Update on pharmacologically‐relevant vitamin D analogues

Abstract

Pharmacologists have been interested in vitamin D since its metabolism was elucidated in the early 1970s. Despite the synthesis of thousands of vitamin D analogues in the hope of separating its calcemic and anti‐proliferative properties, few molecules have reached the market for use in the treatment of clinical conditions from psoriasis to chronic kidney disease. This review discusses vitamin D drugs, recently developed or still under development, for use in various diseases, but in particular bone disease. In the process we explore the mechanisms postulated to explain the action of these vitamin D analogues including action through the vitamin D receptor, action through other receptors e.g. FAM57B2 and dual action on transcriptional processes.

Introduction

Since the metabolism of vitamin D was elucidated in the early 1970s, organic chemists and pharmacologists have sought to design and synthesize vitamin D analogues that would mimic some, if not all, of the biological actions of the parent vitamin. Thousands of molecules have been prepared and tested but only a few of these have been successfully translated into marketed drug forms that can be used to treat bone disease or in other areas of clinical medicine. These synthetic vitamin D analogues have been reviewed extensively in recent years 1, 2, 3, 4, 5 and this chapter will not attempt to duplicate those reviews. Furthermore, this review will not spend much time exploring the precise biochemical mechanism by which the active form of vitamin D₃, 1,25‐dihydroxyvitamin D₃ [1,25‐(OH)₂D₃] is able to modulate gene expression of target genes involved in the physiological effects of vitamin D on calcium/phosphate homeostasis or its effects on cell proliferation and differentiation covered in recent summaries of this subject 6, 7, 8. Rather, this short treatise will focus on the pharmacologically important vitamin D agents recently introduced or are under active development that might be employed ultimately in the treatment of bone disease or related conditions. We start by reviewing new work on two early metabolites of vitamin D₃: 25‐hydroxyvitamin D₃ [25‐OH‐D₃] and 24,25‐dihydroxyvitamin D₃ [24,25‐(OH)₂D₃], then move on to review newer analogues of 1,25‐(OH)₂D₃ that are modified chemically to be more selective in their actions, especially in bone.

25‐OH‐D₃

Prior to any discussion of vitamin D analogues, we must mention the parent vitamin D molecules, vitamin D₃ (cholecalciferol) and vitamin D₂ (ergocalciferol) that are still used in the treatment of vitamin D deficiency and its accompanying bone diseases, rickets and osteomalacia. Doses of 600–800 IU day^–1 have been recommended to maintain serum 25‐OH‐D (defined as the sum of 25‐OH‐D₃ + 25‐OH‐D₂) levels above 20 ng ml^–1 (50 nmol l^–1) in the normal healthy population by the Institute of Medicine 9 but the Endocrine Society suggests higher doses of 1000–2000 IU day^–1 are needed to cure vitamin D deficiency 10. Oral supplements of vitamin D (D₂ or D₃) are usually given at daily, weekly or monthly intervals in order to raise serum 25‐OH‐D levels, though a more frequent dosing regimen seems to be the more effective alternative. Much recent testing suggests that monthly dosing of vitamin D₂ is less effective than vitamin D₃ 11, 12, although the differences are not so evident when the two are compared in daily dosing protocols 13. A recent clinical trial 14 suggests that large doses of vitamin D₃ given every few months is associated with an increase in the frequency of falls and fractures, alhough the exact mechanism for how such side effects manifest is not known. Nevertheless, vitamin D supplements with their low cost and over‐the‐counter availability remain the agent of choice to treat vitamin D deficiency and associated bone diseases.

25‐OH‐D₃ (Table 1), the first metabolite of vitamin D₃ is produced in the liver and constitutes the main circulating form. Although, like vitamin D₃, 25‐OH‐D₃ is biologically inactive until it is 1α‐hydroxylated in the kidney, the provision of this prodrug form overcomes certain barriers sometimes observed with the parent vitamin D₃, namely: malabsorption of orally administered drug; sequestration of the lipophilic molecule by the adipose tissue; and slow or inefficient 25‐hydroxylation by the liver enzyme, CYP2R1 21. A synthetic drug form of 25‐OH‐D₃ (also known as calcifediol) named Calderol was first introduced in the mid‐1970s for the treatment of renal osteodystrophy and other bone diseases and, although it is still available in Europe 22, it was discontinued in North America in the late 1990s because of poor demand in the face of the initial success of 1,25‐(OH)₂D₃ and its analogues. Recently, 25‐OH‐D₃ was reintroduced in North America as an extended‐release form named Rayaldee (OPKO‐Renal, Miami, FL, USA) and approved for the treatment of secondary hyperparathyroidism experienced by stage 3 and 4 chronic kidney disease patients, the forerunner of renal osteodystrophy 23. Data from stage 3 and 4 chronic kidney disease patients show that 25‐OH‐D₃ effectively raises serum 25‐OH‐D₃ levels from low pretreatment values (<20 ng ml^–1) to normal levels (>30 ng ml^–1) and effectively lowers parathyroid hormone without causing hypercalcaemic side effects 24. Since these stage 3 and 4 patients have been reported to still possess residual CYP27B1 (1α‐hydroxylase) enzyme activity, it is believed that they can adequately convert 25‐OH‐D₃ to 1,25‐(OH)₂D₃ in a regulated manner. Thus, the expectation is that the prodrug, Rayaldee might provide a safer alternative to 1,25‐(OH)₂D₃ and its commonly prescribed 1α‐hydroxylated analogues, paricalcitol and doxacalciferol, which are now suspected of causing excessive calcification of soft tissues during long‐term use 25. Early use of Calderol in the 1980s and 1990s suggested that even dialysis patients who have lost renal function and the ability to 1α‐hydroxylate 25‐OH‐D₃ in the kidney could still make 1,25‐(OH)₂D₃ from large doses of 25‐OH‐D₃ in extra‐renal tissues due the presence of the extra‐renal CYP27B1 enzyme 26. Accordingly, it remains to be seen if Rayaldee will generate a sufficient rise in serum 1,25‐(OH)₂D₃ to reduce hyperparathyroidism in dialysis patients or can be used more widely to benefit other patients with bone diseases caused by low serum levels of 25‐OH‐D₃.

Table 1.

Natural metabolites of vitamin D and analogues of 1α,25‐(OH)₂D₃

Vitamin D analogue [ringstructure]a	Side chain structure	Company	Status	Possible target diseases	Mode of delivery	Reference
25‐OH‐D 3 1		Generic   OPKO‐Renal	In use Europe   In use N. America	Vitamin D Deficiency and/or Chronic Kidney Disease Stage 3/4 Hyperparathyroidism	Rapid release, oral Extended release oral	US Patent No. 3 565 924
24 R ,25‐(OH) 2 D 3 1		‐	‐	Bone fracture repair	‐	Martineau et al. 15
Calcitriol, 1α,25‐(OH) 2 D 3 2		Roche, Duphar	In use worldwide	Hypocalcemia, Psoriasis	Systemic Topical	Baggiolini et al. 16
2‐methylene‐19‐nor‐ 20‐epi‐1α,25‐(OH) 2 D 3 (2MD) 3		Deltanoids	Preclinical	Osteoporosis	Systemic	Shevde et al. 17
2‐methylene‐22‐dehydro‐19‐nor‐1,24 R ,25‐(OH) 3 D 3 (WT‐51) 3		Deltanoids	Preclinical	Osteoporosis	Systemic	US Patent Application No. 2014/0206655
ED71 (Eldecalcitol) 4		Chugai	In use Japan	Osteoporosis	Systemic	Nishii et al. 18
19‐nor‐14‐epi‐23‐yne‐ 1,25‐(OH) 2 D 3 (Inecalcitol, TX527) 5		Hybrigenics	Clinical trials	Prostate Cancer Leukemia	Systemic	Ma et al. 19
C8‐alkyl derivatives 6		‐	‐	Hyperproliferative disorders	‐	Gogoi et al. 20

^aStructure of the vitamin D nucleus (secosterol ring structure).

As described above, current dogma suggests that ALL vitamin D analogues have their therapeutic effects by mimicking the action of 1,25‐(OH)₂D₃ through a nuclear vitamin D receptor (VDR)‐mediated transcriptional mechanism, up‐and down‐regulating expression of vitamin D‐dependent genes 6, 7, 8. Although an additional theory involving a 1,25‐(OH)₂D₃ membrane‐receptor has been proposed 27, it has largely been dismissed because of the lack of convincing supporting evidence. Epidemiological studies have previously demonstrated an inverse association between serum 25‐OH‐D levels and lipid levels, as well as severity of metabolic diseases and body‐mass index 28, 29, 30, 31. However, Uesugi and colleagues 32 have recently shown in vitro that 25‐OH‐D₃ can inhibit lipogenesis by mediating the degradation of sterol regulatory‐binding element proteins (SREBPs) involved in transactivation of lipogenic genes, via a novel VDR‐independent mechanism 33. SREBP precursors reside in the membrane of the endoplasmic reticulum, bound to SREBP cleavage‐activating protein (SCAP). Under low cellular concentrations of sterols, the SREBP–SCAP complex translocates to the Golgi, where the N‐terminal transcription factor domain is cleaved by sequential action of site‐1 and ‐2 proteases. The mature form of SREBP translocates to the nucleus to mediate the transcription of lipogenic genes. The presence of cholesterol and 25‐hydroxycholesterol promote binding and sequestration of SREBP–SCAP complex to insulin‐induced genes, preventing translocation to the Golgi, transcription factor maturation and consequently lipogenesis during a sterol‐replete state, as a part of a negative feedback loop 33, 34, 35. 25‐OH‐D₃ inhibits lipogenesis by binding to SCAP, mediating its ubiquitination and degradation. The unbound SREBP is unstable and is ultimately degraded as well 34. While inhibition of lipogenesis via the 25‐OH‐D₃–SCAP–SREBP axis has not yet been tested in vivo, it presents a novel pathway that could potentially be targeted for vitamin D analogue design for use in treatment of metabolic diseases. It is possible that 25‐OH‐D₃ is simply mimicking 25‐hydroxycholesterol, which is a known oxysterol regulator of cholesterol metabolism. However, these preliminary data point to a possible beneficial effect of high levels of 25‐OH‐D₃ to lower lipid levels in bone disease patients treated with 25‐OH‐D₃. In addition to 25‐OH‐D₃, there is ample justification for the Japanese consortium to screen a library of vitamin D analogues to potentially find agents capable of more potent binding to SREBP/SCAP. This work is reportedly in progress 36.

24,25‐(OH)₂D₃

Since its discovery in the early 1970s, 24,25‐(OH)₂D₃ (Table 1) has been considered a degradatory product of 25‐OH‐D₃, devoid of biological activity and like its 1α‐hydroxylated counterpart, 1,24,25‐(OH)₃D₃, the first step on a catabolic pathway to truncated water‐soluble metabolites destined for excretion 37. The enzyme involved is CYP24A1, the deletion or mutation of which causes hypercalceamia in both mice and humans, evidence that is consistent with 24‐hydroxylation being a catabolic step 38. However, for 3 decades there has been the minority opinion that 24‐hydroxylated vitamin D metabolites might play an anabolic role in addition to the catabolic role evident when CYP24A1 exhibits loss of function. This role would be in mineralization or calcification of unmineralized osteoid 39 or tibial fracture repair in chicks 40. Until now, there has been a lack of any plausible mechanism for how 24,25‐(OH)₂D₃ could achieve this effect given the fact that it shows poor VDR binding and other evidence to support this theory has been meagre although a putative 24,25‐(OH)₂D₃ receptor has been claimed 41.

However, a very recent publication 15 has provided, for the first time, evidence for a possible novel mechanism to explain acceleration of bone‐fracture repair. Matrineau et al. have identified and cloned an effector molecule named FAM57B2 that binds 24,25‐(OH)₂D₃ and produces a lipid‐signalling molecule lactosyl ceramide (LacCer) thatmediates endochondral ossification during bone callus formation 15. Ablation of CYP24A1 gene or inactivation of Fam57b2 gene in chondrocytes caused loss of the bone‐healing effect while treatment with exogenous synthetic 24R,25‐(OH)₂D₃ or lactosyl ceramide, but not 1,25‐(OH)₂D₃, restores the effect 15. It is not yet clear if other 24‐hydroxylated vitamin D analogues or even other vitamin D analogues in general can reproduce the same bone‐healing effect as 24R,25‐(OH)₂D₃ at equivalent doses to those used here. 24,25‐(OH)₂D₃ is an allosteric modulator of FAM57B2 that catalyses formation of lactosylceramide from glucosylceramide. Although the specific role of LacCer in the chondrocyte remains unclear, it has been shown to participate in pathways such as cell proliferation, adhesion, apoptosis and angiogenesis. FAM57B2 ablation was found to phenocopy the callus defect observed in cyp24a1 –/– mice. Although the callus defect in cyp24a1 –/– mice could be rescued by either 24,25‐(OH)₂D₃ or LacCer administration; only LacCer treatment could restore the callus defect in Fam57b2 –/– animals indicating the requirement of both 24,25‐(OH)₂D₃ and its effector molecule FAM57B2 in this novel pathway of bone fracture repair. These data offer exciting evidence that 24‐hydroxylated vitamin D metabolites might be used to accelerate bone repair mechanisms.

Analogues of 1,25‐(OH)₂D₃

The physiological actions of 1,25‐(OH)₂D₃ include stimulation of intestinal calcium and phosphate absorption; conservation of calcium in the kidney; and stimulation of bone resorption through a VDR‐mediated mechanism of action 6, 7, 8. All of these functions are potentially useful for an agent used in the treatment of bone disease. Furthermore, although the VDR is present in both osteoblasts and immature cells of the osteoclast lineage and it was thought that the effects of 1,25‐(OH)₂D₃ on mature osteoclasts may be through effects on the rates of osteoclastogenesis and be regulated indirectly via osteoblasts involving osteoblast/osteoclast communication factors 42. At the molecular level, 1,25‐(OH)₂D₃ achieves its multiple effects by modulating the genes for key intestinal, kidney and bone proteins such as calbindins, Ca‐ATPases, osteocalcin, osteopontin and RANKL 6, 7, 8. It is not then surprising that various 1α‐hydroxylated vitamin D analogues (Table 1), such as 1,25‐(OH)₂D₃, 1α‐OH‐D₃ and 1α‐OH‐D₂ have undergone clinical trials for the treatment of osteoporosis 42. Four small trials of 1α‐OH‐D₃, three of which were performed in Japan, suggested that it was effective in reducing hip fractures while in other trials the results were inconclusive 43, 44, 45, 46, 47. In early clinical trials, 1,25‐(OH)₂D₃‐treated patients showed modest gains in bone mineral density and reductions in fracture rates, but a more recent trial demonstrated no statistically significant effect of 1,25‐(OH)₂D₃ on the incidence of nonvertebral fractures or vertebral deformities. Moreover, in these trials, there was an increased risk of hypercalcaemia. This is not a big surprise; 3 decades ago, Hock and colleagues treated rats with pharmacological doses of 1,25(OH)₂D₃ and found that animals developed hypercalcaemia and nephrocalcinosis along with a hyperosteoid or under‐mineralized bone histology 48. Thus, based upon these clinical trials, 1,25‐(OH)₂D₃ has not been recommended as a first‐line drug to treat osteoporosis.

More recently, several newer 1,25‐(OH)₂D₃ analogues (Table 1) have been developed that are claimed to be more selective towards bone, particularly promoting bone formation rather than resorption. As such, they are less calcaemic than 1,25‐(OH)₂D₃, but not noncalcaemic. Of particular note are the C2‐substituted vitamin D analogues such as ED‐71 (now called Eldecalcitol), currently in use in Japan for treatment of osteoporosis 49; and 2‐MD [2‐methylene‐19‐nor, 20‐epi‐1,25‐(OH)₂D₃] and its 22‐dehydro, 24‐hydroxy derivative WT‐51 17, 50, 51. Despite the increased size of the A‐ring, these C2‐substituted derivatives have been shown to still fit neatly into the ligand‐binding pocket of the VDR in X‐ray crystallographic studies and also modulate gene expression. Eldecalcitol appears to have a tighter affinity for the vitamin D‐binding protein and as a result is reported to have a longer half‐life in the plasma 18. Eldecalcitol restored bone mass without causing hypercalcaemia in long‐term studies involving ovariectomized rats 49. Both 2MD and WT‐51 represent newer 2‐substituted vitamin D analogues which promote bone nodule formation in vitro and cause increases in bone mineral density and bone strength in ovariectomized rats in vivo without causing hypercalcaemia 17, 50, 51. However, in a randomized, double‐blind, placebo‐controlled study of osteopaenic women, daily oral dosing with 2‐MD caused a marked increase in bone formation markers but also resulted in a marked increase in bone resorption 51. Consequently, the net effect was increased bone remodelling and to date 2‐MD has not been developed further. WT‐51 has yet to reach the clinic.

Another family of 1,25‐(OH)₂D₃ analogues, the C/D‐ring modified derivatives (Table 1), that are commanding attention are under development for use in other fields include the Hybrigenics drug, Inecalcitol [14‐epi, 23‐yne derivative of 1,25‐(OH)₂D₃], which induces apoptosis in squamous cell carcinoma models and is being tested in a clinical trial for treatment of leukaemia 19, 52; and the C‐8 alkyl, C‐ring‐less, m‐phenylene aromatic D‐ring derivatives of 1,25‐(OH)₂D₃ 20, which are purported to be noncalcaemic. Although there is optimism with every new family of vitamin D analogues, caution prevails as many have claimed that vitamin D analogues are non‐calcaemic or low‐calcaemic during in vitro testing only to find that the agents are rejected during clinical trials due to their calcaemic side‐effects in vivo.

Antagonists, nonsteroidal scaffolds and hybrid vitamin D molecules

The elucidation of the ligand‐binding domain of the VDR by X‐ray crystallography 53 has allowed for rational drug design of potential vitamin D analogues. The 1.8 A resolution crystal structure of the ligand‐binding domain reveals a folded helical arrangement around a pocket harbouring the 1,25(OH)₂D₃ molecule enclosed by a trapdoor composed of a helix‐12 53. During the binding of the ligand, the helix‐12 folds relative to the rest of the VDR structure to take up a different orientation and block exit of the ligand. The natural 1,25‐(OH)₂D₃ hormone occupies only 56% of the available space within the ligand‐binding pocket 53. Certain bile acids e.g. lithocholic acid will also fit into this pocket, suggesting that other molecules might trigger modulation of vitamin D‐dependent genes in vivo 54. Work has not only shown that synthetic vitamin D ligands can be modified in several new directions and still fit comfortably into the ligand‐binding domain (e.g. the 2‐substituted family of useful analogues, as noted earlier) but still allow for the helix‐12 rearrangement of the VDR necessary to trigger the recruitment of co‐activators and transactivation of vitamin D‐dependent genes 6, 7, 8. Agents that bind to the VDR but block the helix‐12 rearrangement such as the extended side‐chain vitamin D analogues (e.g. TEI‐9647 which is a 26,23‐lactone 55, 56 or the Schering ZK‐159222 compound, which is a side‐chain extended ester 57) are VDR antagonists (Table 2). TEI‐9647 was in clinical trial for the treatment of Paget's Disease 58.

Table 2.

Miscellaneous vitamin D compounds

Name	Structure	Name	Structure
TEI‐9647 Teijin   VDR antagonist Dehydration product of 1α,25( R )‐(OH) 2 D 3 –26,23( S )‐lactone   Saito and Kittaka 56 Ochiai et al . 55 Toell et al . 57 Ishizuka et al . 58		ZK159222 Schering   VDR antagonist   Toell et al. 57
LG190090 Ligand Pharmaceuticals   Non‐steroidal VDR agonist   Boehm et al . 59		LY2108491 Eli Lilly   Non‐steroidal VDR agonist   Ma et al. 60
DK 406   VDR‐Histone deacetylase inhibitor hybrid Bijian et al . 62

Modelling of the ligand‐binding pocket of VDR has also allowed for the design of vitamin D analogues based upon nonsteroidal scaffolds (Table 2). Biphenyl derivatives with a variable inter‐ring linker and side chains (e.g. Eli Lilly LY2108491 59 or the ligand LG190090 60) allow for the creation of a rigid molecule with the length, shape and functional groups to mimic the hormone, 1,25‐(OH)₂D₃. The main advantage of such molecules is that, unlike 1,25‐(OH)₂D₃, they lack chiral centres and are cheap to manufacture. Nonetheless, no nonsteroidal scaffold has entered clinical trials for the treatment of vitamin D‐related disease.

Another interesting concept is to make a hybrid vitamin D analogue that combines vitamin D functions with some complementary action. Retiferols combining vitamin D and retinoid functions have been successfully created by the Kutner group 61 could have some utility of the treatment of bone disease, while the VDR agonist has also been combined with a histone deacetylase inhibitor in the form of DK‐406 (Table 2) by Gleason and colleagues 62, its value arising from the fact that both agents are regulators of gene transcription and might be expected to act synergistically on certain vitamin D‐dependent genes.

Conclusions

As this review indicates, the same vitamin D metabolites that formed the original focus of pharmacologists' interest in the vitamin D field are currently undergoing a renaissance. Furthermore, the interest caused by 25‐OH‐D₃ and 24,25‐(OH)₂D₃ is not all through the same 1,25‐(OH)₂D₃‐VDR‐mediated gene transcriptional mechanism that has driven the synthesis of thousands of calcitriol analogues for the past 3 or more decades. Novel ideas about how these agents might be proven to work therapeutically have emerged that involve the extra‐renal CYP27B1 enzyme, binding to the SREBP‐SCAP complex or an effector molecule FAM57B2, which binds 24R,25‐(OH)₂D₃. Although none of these operate in the clinical arena, there is promise that the work may result in novel drugs for treatment of vitamin D deficiency and bone disease. Time will tell.

Competing Interests

There are no competing interests to declare.

Jones G., and Kaufmann M. (2019) Update on pharmacologically‐relevant vitamin D analogues, Br J Clin Pharmacol;85:1095–1102. 10.1111/bcp.13781.