Title: 1,25‐Dihydroxyvitamin D as Monotherapy for XLH: Back to the Future?
Abstract: The identification in the year 2000 of mutations in fibroblast growth factor 23 (FGF23) as the genetic cause of autosomal dominant hypophosphatemic rickets (ADHR) was a seminal discovery in the field of bone and mineral homeostasis.1 Alterations in FGF23 levels were subsequently identified as the common abnormality in a number of phosphate homeostasis disorders, including X-linked hypophosphatemic rickets (XLH),2 tumor-induced osteomalacia (TIO),3 fibrous dysplasia of bone,4 autosomal recessive hypophosphatemic rickets (ARHR),5 cutaneous skeletal hypophosphatemia syndrome (CSHS),6 familial tumoral calcinosis,7 and others. The primary actions of FGF23 are to regulate blood phosphate and 1,25-dihydroxyvitamin D3 (1,25-D) levels by its actions on renal sodium/phosphate cotransporters and 25-hydroxyvitamin D hydroxylation.8 The fact that PTH, like FGF23, is also important in regulating phosphate and vitamin D homeostasis invokes the existence of an FGF23-Vitamin D-PTH axis that is key to many important aspects of bone and mineral biology. It is a complicated axis that involves the interaction of multiple receptors, a coreceptor, Klotho,9 and multiple disparate but interacting signaling and enzymatic pathways.10, 11 Unraveling, in a hierarchical fashion, the roles of these complicated, intersecting, and overlapping pathways is important for both understanding mineral homeostasis physiology, and for the treatment of phosphate disorders. This has proved a significant challenge. Although a large number of very informative mouse models have been developed to understand and query this axis, at times the complexity of these models introduces potential confounders that are difficult to control for, and yield results inconsistent with observed human physiology. Thus, it is often difficult to understand and contextualize the findings and apply them to human mineral physiology and pathophysiology. The number of clinical studies conducted in this area is limited, and often focused on renal failure cohorts, the physiology of which differs significantly from those with normal renal function. Although a complete understanding of the FGF23-Vitamin D-PTH axis is lacking, there are a number of findings that are generally accepted to be true. These include: (1) FGF23 directly inhibits phosphate reabsorption and the generation of active vitamin D, 1,25-dihydroxyvitamin D3 (1,25-D) at the level of the proximal renal tubule leading to a lowering of blood phosphate and 1,25-D levels8; (2) in a classical feedback fashion, increases in blood phosphate and 1,25-D increase FGF23 levels12, 13; and (3) frank or relative FGF23-mediated reductions in serum 1,25-D lead to a compensatory increase in PTH, especially evident in the more severe forms of FGF23 excess such as TIO and renal failure.14, 15 The phosphaturic effects of increased PTH exacerbate the phosphate-lowering effects of FGF23 in patients with intact renal function such as TIO and XLH (reviewed in Blau and Collins11). As for the hierarchical relationship between FGF23 and PTH, FGF23's suppressive effect on 1α-hydroxylase is more potent than PTH's stimulatory effect. For example, 1,25-D levels are suppressed in TIO14 and renal insufficiency,15 despite normal or high PTH, a phenomenon that can be reversed with blockade of FGF23 signaling.16 However, it may be that PTH signaling is needed for the full phosphaturic action of FGF23 because PTH-deficient patients with hypoparathyroidism have elevated blood phosphate in the setting of elevated FGF23.17, 18 In this issue of the Journal of Bone and Mineral Research, Liu and colleagues19 undertake a straightforward, interesting, and potentially clinically relevant approach to the study and treatment of the hyp mouse, the animal model for XLH. The hyp mouse was the first genetic animal model employed to study disorders of FGF23. Of note, the hyp mouse was discovered both before the discovery of FGF23 and the identification of the responsible gene.20 Based on similarities in inheritance pattern and phenotype, the hyp mouse was recognized as an animal model for XLH, and the cause in both species eventually identified as mutations in the gene phosphate-regulating endopeptidase homolog, X-linked PHEX (originally called PEX),21 which encodes a transmembrane endopeptidase. The precise mechanism by which PHEX regulates FGF23 is not known, but it is clear that Phex mutations in hyp mice result in a defect in extracellular phosphate sensing, leading to an increase in FGF23 levels as blood phosphate increases,22 an observation also seen in patients with XLH.23 Regardless of a lack of understanding of the precise mechanism of the underlying pathophysiology, the hyp mouse remains a good model for studying the physiology and treatment of diseases of FGF23 excess. The major phenotypic abnormalities and clinical targets in XLH are short stature, due to growth plate abnormalities in childhood, and long-bone bowing and fragility due to rickets and osteomalacia; although the clinical spectrum is broad and severity of symptoms quite variable.24 Treatment is clearly indicated in the majority of children in an effort to prevent disabling short stature and skeletal complications. The role of treatment in adults is less clear, but is at least indicated for bone pain and fractures.24 Although novel therapies such as anti-FGF23 antibody drugs are in development,25, 26 and PTH-lowering by cinacalcet has been proposed as adjuvant therapy,27, 28 the current consensus from experts has coalesced around combination treatment with phosphate and calcitriol (or other active vitamin D analogues).24 Although this treatment has been shown to improve growth and promote rachitic healing, at the same time it is frequently associated with nephrocalcinosis.29 The increase in circulating FGF23 levels that occurs with increasing blood phosphate exacerbates phosphaturia both directly, by its action on the kidney, and indirectly by increasing PTH, which also promotes phosphaturia. The rise in urinary phosphate increases the risk for phosphate-calcium precipitates in the kidney.29, 30 To prevent secondary hyperparathyroidism, the dose of calcitriol is typically adjusted to normalize PTH levels. This too can contribute to the risk of nephrocalcinosis by increasing gastrointestinal calcium absorption and blood calcium, which increases the load of calcium filtered at the kidney. Hypercalciuria can ensue, especially if calcium is increased enough to suppress PTH below the point necessary to promote adequate renal calcium reabsorption. In this context, Liu and colleagues19 explored the therapeutic effects of calcitriol without phosphate supplementation on several biochemical and skeletal parameters in hyp mice. Hyp mice were treated for 75 days with calcitriol, either daily or twice weekly at a dose to maintain eucalcemia, and compared to mice treated with an anti-FGF23 antibody (FGF23Ab), and untreated hyp and wild-type (WT) mice. All treatment modalities resulted in a significant increase in serum phosphate with decreases in urinary phosphate excretion and serum PTH without inducing hypercalcemia or affecting renal function; urinary calcium excretion was not assessed. Daily calcitriol appeared to be most effective in reversing altered skeletal parameters, including bone microarchitecture, mineralization, and biomechanics. Further, daily calcitriol was superior at normalizing growth plate and metaphyseal organization, body weight, lumbar vertebral height, and femur length. The authors postulate that the beneficial effects of daily calcitriol on the growth plate are partly independent of the observed biochemical improvement, because accompanying in vitro studies of cultured growth plate chondrocytes showed that calcitriol had a phosphate-independent effect on mitochondrial ERK1/2 phosphorylation, a key step in hypertrophic chondrocyte apoptosis, which is necessary for the normal growth plate development. FGF23Ab treatment also stimulated endogenous 1,25-D production, but this increase was felt to be transient because 1,25-D levels had returned to baseline by the end of the study period. The authors attributed poorer skeletal outcomes in the FGF23Ab-treated group, at least in part, to waning serum 1,25-D levels. There are at least two potentially significant design flaws that make contextualization of this study difficult. First, the lack of a phosphate and/or combined phosphate and calcitriol treatment group limits the ability to directly compare calcitriol-alone treatment with that of the current conventional treatment, combined phosphate and calcitriol. This is particularly important given that combined phosphate and calcitriol was previously demonstrated to be superior to calcitriol-alone in hyp mice, at least in terms of improving mineralization.31 Second, it is quite possible that the suboptimal response of the FGF23Ab group was due to an insufficient dose or diminished efficacy of this drug. In a previously published report that utilized a different anti-FGF23 Ab drug,32 but which also lacked appropriate control groups to judge comparison of this drug to conventional treatment (there were no phosphate and/or phosphate + calcitriol comparator groups), the authors saw a significant increase in femur length and a greater degree of improvement and growth plate width and growth plate and metaphyseal organization than was seen by Liu and colleagues.19 Additional aspects that must be considered when interpreting the results of Liu and colleagues'19 study are the inherent limitations of the hyp mouse model to recapitulate certain features of the human disease, in particular adults with XLH. For instance, on histomorphometric analysis hyp mice demonstrate low bone volume, whereas patients with XLH have increased bone volume.33, 34 In addition, with aging, patients with XLH commonly have areas of skeletal sclerosis, which is not seen in the hyp mouse. Another species difference that emphasizes caution in extending findings from studies of the treatment of hyp mice to patients with XLH are the absolute differences in blood phosphate between mice and humans. WT and hyp mice both have blood phosphate levels significantly higher than humans, which may allow mice to achieve a threshold of blood phosphate sufficient to promote mineralization with the small changes in blood phosphate that are achieved with calcitriol-alone therapy. Earlier calcitriol-alone treatment in patients with XLH resulted in a smaller increase in blood phosphate than is seen in mice treated with calcitriol.35 In spite of these issues, the data presented by Liu and colleagues19 are convincing and may have important clinical implications. If these findings prove promising in further preclinical studies with appropriate controls, and confirmed in well-controlled studies in subjects with XLH, it could bring about a major change in the way disorders of FGF23 excess are treated. It is of interest and import to note though that over 35 years ago, in the pre-FGF23-discovery era, the calcitriol-alone approach suggested by Liu and colleagues19 had been assessed and judged to be an effective treatment in both hyp mice31, 36 and XLH.35 Interpretation of these studies is difficult though due to both design flaws in the early clinical studies (lack of systematic treatment and control groups) and lack of modern reagents and assays. However, by histomorphometric analysis, it appeared that osteoid volume and width showed greater improvement with combined phosphate and calcitriol treatment that than with calcitriol alone.35, 37, 38 Nonetheless, a calcitriol-alone treatment was, and sometimes is still used clinically, especially in cases of mild FGF23 excess. What is often an issue in calcitriol-alone treatment though, as was seen in these early studies, is hypercalciuria.35, 39 In conventional combined phosphate and calcitriol treatment of XLH (and other FGF23 excess disorders), nephrocalcinosis is a worrisome complication of therapy.24 However, data from studies analyzing risk factors for nephrocalcinosis in combined phosphate and calcitriol-treated XLH patients are conflicting. Vitamin D dose40, 41 and hypercalciuria42 appear as major factors for nephrocalcinosis in some studies, whereas other studies have found that the phosphate dose is a stronger predictor of nephrocalcinosis.29, 30, 42, 43 Although nephrocalcinosis is a major concern in XLH management, its association with impaired renal function in XLH appears less pronounced40, 42, 44-46 than in hypoparathyroidism.47 Therefore, revisiting this issue in the modern era, as proposed by Liu and colleagues,19 is important. It is not clear whether or not a calcitriol-alone therapy will prove beneficial for other forms of FGF23-mediated hypophosphatemia such as TIO, ADHR, ARHR, and CSHS, which do not appear to have a primary phosphate-sensing defect. Furthermore, although 1,25-D can stimulate gastrointestinal phosphate absorption, which appeared to be the case in the study by Liu and colleagues,19 the extent to which it can stimulate gastrointestinal phosphate absorption may be limited48 and not sufficient in cases of more profound hypophosphatemia. In addition to the previously mentioned novel therapy of FGF23Abs, is the recently reported efficacy of FGFR signaling inhibition by the specific pan FGFR inhibitor, BGJ398, in both preclinical49 and clinical settings.50 FGF23Ab (KRN23) and BGJ398 are currently under investigation and, if eventually available, will add to the armamentarium of drugs to treat FGF23-mediated phosphate-wasting disorders. The latest approach to be considered for treating FGF23 excess is the very recently reported inhibition of renal 24-hydroxylation of 25-hydroxyvitamin D with the 24-hydroxylase inhibitor, CTA102.51 It is clear that in mice FGF23 simultaneously stimulates 24-hydroxylation while it inhibits 25-hydroxylation of vitamin D, shunting vitamin D production along the so-called degradation pathway.8 Bai and colleagues51 show that inhibition of 24-hydroxylation with the 24-hydrxylase inhibitor CTA102 was able to heal the rachitic phenotype in both the hyp mouse and mice overexpressing FGF23. Although FGF23 regulation of 24-hydroxylation appears to play an important role in mice, this may not be true in humans. Dai and colleagues52 failed to find evidence of a significant FGF23-associated effect on catabolism of vitamin D as assessed by 24,25-vitamin D levels. Nor did we find either elevated 24,25-vitamin D or 1,24,25-vitamin D levels in patients with TIO before treatment, or a significant change in 24,25-vitamin D or 1,24,25-vitamin D levels after surgical cure at a time when 1,25-D levels rose dramatically in response to a rapid decline in FGF23 (unpublished data and Chong and colleagues14). Clinicians will eventually find themselves in the desirable position of having multiple treatment options that will allow them to match the treatment with the type and severity of the FGF23-excess disorder. For milder disease, the relatively simple calcitriol-alone therapy may be optimal. For more severe disease, more potent and/or adjuvant treatment may be necessary. This is an exciting time in the field of bone and mineral homeostasis; years of progress now allow us to reinvestigate earlier work with modern tools, and new therapies are around the corner. All authors state that they have no conflicts of interest. This work was funded in full by the Division of Intramural Research, National Institute of Dental and Craniofacial Research, National Institutes of Health, Bethesda, MD, USA. Authors' roles: All authors contributed equally to the content and writing of this manuscript.