Rickets, Vitamin D, and Ca/P Metabolism

Several recent histories of rickets and vitamin D are available [1-4]. Nutritional rickets has affected children since antiquity. Soranus of Ephesus, a Roman physician during the reigns of Trajan and Hadrian, left a remarkably complete text on gynecology, midwifery, and newborn care, stating:

“When the infant attempts to sit and to stand, one should help in its movements. For if it is eager to sit up too early and for too long a period it usually becomes hunchbacked (the spine bending because the little body has as yet no strength). If, moreover, it is too prone to stand up and desirous of walking, the legs may become distorted in the region of the thighs. This is observed to happen particularly in Rome as some people assume, because cold waters flow beneath the city. (Now) if nobody looks after the movements of the infant the limbs of the majority become distorted, as the whole weight of the body rests on the legs … then of necessity the limbs give in a little, since the bones have not yet become strong.”

[Soranus. Gynecology, Book II, 48. Translated by O. Temkin. The Johns Hopkins Press, Baltimore, 1956]

As detailed by Bagley [5], the descriptions provided by both Soranus and soon thereafter by Galen provide evidence that rickets was commonplace, perhaps even typical, among Roman children due to breastfeeding up to a year of age and the common practice of swaddling newborns for the first few months of life, preventing exposure to sunlight. O’Riordan and Bijvoet [6] cite other early, imprecise reports from Theodosius of Bologna (in 1554), and Bartholomaeus Reusner (in 1582), and state that “later texts, from Holland in 1614, might be referring to rickets under the titles of ‘Ailment of Saint Willibrod’ in which children had knobs on their ribs and ‘Ailment of Saint Machutus’ in which children had deformed legs,” thus the classic “rachitic rosary” was known at least 400 years ago. The 1634 Annual Bill of Mortality of the City of London listed 10,900 deaths, ascribing 14 of them to “rickets,” and by 1659, 441 deaths were ascribed to rickets [6]. Most authors date the medical understanding of rickets from 1645, when Daniel Whistler described rickets and osteomalacia in his MD thesis [7], soon followed by Francis Glisson’s extensive text, first in Latin in 1650 [8], then in English in 1668 [9]. In 1772, Levacher de la Feutrie published a French text on rickets, including descriptions of devices to straighten deformed bones [10].

The first person to implicate deficient sunlight in the etiology of rickets and its use as a cure may have been the Polish physician/chemist Jedrzej Sniadecki (1768–1838). As reported by Wlodzimierz Mozolowski [11], in volume 1 of his “Dziela,” written in 1822, but published in 1840, Sniadecki stated:

“If the parents’ financial status permits, it is best to take the children out into the country and keep them as much as possible in the dry, open and pure air. If not, at least they should be carried about in the open air especially in the sun, the direct action of which on our bodies must be regarded as one of the most efficient methods for the prevention and the cure of this disease… Thus strong and obvious is the influence of the sun on the cure of the English disease, and the frequent occurrence of the disease in densely populated towns, where the streets are narrow and the dwellings of the working-class people low and very poorly lit.”

Mozolowski concluded: “From the passages quoted above it is quite clear that J. Sniadecki fully realized the significance of the sun both in prevention and cure of rickets.” Thomas Percival reported the use of cod liver oil for “rheumatisms” in adults, and quotes a letter from a physician, Robert Darbey, the last paragraph of which reports its salutary effect in two children with apparent rickets, but without using that term [12]. Both D. Schütte in 1824 [13] and others (reported by ref. [6]) showed that cod-liver oil was effective in treating rickets. As reported by Chesney in 2012, 60–90% of urban British children had rickets in the late 19th century, but its cause and treatment were controversial [14]. Theobold Palm, a medical missionary, noted the absence of rickets in Asian countries, but a high incidence in Britain, with a much higher incidence in crowded urban areas; he recommended sunbaths and relocating rachitic children from urban to rural areas [15]; Bucholz reported curing rachitic children with incandescent light in 1904 [16].

The science of the etiology and treatment of rickets advanced rapidly after World War I. Mellanby rendered puppies rachitic, then supplemented their diets with various additives, noting that fats, notably cod liver oil, repaired the bony defects, concluding that “fat-soluble A” was responsible for the antirachitic action [17]. In 1922, McCollum et al. [18] showed that heat and oxygen inactivated the fat-soluble vitamin A in cod liver oil, but not the fat-soluble “calcium-depositing vitamin” (i.e., vitamin D), showing that vitamin A was not responsible for the salutary action of cod liver oil. Also at this time, UV light was successfully used to treat rickets [19, 20] (Fig. 1); Hess noted that rickets had a seasonal variation, being most prevalent in late March, after children had been indoors during the winter [21] and that such “heliotherapy” increased circulating phosphate levels [22].

The link between sunlight and dietary supplementation with a “vitamin” came in 1924, when Hess and Weinstock at Columbia University [23] and Steenbock & Black at the University of Wisconsin [24] showed that UV radiation of either certain fats or plants endowed them with antirachitic properties; the use of filters established that “UVB” radiation (∼280–320 nm) was the active form of sunshine [25]. The carefully controlled work of Martha May Eliot led to the acceptance of both cod liver oil and sunlight in the prevention and cure of rickets [26]. Cod liver oil is unpalatable, but milk and other foods were soon irradiated with UV light, nearly eliminating rickets. The material activated by UV light was traced to the “sterol” fraction of both animal and vegetable foods, suggesting that light acted on a cholesterol-like molecule (now known to be 7-dehydrocholesterol) in skin. In 1931, Askew et al. [27] reported the isolation and structural characterization of ergocalciferol (vitamin D₂) from irradiated ergosterol. Desmond Bernal was first to use X-ray crystallography to study biological molecules, showing that cholesterol is a planar, 4-ring structure, and determining the structure of vitamin D [28]. In Göttingen, Adolph Windaus, recipient of the 1928 Nobel Prize in Chemistry for his work with sterols, showed that D₂ was produced by UV radiation of ergosterol [29] and that D₃ is similarly produced from 7-dehydrocholesterol [30]. The antirachitic component of cod liver oil was found to be identical to vitamin D₃, thus closing the loop between fish oil and sunshine. An outline of research discoveries concerning calcium metabolism is presented in Table 1.

Early clinical descriptions of two severe, infantile rachitic disorders presaged the elucidation of the pathways of vitamin D synthesis and action. Resistance to the action of vitamin D, even when administered in huge doses, was reported by Fuller Albright in 1937 and was logically termed “vitamin D-resistant rickets” [31]; in 1961, Andrea Prader described a clinical phenocopy of this disorder that was sensitive to high-dose vitamin D therapy [32], and became known as “vitamin D-dependent rickets.” Such observations raised the question of what chemical forms of vitamin D were biologically active. Using tritiated vitamin D₃ revealed polar metabolites; substantial biochemical effort (silica gel chromatography, reversed-phase chromatography, mass spectrometry, and NMR spectrometry) identified the principal form in human circulation as 25OHD₃, which was more potent than D₃ [33]. From 1969 to 1971, several groups, notably that of Egon Kodicek in Britain, identified 1,25(OH)₂D₃ (calcitriol) from the kidney as the truly active metabolite of vitamin D [34-37]. The corresponding forms of vitamin D₂ were soon identified, as were other metabolites hydroxylated at C24, C26, and the various combinations of these hydroxylations [1, 4]. This work established that both vitamins D₂ and D₃ are 25-hydroxylated in the liver, that 25OHD is the predominant metabolite in the circulation, and that the kidney then 1α-hydroxylates 25OHD to biologically active 1,25(OH)₂D₂ or 1,25(OH)₂D₃. The disease identified by Prader in 1961 was a defect in renal 1α-hydroxylation [38]. With the dawn of the era of molecular biology, most steroidogenic enzymes were cloned between 1984 and 1990, and the identification of the genes and enzymes participating in vitamin D metabolism followed (reviewed in [39]).

Study of hepatic vitamin D 25-hydroxylation was complicated by the presence of multiple enzymes that could catalyze this reaction (at least in vitro), and by the lack of obvious regulation of this step, so that circulating 25OHD concentrations are mainly determined by dietary intake and exposure to UV light. In 1990, two groups reported 25-hydroxylase activity in both mitochondria and microsomes and cloned cDNA for an enzyme initially called P450c25 [40, 41], but that enzyme mainly hydroxylates C26 and C27 in bile acid synthesis; it is now termed CYP27A1 [42], and its mutations cause cerebrotendinous xanthomatosis without disordered calcium metabolism [43], and hence it is not a significant vitamin D 25-hydroxylase. Several hepatic microsomal P450 enzymes can also catalyze some 25-hydroxylase activity, but are clearly minor players [39].

In 1994, Samuel Casella et al. [44] reported two brothers of Nigerian parentage who had hormonal findings suggesting 25-hydroxylase deficiency, but subsequent studies found no mutations in CYP27A1. In 2003, David Russell’s laboratory reported that microsomal CYP2R1 was a potent vitamin D 25-hydroxylase [45]; in collaboration with Michael A. Levine, they used DNA from an EBV-transformed cell line established from one of Casella’s patients to identify a homozygous mutation of CYP2R1 that impaired 25-hydroxylase [46]. Follow-up studies found no CYP2R1 mutations in 27 Nigerian children with sporadic rickets, but missense mutations were identified in affected members in 2 of 12 families [47]. These genetic findings confirm that CYP2R1 is the principal human hepatic enzyme catalyzing 25-hydroxylation of vitamin D; 25-hydroxylase deficiency is very rare, probably because other enzymes contribute to vitamin D 25-hydroxylation in vivo. Polymorphisms in or near the CYP2R1 gene differ in their worldwide distribution and affect vitamin D homeostasis [48].

Most vitamin D is inactivated is by its 23- and 24-hydroxylation by mitochondrial CYP24A1 in the kidney and intestine, initiating the inactivation pathway leading to calcitroic acid. Ohyama et al. [49] first cloned CYP24 by purifying the protein from rat renal mitochondria, raising an antiserum, and screening a rat kidney cDNA expression library; the human cDNA was reported 2 years later [50]. Homozygous or compound heterozygous mutations in CYP24A1 causing hypercalcemia were reported in 2011 [51, 52]. Most patients were infants with weight loss, failure-to-thrive, hypercalcemia, hypercalciuria, and/or nephrocalcinosis, with normal 25OHD levels, normal to moderately elevated 1,25(OH)₂D levels, low 24,25(OH)₂D levels, and low levels of parathyroid hormone (PTH). Once known causes of infantile hypercalcemia are eliminated, the diagnosis is usually “idiopathic infantile hypercalcemia”; however, such infants may have CYP24A1 mutations. In addition, heterozygous CYP24A1 mutations may lead to clinically apparent hypercalcemia in individuals ingesting very large amounts of vitamin D [53, 54]. In the 1950s, there was a dramatic rise in the incidence of “idiopathic infantile hypercalcemia” in Britain shortly after the introduction of excessive fortification of milk and cereal with vitamin D [55]. Britain and most of the world then banned fortification of milk with vitamin D, and the incidence of hypercalcemia decreased, but the USA and Canada, which added much less vitamin D to milk and other foods, continued their successful policy of milk fortification; this history was reviewed in detail in 1967 [56]. Contemporary data show little correlation of serum calcium levels and vitamin D intake in normal adults [57].

1α-Hydroxylation has been of long-standing interest because of its importance in normal physiology and because synthesis of 1,25(OH)₂D is impaired in several clinical disorders. Classic approaches for protein purification failed because of the very low abundance of 1α-hydroxylase in renal mitochondria. However, in 1997, four groups (St-Arnaud & Glorieux [McGill U. Montreal], Takeyama and Kato [U. Tokyo], Miller& Portale [UCSF] and Shinki & Suda [Showa U., Tokyo]) using different approaches reported the cloning of the human, rat, and mouse cDNAs [58-62] and the human gene [62, 63] for the mitochondrial vitamin D-1α-hydroxylase, subsequently termed CYP27B1. Kato’s group used mice lacking the vitamin D receptor (VDR), which thus overproduced the 1α-hydroxylase enzyme, and then screened a cDNA expression library for activation of a VDR construct [59]. The Glorieux and Shinki groups increased rat renal 1α-hydroxylase mRNA by feeding the animals a diet deficient in calcium and phosphorus, then used probes corresponding to the conserved P450 heme-binding site to identify candidate sequences [58, 61].

At UCSF, Walter L. Miller’s laboratory, collaborating with Anthony A. Portale in pediatric nephrology, obtained the first human clone [60]. Human keratinocytes have robust 1α-hydroxylase activity when grown in serum-free, low-calcium medium; they made a keratinocyte cDNA library and screened it with oligonucleotides having sequences corresponding to the ferredoxin-binding and heme-binding sites of other P450s, identifying a putative 1α-hydroxylase cDNA. Multiple experiments indicated that the CYP27B1 encoded by the keratinocyte cDNA was the same as the renal enzyme; most notably, keratinocyte cDNA from a patient with 1α-hydroxylase deficiency carried the same mutations found in the patient’s genomic DNA. Thus, “vitamin D-dependent rickets type I” was vitamin D 1α-hydroxylase deficiency, as expected based on Prader’s report in 1961 [32]. The human CYP27B1 gene was cloned and localized to chromosome 12 [63]; at the same time, St-Arnaud and Glorieux cloned the rat cDNA and used it to map the human gene to 12q13.1-13.3 [58]. This was the predicted location, because in 1991, taking advantage of the very high incidence of 1α-hydroxylase deficiency among the French Canadian population of the Charlevoix-Saguenay-Lac Saint Jean region of Quebec, that laboratory had mapped 1α-hydroxylase deficiency to chromosome 12q13-q14 by linkage analysis [64].

“Vitamin D 1α-hydroxylase deficiency” is a mechanistically descriptive term designating “hereditary pseudo-vitamin D deficiency rickets”, “vitamin D dependent rickets,” or “vitamin D-dependent rickets type I.” Because 1α-hydroxylase deficiency is rare, most studies have reported only a few individuals. However, in 1998, Wang et al. identified the mutations in 19 patients from 17 families in multiple ethnic groups [56]. DNA sequencing and microsatellite haplotyping showed that French-Canadian patients carried a single haplotype and the same CYP27B1 frameshift mutation (958ΔG in codon 88) that ablates enzyme activity. This study also found that multiple families carried a third copy of the 7-base sequence CCCACCC that is normally duplicated in CYP27B1 exon 8, causing an inactivating frameshift mutation. This 7-bp insertion was associated with several different microsatellite haplotypes and was found in several unrelated ethnic groups, indicating that the 7-bp insertion arose recurrently de novo [65]; the recurrent nature of this mutation has been confirmed in other studies [66, 67]. Some rare missense mutations retain partial activity [68, 69], but most patients with classic findings have nonsense mutations [70].

Albright’s paper in 1937 [31] is generally credited as the first description of vitamin D resistance, but he lacked the assays now used to define that disorder clinically; and based on the prominence of renal phosphate wasting and hypophosphatemia in this description, his patient may instead have had X-linked hypophosphatemia. In 1969, Mark Haussler at the University of Arizona provided the first evidence for a VDR that bound to chromatin [71], and in 1974, he reported that a specific cytosolic receptor bound calcitriol and associated with chromatin in chicken intestinal cells [72]. He coauthored a 1978 report described a young woman with hypocalcemia, hyperparathyroidism, osteomalacia, and osteitis fibrosa in whom serum concentrations of 25OHD were normal but those of 1,25(OH)₂D were very high [73]. That report proposed the patient had end-organ resistance to the action of 1,25(OH)₂D and proposed the name “vitamin D-dependent rickets type II” (VDDR-II) to distinguish it from VDDR-I (1α-hydroxylase deficiency). Marx et al. [74] described a brother and sister who developed infantile rickets with similar clinical and laboratory findings, but whose hypocalcemia responded to doses of calcitriol about 20 times higher than those needed to treat 1α-hydroxylase deficiency; this suggested a disorder in the unknown molecule(s) that mediated the action of vitamin D. Rickets with alopecia, now recognized as a classic presentation of vitamin D-resistant rickets, was reported in 1979 with clinical/metabolic characterization [75] (Fig. 2), and a study in 1982 showed absent binding of 1,25(OH)₂D in cells from a similar patient [76]. The human VDR was cloned in 1988 [77], and mutations causing vitamin D-resistant rickets were identified soon thereafter. The human VDR is a typical zinc-finger protein nuclear receptor, encoded by a gene on chromosome 12q13-14, very near the CYP27B1 gene for the vitamin D 1α-hydroxylase. The first mutations in the VDR gene were reported in 1988 [78], and mutations were soon found in all exons and all functional domains of the VDR [79].

The assiduous efforts of Renaissance anatomists failed to reveal the parathyroid glands, which escaped notice until the 19th century. Both Remak in 1851 and Virchow in 1864 apparently mentioned parathyroids briefly [80]. Richard Owen (1809–1882), describing his 1852 autopsy of an Indian rhinoceros that died at the London Zoo, reported “a small, compact, yellow glandular body attached to the thyroid at the point where the veins emerge” [81]. In 1877, Ivar Sandström (1852–1889), then a medical student in Uppsala, described the parathyroids in detail by dissections and histology of dogs, cats, cattle, horses, and rabbits and then confirmed his results in 50 human cadavers. His manuscript was apparently rejected by prominent German journals [82, 83] and finally appeared in a local Swedish language journal in 1880 [84, 85]. Experimental parathyroidectomy (sparing the thyroid) caused tetany and seizures, which had previously been noted in experimental thyroidectomy (which did not spare the parathyroids) [86, 87]. Despite this, most early investigators believed that the parathyroids were functionally, as well as anatomically, associated with the thyroid. In 1899, Gustave Moussu reported that an equine parathyroid extract worsened myxedema in dogs but benefited hyperthyroidism, and suggested that the action of the parathyroids was to restrain the thyroid [88].

The roles of the thyroid and parathyroids were distinguished in 1909 when Berkeley and Beebe ameliorated human tetany with a parathyroid extract [89], and MacCallum and Vogetlin noted that parathyroid destruction caused tetany responsive to intravenous calcium chloride [90]; these studies connected the parathyroids to calcium metabolism. An alternative hypothesis was that the parathyroids were detoxifying organs, as parathyroidectomized animals excreted methyl guanidine [91], and its administration caused tetany [92]. PTH was identified independently by Adolph M. Hanson (1880–1959), a private practitioner working in his garage in Faribault, MN [93-95], and by James B. Collip (1892–1965), chair of biochemistry at the University of Alberta, who was famous for preparing the insulin used by Banting, Best, and MacLeod in Toronto. Hanson lacked Collip’s resources and influence; McCullagh’s 1928 review [96] does not even mention him. Hanson patented his preparation, but was unable to market it, and eventually assigned his patent rights to the Smithsonian Institution [82]. Collip may have been inspired to work on the parathyroids by his 1919 meeting with Noel Paton, the principal proponent of the “methyl guanidine school” [97]. Collip’s preparation was superior to Hanson’s, corrected the tetany of parathyroidectomized dogs [98], and caused hypercalcemia in normal animals [99]; in an era when every little observation could be published, these two papers are surprisingly modern in their thoroughness and scope. Collip’s preparation was then marketed by Eli Lilly with the brand name “Parathormone.” A history of this work, based on personal notes and files as well as the published literature, is provided by Alison Li [97].

The harsh HCl extraction procedure used by Collip resulted in the partial degradation of PTH into multiple peptides; Rasmussen showed that one of these peptides of only about 33 amino acids was active in raising Ca levels in rats [100]. The 84 amino acid sequence of bovine PTH was reported in 1970 [101], the sequence of the bioactive N-terminal 34 amino acids of human PTH was reported in 1972 [102], and the full-length protein in 1978 [103]. Teriparatide, the N-terminal 34 amino acids of human PTH, is marketed by several firms for treatment of osteoporosis. Several reports describe its use in children, but it has not been approved for children because of animal data suggesting a possible risk for osteosarcoma.

Over the next 20 years, clinical disorders of hypo- and hyperparathyroidism were reported, most famously the case of Capt. Charles Martel with hyperparathyroidism due to a mediastinal tumor (Fig. 3) [104]. Studies of hyperparathyroidism did not reveal the site of PTH action, leading to a vigorous debate: J.B. Collip said it acted on bone, while Fuller Albright said it was on the kidney; both were right, but they fought over this issue for 17 years [105]. Albright described “pseudohypoparathyroidism” as an example of the “Seabright-Bantam” (sic) syndrome, referring to the male Sebright bantam chicken and its female feathering pattern; this is now known to be due to an aromatase defect [106], but these animals were long thought to represent “end-organ resistance” to the action of a hormone [107]. In fact, patients with pseudohypoparathyroidism do not have mutations of the PTH receptor, but are a heterogeneous population with different GNAS mutations (for review see [108]). Thus, Albright is generally credited for describing hormone resistance as a mechanism of disease.

In 1968, Angelo DiGeorge (1921–2009), a founding member and past president of the (LW)PES, described congenital hypoparathyroidism in conjunction with thymic aplasia and immune defects [109]. The syndrome was broadened to include multiple congenital anomalies and became known as a developmental field defect of the third and fourth pharyngeal pouches; thus, “DiGeorge Syndrome” became known as “velo-cardio-facial syndrome.” The majority of these patients had a deletion on the long arm of chromosome 22 (chromosome 22q11.2 deletion syndrome) [110]. The acronym CATCH22 (cardiac abnormality/abnormal facies, T-cell deficit due to thymic hypoplasia, cleft palate, and hypocalcemia due to hypoparathyroidism resulting from 22q11 deletion) is also widely used [111]. Botto et al. [112] found an incidence of about 1:6,000 births, making it is one of the most common chromosomal disorders.

It was known from the 1920s that deficiency of either vitamin D or PTH caused hypocalcemia, but through the 1960s, the connection was unclear, despite complex, well-executed experiments [113, 114]. Children with 1α-hydroxylase deficiency (formerly termed “VDDR-I” or pseudo-vitamin D-deficiency rickets) typically have hypocalcemia and secondary hyperparathyroidism [38, 115], and patients with CYP27B1 gene mutations causing 1α-hydroxylase deficiency have elevated PTH values [65], even in those where there is partial retention of function [68]. When the genes for rodent and human 1α-hydroxylase were cloned in 1997 (described above), it was soon shown that PTH activates expression of the mouse Cyp27B1 gene [116], closing the loop between PTH and vitamin D.

Hypercalcemia without elevated circulating PTH was long noted as a paraneoplastic syndrome in some adult patients with malignancies. A long search led to the identification of a parathyroid hormone-related protein (PTHrP) by groups in Melbourne, Australia [117, 118], and in San Francisco [119]. The physiological roles of PTHrP remain under investigation; it is a widely expressed autocrine/paracrine factor, rather than a classic hormone; a major role in fetoplacental calcium transport was suggested by its abundant expression in fetal membranes [120], and an active role in maternofetal calcium transport was shown in 1996 [121]. Thus, PTHrP, discovered in adult cancer, appears to be mainly a fetal hormone with multiple roles in calcium and bone metabolism.

Discoveries about the physiology of phosphate metabolism were facilitated by the existence of several disorders caused by abnormalities related to fibroblast growth factor 23 (FGF23), especially the genetic forms of hypophosphatemic rickets caused by renal phosphate wasting. While not a complete assessment of every discovery, we provide a general overview of the journey highlighting several important milestones and advances in phosphate disorders in the last century, leading to the discovery of FGF23 and directed treatment of patients with these disorders.

Although earlier authors had described case reports of rickets that was difficult to treat, in 1937 Fuller Albright provided the first detailed investigation of a child with hypophosphatemic rickets due to renal phosphate wasting and without other evidence of renal tubulopathy. He characterized this disorder as being a “vitamin D-resistant rickets,” based on the inability to effectively heal the rickets with doses of vitamin D usually used for nutritional rickets, and the ability to demonstrate some skeletal improvements with higher vitamin D doses than usual [31]. In 1957, Winters et al. [122] then highlighted several published cases of different forms of rickets to which “vitamin D resistance” was attributed, but noted especially the X-linked dominant familial form of hypophosphatemic rickets as a “vitamin D-resistant” form of rickets. Shortly thereafter, Prader et al. [123] described a patient with hypophosphatemic tumor-induced osteomalacia.

Early efforts to treat patients with hypophosphatemic rickets and osteomalacia focused on the application of high doses of vitamin D, which carried a risk for vitamin D toxicity. Frame and Smith noted that adult patients with acquired hypophosphatemic osteomalacia could improve using phosphate salts [124]; in children, the rickets could also be refractory to oral phosphate salts [125]. Given the limited response to vitamin D, Stickler et al. [126] tried to treat familial hypophosphatemic rickets by increasing the calcium-phosphate product through increasing intake of phosphate and calcium (on alternating days to ensure absorption), without vitamin D. However, this approach also proved unsuccessful, and they concluded that, at least to some degree, gastrointestinal absorption was also impaired. However, based on later understanding of the underlying physiology of X-linked hypophosphatemia, true “vitamin D resistance” at the level of the receptor is not actually part of the pathophysiology. Thus, the term “vitamin D resistant rickets” is misleading and in the modern setting should be abandoned in patients with X-linked hypophosphatemic rickets (XLH) and applied only when the VDR gene is mutated (see above).

In 1978, Charles Scriver provided an important link to the apparent abnormalities in vitamin D responsiveness. He and his coauthors used a competitive binding assay to demonstrate that 1,25-dihydroxyvitamin D (1,25[OH]₂D or calcitriol) levels were impaired in children with hypophosphatemic rickets [115]. With the combination of hypophosphatemia and low or low-normal calcitriol levels being key components of this renal phosphate-wasting disease, studies were then conducted to confirm that addition of calcitriol was critical to healing the osteomalacia and rickets [127-129]. Combining phosphate salts with high doses of calcitriol (or other active vitamin D analogues) became the standard of care for children with XLH. Both the effects of XLH on phosphorus and on calcitriol are mediated through the kidney, but neither administration of phosphate nor calcitriol corrected the renal phosphate leak [127]. The underlying cause of the phenotype in XLH remained elusive.

A key step in further understanding of XLH was the generation of the hyp mouse model, published in 1976 by Eicher et al. [130]. This mouse model recapitulated the X-linked dominant inheritance pattern and the biochemical and skeletal phenotype of XLH. Even though the genetic defect had not yet been discovered, this model enabled studies that eventually demonstrated clearly that a hormonal factor was responsible for the renal phosphate wasting, rather than a kidney factor. Through parabiosis experiments, Meyer et al. [131] were able to demonstrate that the wild-type mouse developed phosphaturia when parabiosed to the hyp mouse, indicating that crossing of a humoral factor was causative [131]. Further confirmation of a humoral factor came when Nesbitt et al. [132] conducted crossed renal transplant experiments between hyp and wild-type mice. The hyp mouse kidney did not develop phosphaturia when transplanted in the wild-type mouse, but the wild-type kidney did develop phosphaturia when transplanted in the hyp mouse.

A consortium of investigators was convened to find the genetic cause of XLH, through linkage studies in kindreds. Eventually this led to the discovery of the phosphate-regulating gene with homologies to endopeptidases on the X-chromosome or PHEX (originally published in 1997 as PEX and later changed) [133]. However, the protein PHEX is expressed in bones and teeth and is not a humoral factor.

A kindred with an autosomal dominant form of hypophosphatemic rickets had been described in 1971 by Bianchine et al. [134], having similarities to XLH. Michael Econs later studied another kindred with this inheritance pattern, which was described as autosomal dominant hypophosphatemic rickets (ADHR) [135]. Persons with ADHR presented with a similar biochemical phenotype to XLH: hypophosphatemia, impaired tubular reabsorption of phosphate, and low or inappropriately normal calcitriol levels. However, it was clear from this kindred that not everyone developed “rickets.” This family demonstrated incomplete penetrance. Additionally, some children presented similarly to XLH and then normalized their biochemistry over time. Others were confirmed as normal during childhood and developed new onset severe hypophosphatemia and osteomalacia as adolescents or adults. Genetic linkage studies in this and other ADHR kindreds led to the discovery of FGF23 in 2000 [136]. The name FGF23 derives from its structural relationship to other FGF (fibroblast growth factor) molecules and does not imply an action on fibroblasts. FGF23 was also cloned from tumors causing tumor-induced osteomalacia [137]. FGF23 proved to be a humoral factor normally expressed in bone, and the mutation causing ADHR occurred in a cleavage motif, leading to impaired proteolytic cleavage preserving intact, active FGF23 [138, 139]. FGF23 appears to fulfill all the characteristics that had been expected of the hypothetical hormone “phosphatonin” [140].

Multiple causes of autosomal recessive hypophosphatemic rickets or ARHR were also discovered starting in 2006, through studies of clinically affected patients. These linked abnormalities of 3 additional genes, DMP1, ENPP1, and FAM20C, to increased gene expression of FGF23 [141-144]. Thus, the common feature of each of these genetic diseases (XLH, ADHR, and the forms of ARHR) is high FGF23 expression, along with the resulting biochemical consequences of excess FGF23.

Klotho had been discovered in 1997 in a mouse model that had multiple aging-related phenotypes [145]. While klotho had been thought to be a gene conferring an antiaging phenotype, several of the features of klotho deficiency actually recapitulated mineral metabolism disorders, including hyperphosphatemia and high calcitriol. The deficiency of klotho in mouse models was strikingly similar to that of FGF23 deficiency [146]. In 2006, Kuroso et al. [147] made the critical connection when klotho was found to be a crucial cofactor in the signaling of FGF23 through FGF receptors. Alpha klotho enables signaling of FGF23 through the FGF receptors, FGFR1, 3, and 4, and provides tissue specificity of FGF23 activity, primarily in the kidney. Ichikawa et al. [148] demonstrated the first pathogenic variant in klotho causing human disease. This patient had hyperphosphatemic tumoral calcinosis along with extreme elevation of intact FGF23 consistent with a response to FGF23 resistance [148].

It is important to note that treatment with calcitriol and phosphate does not lower FGF23, but rather stimulates FGF23 levels as demonstrated by multiple studies in normal humans and in patients with XLH [149-151]. Since FGF23 mediates the hypophosphatemia seen in XLH, it is logical to target FGF23 itself to improve the phenotype of XLH. A placebo-controlled trial of anti-FGF23 antibody (burosumab) in adults with XLH demonstrated increased phosphorus levels, fracture healing, and other salutary patient-reported outcomes [152], and in children burosumab improved rickets more than conventional therapy.

Much has been learned regarding phosphorus metabolism since the first description of children and adults with hypophosphatemic rickets. FGF23 has become critically important in our understanding of these disorders and also has relevance to more common conditions such as chronic kidney disease. Future studies will expand our understanding of mechanisms and the consequences of too much and too little FGF23, phosphorus, and their influence on overall mineral metabolism.

Not applicable.

Dr. Erik A. Imel receives research funding and consulting from Ultragenyx Pharmaceuticals and consulting fees from Kyowa Kirin Pharmaceuticals.

The authors have not received any funding or financial support for this work.

Walter L. Miller wrote the material concerning the early history of rickets, vitamin D, and PTH; Erik A. Imel wrote the material concerning phosphorus and FGF23; both authors reviewed and approved the entire manuscript.

No new data were generated for this report.