Objective Dickkopf-1 (DKK-1) is an inhibitor of osteoblastogenesis, and its lower levels are linked to new bone formation. The aim of this study was therefore to explore serum levels of DKK-1 and to evaluate DKK-1's association with the severity of spinal involvement in diffuse idiopathic skeletal hyperostosis (DISH).
Methods Serum levels of total and functional DKK-1 and C-reactive protein (CRP) were measured in 37 patients with DISH and 22 healthy age and sex-matched controls. Plain radiographs of the cervical and thoracic spine were performed, and the diagnosis of DISH was defined using the Resnick criteria. Patients were divided into three groups based on spinal involvement. Bone mineral density (BMD) and bone turnover markers were evaluated in patients with DISH.
Results The levels of total serum DKK-1 were significantly lower in patients with DISH than in healthy controls (p<0.0001). Importantly, low serum levels of DKK-1 were associated with more severe spinal involvement in DISH, independent of age, sex, disease duration, CRP, bone turnover markers or BMD. However, these findings were less significant for functional DKK-1.
Conclusion These observations indicate that DKK-1 may play a significant role in bone formation during DISH.
Statistics from Altmetric.com
Diffuse idiopathic skeletal hyperostosis (DISH) is characterised by new bone formation, calcification and ossification of the anterior longitudinal ligament of the spine and various extraspinal ligaments.1 The axial skeleton, most often the thoracic spine, is frequently involved but involvement of peripheral joints also occurs. The main clinical feature represents a limitation of spinal mobility and/or pain and stiffening of peripheral joints. DISH is a common disorder among older adults, often associated with diabetes mellitus, dyslipidaemia and hyperuricaemia; however, the cause of the disease is not completely understood.1 It has been hypothesised that structural changes of the spine are related to the effect of metabolic factors that lead to new bone formation.2 Natural inhibitors of bone formation, such as matrix Gla protein, have been suggested to be deficient in DISH patients, thus contributing to hyperostosis.2
Dickkopf-1 (DKK-1) is recognised as a key regulator of bone remodelling by inhibition of the Wnt signalling required for new bone formation.3 It was recently demonstrated that DKK-1 blockade reversed the bone-destructive pattern of an inflammatory arthritis model to the bone-forming pattern of osteoarthritis4 and promoted ankylosis of sacroiliac joints in an experimental model.5 Diarra et al4 showed that the levels of serum DKK-1 are significantly decreased in patients with ankylosing spondylitis (AS) in comparison with healthy individuals. Although DISH and AS appear as different diseases, they share a common feature of new bone formation.6 Therefore, we investigated serum levels of DKK-1 and evaluated its association with the severity of spinal involvement in DISH.
The inclusion criteria for enrolling patients in this study were age greater than 45 years and fulfilment of the definition of DISH.7 Patients with rheumatological symptoms suggestive of DISH were screened at the Institute of Rheumatology in Prague. In total, this study consisted of 37 patients with newly diagnosed DISH and 22 healthy age and sex-matched controls without a history of symptoms resembling DISH. The exclusion of potential patients with AS was based on patient history and normal radiographic findings on sacroiliac joints at baseline. To exclude osteoporosis or other metabolic osteopathies, the bone mineral density (BMD) of the lumbar spine and proximal femur was measured in patients with DISH (LUNAR, Prodigy; GE Healthcare, Madison WI, USA). All patients gave informed written consent, and the study was approved by the local ethics committee.
All patients had plain radiographs (anteroposterior and lateral) of the thoracic and cervical spine performed. The diagnosis of DISH was based on the Resnick classification criteria that require involvement of at least four contiguous vertebrae of the thoracic spine.8 Scoring of the spine was performed by an independent and experienced radiologist (JG). Patients were divided into three groups based on spinal involvement characterised by: (1) four contiguous vertebrae of the thoracic spine; (2) more than four contiguous vertebrae of the thoracic spine; and (3) more than four contiguous vertebrae of the thoracic spine plus an affected cervical spine (see supplementary figure 1A–C, available online only).
Serum biomarker measurement
Fasting blood samples were collected from all patients with DISH at the time of diagnosis and from healthy individuals. Serum samples were immediately centrifuged and stored at −80°C until analysed. Total serum DKK-1 concentrations were measured by commercially available ELISA according to the manufacturer's protocols (Biomedica, Vienna, Austria). The sensitivity for DKK-1 was determined to be 10.90 pg/l. Functional DKK-1 was measured by binding of DKK-1 to its receptor (LRP-6) using a commercially available ELISA (human Dkk-1 DuoSet ELISA kit) according to the recommendations of the manufacturer (R&D Systems, Minneapolis, Minnesota, USA). Briefly, microtitre plates were coated with 2.5 µg/ml of human LRP-6-Fc chimera (recombinant human LRP6/Fc chimera; R&D Systems) before the addition of serum samples, and detection was performed using 50 ng/ml human anti-DKK-1 antibody (R&D Systems). All measurements were performed in duplicate for each sample, and the mean value was calculated. Absorbance was detected using the Sunrise ELISA reader (Tecan, Salzburg, Austria) with 450 nm as the primary wavelength. Serum C-reactive protein (CRP) levels were measured by routine nephelometry in patients with DISH and in healthy controls. Enzyme immunoassays were used for analysis of the levels of serum bone alkaline phosphatase (Metra Biosystems, Santa Clara, California, USA), osteocalcin and deoxypyridinolin (both Diagnostic Products Corporation, Los Angeles, California, USA) in patients with DISH.
Statistical analysis was performed using GraphPad Prism 5 and SPSS 14.0. For analysis of differences between groups, a non-parametric Kruskal–Wallis (multiway analyses of variance) with Dunn's post-hoc test and a Mann-Whitney U test were performed. Spearman's rank order test (r) was used for correlations of parameters. Contingent tables were evaluated with Fisher's exact test. A p value equal to or less than 0.05 was considered statistically significant. Data are presented as mean (SD) or median (range) values.
Characteristics of patients with DISH
Of all the patients with DISH, 15 had involvement of four contiguous vertebrae of the thoracic spine (group 1), 13 had involvement of more than four contiguous vertebrae of the thoracic spine (group 2) and nine patients had additional radiographic evidence of cervical spine involvement (group 3). Although the diagnosis of DISH was established within the inclusion to this study, the median symptom duration of axial involvement was 10 years (range 2–57). In total, 10 patients had diabetes mellitus. Five of the patients with diabetes had cervical spine involvement. None of the patients with DISH had osteoporosis, and the mean BMD of the lumbar spine was 1.244±0.251 g/cm2 and of the total femur was 1.004±0.170 g/cm2. The baseline characteristics of the patients and healthy individuals are shown in table 1.
Association between low DKK-1 levels and spinal involvement in DISH
The levels of serum DKK-1 were significantly lower (p<0.0001) in patients with DISH (in male patients, median 1015.73, range 207.04–8051.77 pg/ml; in female patients, median 674.02, range 230.96–4446.867 pg/ml; NS) compared with those in healthy controls (in male patients, 4569.49, range 3342.05–8988.51 pg/ml; in female patients, 3331.35, range 1371.08–7022.63 pg/ml; NS; figure 1). More interestingly, we found that the levels of serum DKK-1 were lower in patients with more severe spinal involvement (figure 2A). There were significant differences in the levels of DKK-1 among the groups (p=0.008). Dunn's post-hoc analysis showed that patients with DISH from group 3 (most severe spinal involvement) had significantly lower levels of DKK-1 compared with patients from group 1 (least severe spinal involvement) (median 536.77, range 207.04–968.58 vs median 1466.29, range 434.16–8051.77 pg/ml; p<0.001).
To analyse functional DKK-1, we used functional ELISA measuring the binding of DKK-1 to its receptor LRP-6. The levels of functional DKK-1 positively correlated with the levels of total DKK-1 (r=0.743, p<0.0001). In contrast to total DKK-1 levels, we found only a trend towards the lower levels of functional DKK-1 in patients with DISH (median 542.97, range 9.29–5744.60 pg/ml) compared with healthy controls (median 861.40, range 91.00–4253.50 pg/ml; p=0.071). However, the difference in the levels of functional DKK-1 among the groups of patients with DISH remained significant (p=0.044), with lower DKK-1 levels in patients with more severe spinal involvement (group 3, median 347.33, range 9.29–767.09 vs group 2, median 729.58, range 310.0–2273.50 vs group 1, 660.65, range 40.7–5744.60 pg/ml). Using Dunn's post-hoc analysis, however, the differences between studied groups was no longer significant.
DKK-1 levels, CRP and bone turnover markers
Serum levels of CRP were significantly higher (p<0.0001) in patients with DISH (median 4.10, range 0.70–22.90 mg/l) than in healthy controls (median 1.60, range 0.48–7.20 mg/l). The levels of CRP increased with increasing severity of spinal involvement (p=0.034; figure 2B). Patients with DISH from group 1 had significantly lower (Mann–Whitney p=0.023) levels of serum CRP (median 3.85, range 0.700–6.90 mg/l) compared with patients from group 3 (median 6.50, range 3.30–22.90 mg/l). Using Dunn's post-hoc analysis, however, this association no longer remained significant. The levels of DKK-1 were not associated with age, disease duration or the presence of diabetes mellitus. In addition, serum levels of DKK-1 were not associated with the levels of serum CRP and bone turnover markers or BMD in any patients with DISH, but they positively correlated with serum osteocalcin (r=0.527, p=0.025) and bone alkaline phosphatase (r=0.561, p=0.015) in female patients with DISH, which was no longer significant when corrected for multiple testing.
To our knowledge, this is the first study showing that total circulating DKK-1 levels are lower in patients with DISH and that they are associated with the severity of spinal involvement.
Higher levels of DKK-1 are related to bone resorption, whereas lower levels are linked to new bone formation.9 Patients with multiple myeloma–characteristic lytic bone disease have elevated circulating DKK-1 levels.10 Serum levels of DKK-1 are also elevated in patients with rheumatoid arthritis (RA), a typical autoimmune disease characterised by the development of bone erosions.4 11 In addition, serum DKK-1 levels correlate with RA disease activity.4 11 Conversely, rather lower levels of functional DKK-1 and no association with disease activity have been found in patients with AS.4 In agreement with these data, we found lower levels of total serum DKK-1 in patients with DISH. This finding suggests that DKK-1 is a potential link between common features of new bone formation and these two diseases. On the other hand, a recent study showed that total levels of DKK-1 are rather high, but functional DKK-1 is low in patients with AS.12 The authors suggest that DKK-1 is dysfunctional in AS. In our study, the levels of functional DKK-1 in patients with DISH were numerically lower, but did not differ statistically from those in healthy controls, which is in agreement with a recent report by Aeberli et al.13 This finding might, on the other hand, question the role of DKK-1 in the development of hyperostosis in DISH.
However, we showed that patients with DISH and with the most advanced disease with cervical spine involvement had significantly lower total and also functional serum DKK-1 levels than patients with less advanced spinal involvement. The finding of lower DKK-1 in DISH patients may be reciprocal to the finding that myeloma patients without lytic lesions had significantly lower DKK-1 levels than patients with lytic bone disease.10 Consistently, Garnero et al14 found that elevated DKK-1 levels are associated with a higher risk of the progression of bone erosion in patients with RA. Based on these data, it can be suggested that lower circulating DKK-1 levels in patients with DISH reflect the bone anabolic phenotype of the disease.
Consistent with previous data from patients with AS,4 serum levels of DKK-1 were not associated with the level of serum CRP, either in patients with DISH or in healthy controls. It is interesting to note that DKK-1 may be associated with markers of joint damage. In patients with RA and osteoarthritis of the knee, the opposite relationship between serum levels of DKK-1 and cartilage breakdown exists.15 In our study, we tested the hypothesis that serum DKK-1 levels might be associated with bone turnover markers. However, we have not observed serum DKK-1 reflecting bone turnover or bone mineral content in patients with DISH. It could be suggested that this discrepancy can be caused by different pathogenesis and/or mechanisms controlling bone metabolism that are involved in these diseases.
This study has some limitations. We cannot exclude spinal involvement in some individuals in the control group because these patients were not screened for DISH or spondylosis, although they were asymptomatic. The radiographic evaluation of extraspinal involvement has not been performed. Furthermore, we did not perform radiographs of the lumbar spine, although it is less often involved in patients with DISH; therefore, we believe it is a limitation with a minor impact on the results of this study.
In conclusion, the data from this study show that low serum levels of DKK-1 are associated with more severe spinal involvement in DISH. We suggest that DKK-1 may play a significant role in bone formation during DISH.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Web Only Data - This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Funding This study was supported by the Ministry of Health of the Czech Republic, research project no 00023728.
Competing interests None.
Patient consent Obtained.
Ethics approval Ethics approval was obtained from the local ethics committee at the Institute of Rheumatology, Prague.
Provenance and peer review Not commissioned; internally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.