Background
The spondyloarthropathies (SpA) form a class of joint diseases which specifically affect the vertebral column. Ankylosing spondylitis (AS), the prototypic form of SpA, is a chronic inflammatory arthritis of the primarily axial skeleton. One of the most debilitating aspects of AS is the progression from inflammation to bone formation, where inflammation is triggered by an unknown mechanism at the site of tendon and ligament attachments to bone, resulting in enthesitis, followed by the formation of bony projections (syndesmophytes) which ultimately may join and result in ankylosis. Inflammation can be well controlled in many patients by anti-TNF therapy, however ankylosis may still progress [
1‐
4]. Although a recent Cochrane review determined that the effects of long term high-dose NSAID treatment on syndesmophyte formation was unclear [
5], other recent studies have indicated chronic high-dose NSAID treatment [
6‐
8] or long-term TNF blockade [
3,
9] may be effective at slowing progression. No treatment to date has managed to halt or reverse radiographic progression in AS.
A major factor in the poor treatment options available for advanced AS is the lack of understanding of the molecular mechanisms driving disease progression. Several recent studies in both patients and mouse models have identified the Wingless (Wnt) pathway, key for bone development and homeostasis, as disturbed in AS. Sclerostin (SOST) is specifically expressed in osteocytes and chondrocytes providing skeletal tissue specific inhibition of Wnt signalling. Inactivating mutations in SOST which lead to increased Wnt signalling result in increased bone mass and bone strength, as demonstrated in both mouse models and human disease [
10].
SOST’s tissue-specific expression makes it an attractive therapeutic target in bone disease. In patients with SOST-inactivating mutations, the only phenotypes which develop are a direct consequence of high bone mass [
11]. Antibody therapies to SOST have recently successfully completed phase 2 clinical trials, demonstrating an increase in bone density and a decrease in bone turnover markers in osteoporosis patients [
12,
13]. In AS, which exhibits excess bone formation and where dysregulated Wnt signalling has been established, enhancing SOST activity presents an intriguing therapeutic possibility. To date no studies have investigated treatment with exogenous SOST to inhibit bone formation. In this report we have tested the effects of treatment with recombinant SOST to reduce osteoproliferation in the proteoglycan-induced spondylitis (PGISp) mouse model of ankylosing spondylitis.
Discussion
No treatment currently available is able to reverse or significantly slow the excessive bone formation which occurs in AS. Recent studies have demonstrated that Wnt signalling appears to be up-regulated in AS evidenced by decreased levels of the bone-associated Wnt inhibitor SOST in both human and mouse studies [
15,
21,
22]. Inhibiting Wnt signalling through SOST supplementation therefore presents a potential therapeutic target for the specific treatment of this bone formation. However, daily treatment with rSOST did not alter disease progression or severity in the peripheral joints or axial skeleton joints of PGISp mice.
There are a number of factors which may have contributed to the lack of efficacy of rSOST at inhibiting excess tissue formation in this study, including efficacy of a human recombinant SOST in mice, low stability of rSOST, failure of rSOST to reach joints, or insufficient dosage. Human SOST has previously demonstrated activity in a mouse system; the human SOST transgenic mouse, in which a bacterial artificial chromosome (BAC) containing the human
SOST gene was inserted, is osteopenic [
11]. In this model overexpression of human
SOST was effective at blocking Wnt signalling, resulting in a marked decrease in osteoblast activity and thus bone formation.
Recombinant SOST degradation appeared to be relatively slow in PGISp mice, as circulating SOST levels were significantly increased 8 h after injection and remained elevated after 24 h. Further over an 8 week injection protocol it appeared that circulating levels of SOST were increasing suggesting an accumulation of the protein. This suggests that injection of rSOST might be a viable approach if therapeutic efficacy could be demonstrated. Alternatively, the circulating SOST increases might be due to increased endogenous expression. Unfortunately, the ELISA used did not differentiate between human and mouse SOST isoforms.
rSOST may have been bound to other proteins in the bloodstream which prevented it from exerting its actions at the joint. However, changes in SOST levels within the joints were observed. rSOST treatment increased the number of SOST positive osteocytes in unaffected joints. In vitro, application of recombinant human SOST results in decreased expression of SOST in human pre-osteocyte cells [
23]. However, the same levels of recombinant human SOST caused increased expression of SOST in less mature cultures thus application of exogenous human SOST in our study may have influenced endogenous mouse SOST levels.
The failure of injected SOST to have any effect on bone formation in this experiment does not exclude SOST as a potential therapeutic in AS. There is strong suggestive evidence from both mouse models and human patients that regulation of SOST is an important factor in bone formation in AS. Markedly reduced SOST expression has been shown both in serum and biopsies from AS patients [
21,
22]. High SOST levels occur in hTNFtg mice, as TNF-α stimulates SOST expression [
24,
25] but if the mice are treated with Dikkopf-1 neutralising antibodies, which also results in reduced levels of SOST, spondylitis occurs [
24]. Additionally, SOST expression is downregulated in the intervertebral joints in proteoglycan induced spondylitis (PGISp) mice [
15]. SOST may need to be further stabilized or targeted to the bone to increase its potency as a treatment. SOST’s bone specificity in non-pathological situations comes from its specific expression in osteoblasts and osteocytes rather than it having any endogenous bone-targeting properties. One possible strategy is to complex SOST with a bone-targeting moiety, such as a bisphosphonate, to allow the specific accumulation of SOST in the bone. This approach of attaching a therapeutic to a bisphosphonate for bone targeting has previously proven successful in a mouse model of metastatic breast cancer [
26].
Alternatively, the attachment of an FcD
10 motif to a recombinant human alkaline phosphatase protein has shown success in the treatment of hypophosphatasia [
27,
28]. The FcD
10 motif contains the human IgG
1 Fc domain and a deca-aspartate motif. This motif created a strong affinity for hydroxyapatite while not reducing the activity of the enzyme. This targeting strategy has been successful in ameliorating hypomineralization in hypophosphatasia patients where direct infusion or injection of the alkaline phosphatase enzyme have previously failed [
29].
This is the first study to report the injection of rSOST in a mouse model. There have been very few reports of in vivo studies targeting the skeleton utilising recombinant proteins. The dosage of 2.5 µg is similar to that used in previous studies. The lack of efficacy of rSOST to alter the disease course in treated mice may have therefore been a contributed to by an insufficient dosage administered. Future studies should investigate the impact of higher dosages, including safety and toxicity profiling.
Encouragingly, a recent study examining Wnt inhibition through inhibiting β-catenin function in murine systemic sclerosis, another mouse model of rheumatic disease, demonstrated efficacy in reducing and reversing the primary disease symptom of fibrosis. Despite the administration of the Wnt inhibitors systemically, no other side effects such as hair loss or gut abnormalities were seen [
30].
The aim of this study was to specifically focus on the potential role of Wnt signalling inhibition in a mouse model of AS. Current knowledge suggests that Wnt signalling will play a role in AS through the osteoproliferative phase of the disease. The PGISp mouse is a good choice for such a study as it has been well established that this model exhibits axial osteoproliferation similar to the human condition. However there are a number of differences between the PGISp mouse model and AS. Disease in PGISp mice favours females rather than males and there is no strong evidence for a significant role for HLA-B27. A different model could be used to examine the effect of rSOST on bone formation, such as transgenic HLA-B27 rats which spontaneously develop spondylitis and also exhibit axial osteoproliferation [
31]. Alternatively, a non-AS model could be used, such as the spinal cord injury induced model of neurological heterotopic ossification where local inflammation leads to ectopic bone formation [
32]. Application of exogenous rSOST in these models may lead to additional insights into the role and importance of Wnt signalling in pathological bone formation and how this may be manipulated in a therapeutic setting.
Competing interests
MK is a current employee of the Novartis Institutes for Biomedical Research. None of the other authors report any competing interests.
Authors’ contributions
KRH designed and performed the experiments, and prepared the manuscript. HWT performed the experiments. MK and TTG provided reagents and the animal model. MAB and GPT prepared the manuscript and designed the experiments. All authors read and approved the final manuscript.