Denosumab for Fibrous Dysplasia: Promising, but Questions Remain

Fibrous dysplasia (FD) is a rare skeletal dysplasia caused by somatic, gain-of-function mutations in the cAMP-regulating, Gαs transcript of GNAS. The spectrum of disease burden is broad; depending on the extent and location of the disease, patients can experience pain, fractures, impaired mobility, and loss of function and quality of life (1). The pathophysiological driver is cAMP-mediated skeletal stem cell dysfunction, wherein normal bone and marrow are replaced by fibro-osseous tissue and mechanically unsound woven and undermineralized bone, rich in osteoclasts but lacking adipocytes and hematopoiesis (2). A central feature is cAMP-dependent RANKL expression by FD cells, which stimulate osteoclastogenesis by binding to cell surface RANK in precursor cells. The importance of osteoclasts in FD pathogenesis is supported by their prominence and ectopic location in areas of rapidly remodeling, highly active lesions (3). Targeting osteoclast-mediated high turnover in FD with bisphosphonates has shown efficacy in relieving pain in some patients, but has not demonstrated a direct effect on FD in terms of lesion appearance, progression, or regression (4). Lack of a direct effect by bisphosphonates may be because their action requires incorporation into the mineralizing matrix, which is significantly diminished in FD.

Denosumab, which is a potent monoclonal antibody drug that blocks RANKL/RANK interaction, is a logical choice for FD because it can bypass the need for matrix incorporation to directly target ectopic osteoclasts in FD lesions. Support for a direct role of osteoclasts in FD pathogenesis comes from findings in a recent study in a mouse model of FD. Treatment with an anti-mouse RANKL antibody (a denosumab analogue) not only prevented lesion formation but promoted differentiation of stromal cells to bone-forming osteoblasts that produced highly mineralized lamellar bone (5). Lending support for direct effects in patients were the findings from the first patient treated with denosumab that showed inhibition of FD lesion growth (6). However, in both the mice and the patient, discontinuation resulted in rebound in bone turnover to levels above pretreatment; in the patient, rebound was associated with life-threatening hypercalcemia.

That denosumab discontinuation was associated with a rebound increase in bone turnover was evident even in the earliest clinical studies for osteoporosis, but that it can result in significant morbidity has only recently been appreciated. In fact, this phenomenon is still not mentioned in the package insert of denosumab products. It appears the higher the pretreatment bone turnover, the greater the degree of rebound and risk for morbidity, an important consideration when using these drugs in high turnover diseases such as FD. Discontinuation rebound, together with the risk for osteonecrosis of the jaw, which has been reported in FD patients treated with bisphosphonates, and which is greater with denosumab, are the 2 main safety issues associated with denosumab use in FD—a concern magnified by the fact FD appears and expands in childhood and is a lifelong disease. An additional important concern is that FD is a mosaic disease and treatment will affect normal bone as well. In patients with FD in whom the skeletal disease burden is low (a skeletal disease burden score < 10), changes in bone turnover markers that occur during treatment likely represent the effect of the drug on non-FD normal bone; tissue which is at risk for antiresorptive-associated morbidity such as osteonecrosis of the jaw and osteopetrosis-like changes.

Considering the above, questions to be answered before denosumab can be recommended for treating FD are (1) efficacy: Does treatment relive pain in the subset of patients with FD who experience pain? Does it prevent lesion formation, and/or growth, and/or cause lesion mineralization? (2) safety: Can discontinuation-associated rebound osteonecrosis of the jaw or osteopetrosis-like changes in growing children be prevented or controlled? (3) regimen: What dose and interval between doses? In other words, who should we treat, and how to treat them?

The accompanying case series by Majoor begins to answer question #3 about regimen (7). Figure 2, which shows the results of almost 15 years of treatment of a patient with significant FD, provides useful data. We do not know if she had pain (surprisingly, many patients with significant FD burden and high bone turnover do not experience pain), and/or if the previous bisphosphonate treatment affected her pain. But we do get the sense that an interval of 3 months and a dose of 120 mg may be needed to decrease bone turnover in patients with significant disease. Unfortunately, not in this patient, nor in any of the other patients presented in the report, can we adequately compare the effects of bisphosphonates to denosumab. The specific regimens are not shown, and olpadronate, a drug used almost exclusively only in The Netherlands and never compared with other bisphosphonates, was the drug most used. It is difficult to generalize from this series; it is retrospective, open label, patients were heterogeneous in terms of spectrum of disease (one-third had a disease burden score < 10, and 1 patient had nearly panostotic disease), treatment regimens were heterogeneous, pain was not measured, neither anatomical nor histological data were assessed, and discontinuation was not observed. What may have been learned was lost by presenting the results as grouped data, a significant flaw in the analysis of a small group with heterogenous characteristics.

As the authors astutely point out: “Until additional data become available from randomized controlled trials, caution should be exerted in the treatment of patients with FD/MAS [McCune-Albright syndrome] with denosumab.” So where do we stand? From clinical and preclinical studies we know RANKL is prominent in FD (3,5); from preclinical studies, it appears that RANKL-stimulated osteoclasts are likely pathogenic, and blocking RANKL/RANK interaction allows for the conversion of FD to lamellar bone (5). From clinical studies, we have a framework for an effective dose and regimen (7), and from an ongoing study (https://clinicaltrials.gov/ct2/show/NCT03571191), we will likely have a better handle on discontinuation rebound and histologic effects on FD tissue. If the preclinical study translates and we can prevent lesion formation and convert active lesions to bone, we must begin to study denosumab in childhood, this is when lesions are forming and active. And we must develop a strategy to prevent discontinuation-associated resorption of the bone formed during denosumab treatment. Loading the bone formed during denosumab treatment with bisphosphonates has been suggested as a way to prevent discontinuation-associated resorption (8). Is this feasible or safe, especially in children? The road forward is mapped; the journey will require the organized efforts of scientists and clinicians, internists and pediatricians, but for the first time a treatment for FD with mechanistic promise may be in the offing.

Additional Information

Disclosure Summary: The NIDCR receives support from Amgen for a study of denosumab in fibrous dysplasia of bone.