Is it possible to predict a risk of osteoporosis in patients with juvenile idiopathic arthritis ? A study of serum levels of bone turnover markers

Background: Low bone mineral density is a common finding in children with systemic connective tissue diseases, including juvenile idiopathic arthritis (JIA). The influence of the ongoing process of bone remodeling on the disease course merits further investigation. The aim of this study was to assess the clinical relevance of markers of bone turnover and their potential role as predictors of higher fracture risk and, by extension, risk of osteoporosis. Materials and Methods: Blood samples were collected from 59 patients diagnosed with JIA in order to determine serum levels of the following markers of bone turnover: Beta-Crosslaps, osteocalcin, bone alkaline phosphatase, osteoprotegerin and receptor activator for nuclear factor kappa-B ligand. The values were analyzed with laboratory parameters and results of dual X-ray absorptiometry (DXA). Results: Osteoprotegerin and bone alkaline phosphatase levels were age-dependent. BetaCrosslaps values were significantly higher in patients with positive JADAS27 score (p=0.0410). Osteoprotegerin levels were higher in patients treated with biological agents than only with disease-modifying anti-rheumatic drugs (p=0.0273). There was no relation between markers of bone turnover and sex, DXA results, dosage of glucocorticosteroids and disease duration. Conclusions: The authors postulate performing DXA measurements every 6 months in patients with higher disease activity. The potential lower fracture risk in children with JIA within biological treatment needs further assessment. Ageand sex-adjusted reference rates of bone turnover markers need to be developed for Central European patients in order to assess individual values properly.


BACKGROUND
Low bone mineral density (BMD) is an everyday finding in clinical practice of pediatric rheumatologist.Children diagnosed with systemic connective tissue diseases, including juvenile idiopathic arthritis (JIA), juvenile systemic lupus erythematosus (SLE), and systemic vasculitides are normally expected to develop a decrease in bone mass by both, the primary and secondary mechanisms.These include: negative influence of proinflammatory cytokines -specifically: tumor necrosis factor alpha (TNF-α), interleukin (IL) IL-1, IL-6, IL8, IL-12, IL15, IL-18 and granulo-macrophage colony stimulating factor (GM-CSF) -on bone metabolism (de Jager et al., 2007;Jelusic et al., 2007;Kim & Moudgil, 2008;Macaubas et al., 2009); nutritional insufficiency of vitamin D3 and calcium; low physical activity, which is regularly caused by avoiding of every risk of trauma.Low bone mass may also constitute one of the side effects of pharmacotherapy, mainly usage of glucocorticosteroids (GCS).Inhibition of bone formation is the principal mechanism in which GCS affect bone tissue.Even moderate doses of GCS limit both, proliferation and maturation of osteoblast progenitor cells (osteoblastogenesis), and shorten the survival time of osteoblasts by induction of apoptosis (Weinstein et al., 1998).These effects result in a significant depletion in the number of osteoblasts.GCS usage may also affect excretion of several cytokines and growth stimulating factors, including: IL-6, IL-11 (Weinstein et al., 1998), insulin-like growth factors I and II (Delany et al., 2001), and insulin-like growth factor binding proteins (Okazaki et al., 1994).Consequently, synthesis of type I collagen and osteocalcin is impaired, which results in lower production of the osteoid (unmineralized portion of bone matrix).Additionally, GCS increase expression of collagenase-3 (matrix metallopeptidase 13) and thereby exacerbate osteoid deficiency (Canalis & Delany, 2002).By another mechanism, GCS suppress expression of the gene for osteoprotegerin (OPG) and in that way escalate osteoclastogenesis and breakdown of bone tissue (Hofbauer, 1999).Post-steroid insufficiency of estrogens, testosterone and suprarenal androgens caused by inhibition of the gonadotrope and corticotrope cells in the pituitary gland also results in increased bone resorption (Lukert & Raisz, 1994;Hampson et al., 2002).Induction of apoptosis in osteocytes provoked by GCS may even lead to aseptic osteonecrosis (Weinstein et al., 2000).
GCS-induced lowering of BMD may lead to secondary osteoporosis which is defined as impaired bone me-chanical resistance and elevated susceptibility to fractures caused by disrupted microarchitecture of the bone.According to diagnostic criteria developed by International Society for Clinical Densitometry (ISCD) (Gordon et al., 2014), in the absence of vertebral compression fractures, both -a decreased BMD and a positive fracture history are necessary to make an osteoporosis diagnosis in children (Fig. 1).Dual X-ray absorptiometry (DXA) is a widely agreed method to assess BMD in both, the diagnostic process and monitoring of effectiveness of treatment.Because of continuous growth and alterations in ratios between tissues, age-adjusted reference rates of BMD and composition of soft tissues are essential in proper interpretation of single and repeated tests (Ward et al., 2009).Lumbar spine and total body less head (TBLH) measurements are recommended to assess BMD in the pediatric population (Gordon et al., 2014).
The human skeletal system undergoes continuous metabolic remodeling through processes of bone formation and its resorption.Alterations in serum concentrations of biochemical markers of bone turnover may indicate whether the bone metabolism has been dysregulated (Pereira et al., 1999;Skowronska-Jozwiak & Lorenc, 2006).Van der Sluis and coworkers ( Van der Sluis et al., 2002) postulated that the measurements of these markers may constitute independent predictors of fracture risk.Betaisomerized carboxy terminal telopeptide of type I collagen, frequently referred to as BetaCrosslaps (bCTx), is released to the bloodstream within bone resorption as a product of collagen's degradation.Gorska and coworkers (Gorska et al., 2008) reported elevated bCTx concentrations in JIA patients.Although bCTx is considered as a bone-specific marker, leukocyte lysate may decrease its serum concentration.Thus, serum should be separated from the whole blood immediately after coagulation and then stored until analysis (Herrmann et al., 2002).Bone Alkaline Phosphatase (BAP) is a protein located on cellular membrane of osteoblasts and hypertrophic chondrocytes.BAP is secreted within bone mineralization, therefore it may be considered as a marker of bone formation (Vasikaran et al., 2011).Caparbo and coworkers (Caparbo et al., 2009) found a decrease in BAP levels in JIA patients.Osteocalcin (OC, also known as bone gamma-carboxyglutamic acid-containing protein, BGP) is one of non-collagenic proteins of bone matrix synthesized by osteoblasts, odontoblasts and hypertrophic chondrocytes (Vasikaran et al., 2011).The function of OC is not fully elucidated yet.As Ducy and coworkers (Ducy et al., 1996) observed increased bone formation in OC-deficient mice, OC is postulated to be considered as the marker of bone turnover rather than bone formation.There are conflicting findings in the literature on the OC level in JIA patients (Caparbo et al., 2009;Gorska et al., 2008).Active osteoblasts release a ligand protein (RANKL) for receptor activator for nuclear factor kappa-B (RANK) expressed on the membranes of preosteoclasts in order to promote their maturation to osteoclasts and stimulate bone resorption (Molines et al., 2010).RANKL may also be expressed in activated T lymphocytes and synovial cells after their stimulation by proinflammatory cytokines (Boyle et al., 2003).Osteoprotegerin (OPG) is a soluble decoy receptor for RANKL (Zdzisinska & Kandefer-Szerszen, 2006) released by osteoblasts in order to keep balance of the bone metabolism by neutralization of RANKL signaling pathway.OPG is postulated to prevent cartilage degradation through reduction of chondrocyte apoptosis (Shimizu et al., 2007).
General objective of this study was to assess the clinical significance of measuring serum concentrations of the listed markers of bone turnover in patients with JIA.Specific aims included evaluation of GCS' influence on the bone metabolism and differentiation of patients with higher fracture risk and, by extension, risk of osteoporosis.

Study participants.
The study involved 59 patients already diagnosed with JIA, who were admitted to the Department of Pediatric Cardiology and Rheumatology, Medical University of Lodz, Poland between September 2014 and May 2017.The group included 40 girls and 19 boys between the ages of 5 and 18 years (mean age 12.7 ± 3.9 years).Patients' records were thoroughly reviewed in order to obtain demographic data (weight, height, body mass index, body surface area), previously assessed serological and genetic markers (rheumatoid factor, anti-cyclic citrullinated peptide autoantibodies, human leukocyte antigen B27) and treatment history (dosage and duration of the GCS therapy, usage of disease-modifying anti-rheumatic drugs).Patients diagnosed with JIA were also assessed according to the Juvenile Arthritis Disease Activity Score 27-joint Reduced Count (JADAS27) and subtype of the disease (Petty, 2001;Bazso et al., 2009).
Assessment of bone mineral density.Horizon-Wi Densitometer (Hologic Inc, Canada) was utilized for DXA examinations.All scans and analyses were performed by a physician with a certification in the field of bone densitometry and morphometry (MJ-Sz).The obtained results were interpreted according to the ISCD Official Pediatric Positions (Gordon et al., 2014).Measurements of TBLH and lumbar spine BMD included in the database were expressed as Z-Score according to the Bone Mineral Density in Childhood Study (BMDCS) database provided by the manufacturer.
Statistical analysis.All statistical calculations were carried out using Statistica 13.1 software (Statsoft Polska, Krakow, Poland).The values were expressed as mean ± standard deviation (S.D.).The Shapiro-Wilk test was utilized to verify the normality of continuous variables.Spearman's rank correlation coefficients were calculated for variables not normally distributed.The Mann-Whitney U test was used for group comparisons.P values below 0.05 were considered significant.
The study was approved by the Bioethics Committee of the Medical University of Lodz (Approval No. RNN/520/14/KB).

RESULTS
Characteristics of the study group are shown in Table 1.All patients were previously diagnosed with JIA.
The mean age at diagnosis was 9.0 ± 4.3 years and the  mean disease duration was 3.7 ± 3.3 years (55.9% were diagnosed more than 2 years before evaluation).54 patients (91.5%) had positive history of GCS intake.Cumulative GCS dose converted to prednisone-equivalent dose was 9.8 ± 6.0 mg/kg of body weight.
Regarding activity of JIA at evaluation, 74.6% of patients had positive JADAS27 score.CRP and ESR levels were elevated in 27.1% and 37.3% of patients, respectively.In respect to DXA results, mean TBLH Z-Score was -0.86 ± 1.25 (16.9% of patients with a result below 2.0), mean lumbar spine Z-Score was -0.86 ± 1.38 (20.3% of patients with a result below 2.0).
Mean values of the tested bone turnover markers are presented in Table 2.There was no association between concentrations of markers and DXA results or levels of the listed parameters: calcium, magnesium, phosphate, 25-hydroxyvitamin D3 and 1,25-dihydroxyvitamin D3, parathormone and thyroid-stimulating hormone.OPG and BAP levels were negatively correlated with age at evaluation (OPG: r=-0.295p=0.0286;BAP: r=0.514 p=0.000044) and age at diagnosis (OPG: r=-0.389p=0.00329;BAP: r=-0.311p=0.0187).Correlation between OC and age at evaluation was marginally significant (r=0.260p=0.0573).Concentrations of bCTx were independent of age.RANKL was excluded from further analysis as it turned out to be the only marker without any relation to the remaining ones.
Significant results of group comparisons are illustrated in Fig. 2. Concentrations of bCTx were higher in patients with positive JADAS27 score (p=0.0410).Concentrations of OPG were higher in patients treated with biological agents than those treated only with disease-modifying antirheumatic drugs (p=0.0273).Lumbar spine Z-Score was higher in a small group of patients without a history of GCS intake (p=0.0282).Although all of the remaining parameters included in the dataset were involved in group comparisons, they did not show statistically significant differences.

DISCUSSION
Risk of osteoporosis in children with JIA is difficult to assess.While 16.9% (TBLH) and 20.3% (lumbar spine) of patients had BMD Z-Score below -2.0,only one patient included in the study had a positive history of vertebral compression fracture and required further treatment with disodium pamidronate.Hence, serum concentrations of bone turnover markers were measured in the study group in order to evaluate their clinical relevance.In a study conducted by Masi and coworkers (Masi et al., 2004), serum OPG levels were higher in children with polyarticular and extended oligoarticular JIA than in healthy controls and patients with oligoarticular-persistent disease.Conversely, OPG levels were lower in patients with early JIA (disease duration 19.4 ± 12.3 months) than in healthy individuals according to Lien and coworkers (Lien et al., 2010).Comparing results of the recent study to the findings of Gajewska and coworkers (Gajewska et al., 2006), who evaluated markers of bone turnover in healthy Polish children, JIA patients had higher concentrations of OPG than the cited reference group; postpubertal healthy children had lower bCTx levels than JIA patients.Lambrinoudaki and coworkers (Lambrinoudaki et al., 2013) reported OPG and RANKL levels in Greek children, including patients with diabetes mellitus type 1 (DM1) and healthy controls.Both parameters were lower in these groups when compared to JIA patients included in the recent study.Levels of the measured parameters were independent of sex, however, previous reports found significantly higher levels of OC  in females than males in healthy children (Csakvary et al., 2013) and DM1 patients (Napoli et al., 2013).Age of study participants was negatively correlated with OPG, BAP and nearly significantly correlated with OC.Such relations were also observed in DM1 patients and healthy controls (Lambrinoudaki et al., 2013;Napoli et al., 2013).However, markers of bone turnover are not specific for bone remodeling only.Their concentrations may be influenced by primary bone modeling as well (von Scheven et al., 2014), therefore interpretation of the obtained results without proper age-and sex-adjusted reference rates may become challenging.
Concentrations of bCTx in the recent study were higher in patients with positive JADAS27 score, in other words, with higher disease activity.Similarly, bCTx levels were positively associated with disease activity in young male adults newly diagnosed with SLE (Guo et al., 2017).Nevertheless, large population-based study of German children showed seasonal variation of bCTx levels (Thiering et al., 2015), which was also approaching significance in the recent study group (data not shown).Therefore, clinical significance of bCTx measurements remains questionable.
Serum OPG levels were higher in children treated with biological agents -etanercept, adalimumab and tocilizumab -when compared to patients solely receiving disease-modifying anti-rheumatic drugs.Billiau and coworkers (Billiau et al., 2010) observed a nearly significant (p=0.07)increase in OPG concentrations after 6 months of treatment with etanercept and methotrexate, unlike in patients treated only with methotrexate.These results may indicate efficient blocking of anti-osteoblastic effect of TNF-α and IL-6.When administered in vitro, TNF-α inhibits synthesis of type I collagen, OC expression and BAP activity (Corrado et al., 2017), whereas IL-6 interacts negatively with the Wnt signaling pathway (Malysheva et al., 2016).As higher OPG levels may prevent skeletal system from bone destruction through inhibition of osteoclast differentiation, activation and stimulation of their apoptosis (Romas et al., 2002), a potential protective role of biological therapy needs to be further evaluated in the future.
Regarding OPG/RANKL axis, OPG is postulated to be a more relevant marker of the system activity (Lambrinoudaki et al., 2013).Sarma and coworkers (Sarma et al., 2008) suggested that RANKL levels increase in the early phase of JIA and persist elevated throughout the course of the disease.According to Lien and coworkers (Lien et al., 2010) higher RANKL concentrations are predictors of a more severe disease.RANKL levels were also elevated in patients with active polyarticular JIA with bone erosions (Spelling et al., 2008).Similarly to the recent study, RANKL levels were not associated with any factor in DM1 patients (Lambrinoudaki et al., 2013).
Although all study participants had positive history of vitamin D intake, 59.3% of patients had serum 25-hydroxyvitamin D3 concentrations below 30 ng/ml.However, there was no association between markers of bone turnover and vitamin D. These findings are consistent with other authors' results (Schou et al., 2003;Ginty et al., 2004).Influence of vitamin D on bone development in adolescents may be dose-dependent, though.In a study of HIV-infected youth, decrease in the levels of markers of bone resorption was reported only in a group receiving monthly 120 000 IU (Eckard et al., 2017).
In the study presented here, there were no associations between markers of bone turnover and DXA results.However, Masi and coworkers (Masi et al., 2004) found positive correlation between BMD and OPG lev-els in JIA patients.Furthermore, BMD values were positively predicted by OC concentrations in healthy Hungarian children (Csakvary et al., 2013).
Influence of using GCS on bone turnover needs to be further discussed.Wasilewska and coworkers (Wasilewska et al., 2010) suggested that long-term use of GCS leads to a dose-dependent increase in the RANKL level.Although 91.5% of patients were treated with GCS, concentrations of markers of bone turnover were independent of the GCS dose in the study group.However, a GCS-associated reduction in the OC levels was observed in SLE patients (Dhillon et al., 1990), and a decrease in OPG concentrations in rheumatoid arthritis patients treated with prednisolone was reported (Engvall et al., 2013).
A lack of age-and sex-adjusted reference rates was the main limitation of our study.Age-dependent differences in the mentioned parameters might have been evaluated more accurately with Tanner stages.A small group of JIA patients who were not treated with GCS was not enough to perform a reliable analysis of GCS' influence on bone turnover.

CONCLUSIONS
Findings of this study suggest a potential association between the disease activity and a higher fracture risk.Therefore, the authors postulate performing DXA tests every 6 months in patients with higher disease activity in order to assess the dynamics of bone mass lowering.The potential lower fracture risk in children with JIA treated with biological agents needs further evaluation.A population-based study of healthy controls needs to be performed in order to assess concentrations of markers of bone turnover more properly.

Figure 1 .
Figure 1.Diagnostic criteria of osteoporosis according to the International Society for Clinical Densitometry (ISCD) (Gordon et al., 2014)