Background: Fibrous dysplasia/McCune-Albright syndrome (FD/MAS) is a rare genetic bone disease caused by a somatic mutation in the GNAS gene. Currently used bone turnover markers (BTMs) do not correlate with the clinical picture and are not useful to predict or monitor therapy success. This study assessed the correlation of RANKL, OPG, RANKL/OPG ratio, IL-6 and sclerostin with the classic BTMs alkaline phosphatase (ALP), procollagen type 1 propeptide (P1NP) and beta crosslaps (CTX), with pain, skeletal burden score (SBS) and response to bisphosphonate or denosumab treatment.
Methods: Ninety-six serum samples of adult patients >18 years of age with any subtype of FD/MAS were included from the biobank facility of the Leiden University Medical Center, Center for Bone Quality between 2015 and 2021. Standard laboratory assessments were assessed as part of usual care. The concentrations of potential biomarkers RANKL, OPG, sclerostin, IL-6 were analyzed. Data on FD/MAS subtype, age, pain, treatment history and treatment response were retrieved from the electronic patient files. Baseline characteristics were summarized by descriptive statistics. Correlations of the concentrations of the potential biomarkers with classic bone turnover markers, SBS and pain scores were cross-sectionally assessed by Spearman rank order correlation. Correction for multiple testing was performed by Benjamini and Hochberg False Discovery Rate. A sensitivity analyses was performed by excluding patients with SBS below 15 and patients using antiresorptive medication at the time of blood withdrawal or within the wash-out period. In patients treated with bisphosphonates or denosumab after blood withdrawal, pre-treatment concentrations were compared in patients with and without therapy response by Mann Whitney U test.
Results: The median age of the patients was 41.2 (Q1-Q3 25.9–52.2) years, 62.5 % was female. Median SBS was 2.5 (Q1-Q3 0.5–7.8). RANKL level correlated weakly with ALP (Spearman rho 0.309, p = 0.004, n = 84), but not with P1NP or CTX. The RANKL/OPG ratio, OPG, IL-6 and sclerostin did not correlate with ALP, P1NP or CTX. None of the potential biomarkers correlated with SBS or pain. Results of the sensitivity analyses were comparable. Pre-treatment biomarker levels were similar in patients with and without improvement in pain scores following bisphosphonate therapy. Pre-treatment RANKL and sclerostin were comparable between patients with and without improvement in pain scores after denosumab therapy. Pre-treatment IL-6 level and the RANKL/OPG ratio seemed to be higher in patients with response to denosumab (IL-6: median 0.64 (Q1-Q3 0.53–0.74) pg/mL, n = 6, RANKL/OPG: median 0.062 (Q1-Q3 0.016–0.331), n = 5) compared to patients without response (IL-6: median 0.35 (0.20–0.54) pg/mL, n = 5, RANKL/OPG: 0.027 (0.024–0.046), n = 4). Pre-treatment IL-6 correlated with the improvement in maximum pain scores (rho 0.962, p < 0.001, n = 9) and average pain scores (rho 0.895, p = 0.001, n = 9) reported during denosumab therapy.
Conclusion: Increased concentrations of RANKL, IL-6, sclerostin and of the RANKL/OPG ratio do not indicate severity of FD/MAS, as no correlation was observed of these potential biomarkers with the classic BTMs and SBS. Biomarker levels did not correlate with pain and had no value in predicting bisphosphonate treatment response. These biomarkers are not superior over the currently used methods of assessing ALP, P1NP and CTX or evaluating SBS to establish disease extent or activity and provide no reliable results. Yet, possibly pre-treatment IL-6 and the RANKL/OPG ratio may have some predictive value for clinical response to denosumab. Therefore, studies investigating disease activity and treatment response should include lesional imaging and patient-reported outcome measures.