiocarpal and midcarpal joints) (r: radius bone, L: lunate bone, C: capitate bone) shows severe synovitis marked by synovial hypertrophy grade 3 and power doppler activity grade 3 in RA patient with increased serum CXCL13 level of 803 pg/ml. CXCL13: serum B lymphocyte chemoattractant protein13. PDUS power doppler ultrasound marker of disease activity and implementing a role for CXCL13 as an inflammatory marker early in the course of RA [24]. CXCL13 is superior to the conventional acute phase reactants (ESR and CRPs) in evaluating and assessing the degree of synovial inflammation. CXCL13 is constitutively generated in the lymph nodes but moreover, it originates from the inflamed synovial tissue that causes its upregulation and leads to enhancement of the lymphoid follicles′ development and germinal center reactions. On the other side, ESR and other acute phase reactants are induced by the systemic released interleukins and produced by hepatocytes only during inflammation so they are not directly linked to local joint inflammation [6,[25][26][27].
In this study, serum CXCL13 levels were significantly increased in RF and Anti-CCP Abs seropositive RA patients compared with seronegative patients. Our findings matched the previous results reported by Bugatti et al., who found that CXCL13 levels were significantly increased in Anti-CCP Abs-positive compared to seronegative patients and Jones et al, who stated that CXCL13 was associated with RF in RA patients [28,29]. Although RF and Anti-CCP Abs help to classify and address the autoimmunity in RA but unfortunately they do not counter rapidly to the changes of disease activity and are mostly used for prognostic assessment [30].
It is widely accepted now that MSUS is beneficial and especially PD techniques used to detect the active signals within the inflamed synovium added value to the clinical examination in detecting synovitis [31]. The sensitivity of MSUS to detect the erosion compared to radiography had been evaluated before as MSUS-GS detected 6.5fold more erosions in a study of RA patients with disease duration less than one year and 3.4-fold more erosions were distinguished by MSUS in the established RA patients compared with MRI [32].
In our study, periarticular erosions were detected in 10 patients (33.33%) by conventional X-ray and in 16 patients (46.67%) by MSUS. The mean serum CXCL13 level was increased in the RA patients with erosions but its levels were not correlated to Larsen scores. This is as per Greisen et al. whose study demonstrated that serum CXCL13 levels and radiographic progression were not significantly correlated [24] but are of controversial in comparison with others who found that there were definite association between serum CXCL13 levels and increased rates of joint destructions [28,33]. These different results can be explained on the base that RA patients had different disease durations in the different studies.
Assessment of the inflamed joint for the degree of activity and severity by MSUS is a non-invasive, easily used, and reproducible tool that can be used for follow up the synovitis in RA patients. It was decided and agreed by the EULAR-OMERACT US taskforce consensus to characterize gray-scale US synovitis as hypoechoic SH regardless of the presence of effusion and any grade of PD signal, they concluded that effusion occurs more frequently in certain joint depending on patients' weight and level of activity and did not add extra weighting to the definition and severity of an US-proved synovitis [15].
Consistent with previous observations, our study showed that CXCL13 levels were significantly correlated with MSUS synovial hypertrophy expressing the severity of synovitis [34,35]. Interestingly, we have not found any significant association between CXCL13 levels and PD activity scores, which is discordant with the preceding data. Several explanations are possible for such discrepancy. First, PD examination is extremely sensitive to tissue movement and pressure, additionally, with lack of standardization of methods used in PDUS examination and various settings of PDUS, the interobserver reliability remains a concern especially in image acquisition and interpretation. Second, PD activity might be affected with different treatment regimens in the different studies.
Previously, it was reported that CXCL13 expression were significantly correlated with histological grading of synovitis in early DMARD-naïve RA patients and pathogenic significance of CXCL13 in RA synovitis had been investigated in animal model of arthritis and it was found that CXCL13 neutralization had improved the disease outcomes and histological severity of synovitis [35][36][37]. Also, Manzo et al. assessed CXCL13 expression in correlation to its circulating levels and found that circulating CXCL13 was significantly correlated with synovial CXCL13 protein (r = 0.3, P = 0.04) and mRNA (r = 0.56, P = 0.02) expression in early RA [38].
In our study, regression analysis showed that serum CXCL13 level was the only factor that is strongly associated with SH score and hence it can be used to predict the severity of the actively inflamed joints which confirmed what was stated before about its usefulness in grading the severity of synovitis in comparison to ESR, C-reactive proteins, RF, or Anti-CCP abs [16].
In our study, we found that serum CXCL13 at a cutoff point of 111 pg/ml had sensitivity of 96.67% and specificity of 95.0% as a diagnostic test of RA that is in line and supports previous researcher who concluded that serum CXCL13 levels could be helpful confirming the diagnosis of active recent onset RA when they detected elevated levels in a remarkable number of recent onset seronegative RA patients [22].

Conclusion
We concluded that serum CXCL13 is correlated with clinical disease activity and MSUS disease severity; thus, Authors' contributions EAB, SAE, GAH, MAK, SHA, and MME had contributed to the conception, design of the work and definitions of the intellectual contents. EAB, MAK, SHA, and MME were concerned with literature review, run the clinical studies, collecting data, manuscript preparation, and editing. EAB and MAK were concerned with analysis, and interpretation of statistical data and writing the results. The manuscript has been read and approved by all the authors, the requirements for authorship have been met, and each author believes that the manuscript represents honest work.