Vitamin D Signalling Pathways in Cancer: Potential for Anticancer Therapeutics

REVIEWS

Vitamin D signalling pathways in cancer: potential for anticancer therapeutics

Kristin K. Deeb*, Donald L. Trump‡ and Candace S. Johnson* Abstract | Epidemiological studies indicate that vitamin D insufficiency could have an aetiological role in various human cancers. Preclinical research indicates that the active

metabolite of vitamin D, 1α,25(OH)2D3, also known as calcitriol, or vitamin D analogues might have potential as anticancer agents because their administration has antiproliferative

effects, can activate apoptotic pathways and inhibit angiogenesis. In addition, 1α,25(OH)2D3 potentiates the anticancer effects of many cytotoxic and antiproliferative anticancer agents. Here, we outline the epidemiological, preclinical and clinical studies that support the

development of 1α,25(OH)2D3 and vitamin D analogues as preventative and therapeutic anticancer agents.

The most widely accepted physiological role of vitamin The seminal finding by Garland and Garland12 of

D, which is mediated primarily by 1α,25(OH)2D3 (also higher mortality rates from colon cancer in the northeast known as calcitriol), the most active product of vitamin and lower rates in the south, southwest and west in the D synthesis, is in the physiological regulation of Ca2+ and United States led to the important concept that exposure 1 Pi transport and bone mineralization . The importance of to ultraviolet B or sunlight, which leads to vitamin D this role is demonstrated by studies with knockout mice synthesis, can reduce the risk of colorectal cancer. Several deficient in key members of the vitamin D metabolic epidemiological observations have shown an associa-

pathway, such as 25-hydroxyvitamin D3‑1α-hydroxylase tion between low serum 25(OH)D3 levels (the accepted (1α-OHase; encoded by Cyp27b1), an enzyme that gener- measure of vitamin D body stores) and increased risk for 13 14 15 ates 1α,25(OH)2D3, and 25-hydroxyvitamin D 24-hydrox- colorectal , breast and prostate cancers. There are ylase (24-OHase; encoded by Cyp24a1), the enzyme that many epidemiological studies that have sought to

degrades 1α,25(OH)2D3 (catabolism), and the vitamin D determine associations between vitamin D status and 16–19 receptor (Vdr), which binds 1α,25(OH)2D3 to affect target the risk and mortality rates of a number of cancers . gene transcription2–6 (TABLEs 1,2). Loss of these genes in Giovannucci and colleagues16 recently performed an mice led to phenotypes with abnormal bone morphol- extensive analysis that combined major determinants ogy. However, recent observations indicate a much of vitamin D status on cancer risk and mortality with

broader range of action for 1α,25(OH)2D3, including the 51,529 men that were enrolled in the Health Professionals regulation of differentiation, proliferation and apoptosis. Follow-up Study (HPFS). Individuals were prospectively Furthermore, altered expression and function of proteins followed for almost 20 years; diet, exercise and lifestyle crucial in vitamin D synthesis and catabolism have been characteristics were analysed and health outcomes, observed in many tumour types (TABLE 3). Interestingly, including cancer and cancer-related deaths, were Vdr–/– mice show hyperproliferation and increased mitotic assessed. Giovannucci et al. developed a model to pre-

Departments of activity in the descending colon, suggesting a role for dict 25(OH)D3 levels based on the relationship between Pharmacology and 1α,25(OH)2D3-mediated signalling in tumour suppres- dietary and supplementary vitamin D, physical activity, Therapeutics* and Medicine‡, sion7. Zinser and colleagues8–10 showed that Vdr ablation body mass index and sunlight exposure (a source of vita- Roswell Park Cancer Institute, in the mouse increased chemical carcinogenesis in mam- min D). The model was applied to 47,800 individuals in Buffalo, New York, USA. mary, epidermis and lymphoid tissues but not in ovary, the HPFS, and the analysis indicated a strong association Correspondence to C.S.J. (TABLE 2) e-mail: candace.johnson@ uterus, lung or liver . However, mice deficient in between low levels of predicted 25(OH)D3 and increased roswellpark.org key members of the vitamin D synthesis and catabolic cancer incidence and cancer-related mortality, particu- doi:10.1038/nrc2196 pathways do not develop spontaneous cancer2,3,11. larly for cancers of the digestive system16. Furthermore,

REVIEWS

At a glance resultant 25-hydroxycholecalciferol (25(OH)D3) is 1α-hydroxylated in the kidney by mitochondrial • Epidemiological studies point to a relationship between vitamin D deficiency and 1α-hydroxylase (1α-OHase; encoded by the gene cancer risk. CYP27B1), this yields the hormonally active secosteroid • Alterations in vitamin D receptor expression, and in the synthesis (25-hydroxylase 23 1α,25(OH)2D3 (calcitriol) . 24-hydroxylation of and 1α-hydroxylase) and catabolism (24-hydroxylase) of vitamin D metabolites are 25(OH)D and 1α,25(OH) D by the cytochrome involved in the growth regulation of tumours; thus, compromising 1α,25(OH) D 3 2 3 2 3 P450 enzyme 25-hydroxyvitamin D 24-hydroxylase (also known as calcitriol; the active metabolite of vitamin D signalling) sensitivity (24-OHase; encoded by the gene CYP24A1), to the and 1α,25(OH) D signalling. 2 3 metabolites 24,25(OH) D and 1α,24,25(OH) D , • The antiproliferative effects of 1α,25(OH) D have been demonstrated in various 2 3 2 3 2 3 respectively, is the rate-limiting step for 25(OH)D tumour types, as determined by preclinical trials. 3 and 1α,25(OH) D catabolism23. Additionally, • The anti-tumour effects of 1α,25(OH) D involve mechanisms that are associated 2 3 2 3 1α,25(OH) D concentrations are feedback regulated: with G0/G1 arrest, differentiation, induction of apoptosis and modulating 2 3 an increase in 24,25(OH) D induces the synthesis of different signalling pathways in tumour cells, as well as inhibiting tumour 2 3 α 2+ α angiogenesis. 1 ,25(OH)2D3; whereas Ca , Pi and 1 ,25(OH)2D3 itself suppress 1α,25(OH) D synthesis23–26. CYP27B1 • Glucocorticoids potentiate the anti-tumour effects of 1α,25(OH) D and decrease 2 3 2 3 α 1 ,25(OH) D -induced hypercalcemia. 1 ,25(OH) D also potentiates the anti- (which encodes 1 -OHase) expression is induced α 2 3 α 2 3 25 tumour effects of many chemotherapeutic agents such as platinum analogues, by parathyroid hormone (PTH) and repressed 24,27 taxanes and DNA-intercalating agents. by 1α,25(OH)2D3 . Furthermore, CYP24A1 is strongly induced by 1α,25(OH) D to produce the less • Given that the major vitamin D catabolizing enzyme, CYP24A1 (24-hydroxylase), is 2 3 often amplified and overexpressed in tumour cells, agents that inhibit this enzyme active vitamin D metabolites 1α,24,25(OH)2D3 and 24,25(OH) D 23. can potentiate 1α,25(OH)2D3 anti-tumour effects. 2 3 • Preclinical data indicate that maximal anti-tumour effects are seen with There are instances of tissue-specific regulation of the vitamin D synthetic enzymes. 1α,25(OH) D pharmacological doses of 1α,25(OH)2D3, and can be safely achieved in animals 2 3 using a high-dose, intermittent schedule of administration. Some clinical trial data functions in an autocrine and paracrine manner to α indicates that 1α,25(OH)2D3 is well-tolerated in cancer patients within a proper modulate vitamin D function and signalling. 1 - dosing schedule. OHase is expressed at extra-renal sites such as nor- • Data support the hypothesis that vitamin D compounds may have an important mal colon, brain, placenta, pancreas, lymph nodes 28 role in cancer therapy and prevention, and merit further investigation. and skin , allowing local conversion of 25(OH)D3

to 1α,25(OH)2D3. Importantly, increased CYP27B1 expression is observed in breast29 and prostate30 can- Giovannucci et al. reported that an increase of 25 nmol cers and during early colon tumour progression in

per L in predicted 25(OH)D3 level is associated with a well-to-moderately differentiated states, but decreased 29% reduction in cancer-related mortality and a 17% in poorly differentiated colon carcinomas 31–33. reduction in cancer incidence, suggesting that high Increased expression of CYP27B1 in cancer tis-

25(OH)D3 levels might be associated with a decreased risk sues could provide local conversion of 25(OH)D3 16 of some cancers . A meta-analysis of case–control and to 1α,25(OH)2D3, and may support the notion that

cohort studies found that individuals with ≥ 33 ng per ml 25(OH)D3 and 1α,25(OH)2D3 might have a role in the

(82 nmol per L) 25(OH)D3 had a 50% lower incidence of chemoprevention of these cancers. However, CYP24A1 colorectal cancer20. Additionally, patients with early stage (encoding 24-OHase) mRNA expression is upregu-

non-small-cell lung cancer with high 25(OH)D3 levels lated in tumours, and may counteract 1α,25(OH)2D3 and high vitamin D intake at the time of diagnosis and antiproliferative activity, presumably by decreasing 34,35 (TABLE 3) 35 initiation of treatment had improved overall and recur- 1α,25(OH)2D3 levels . Cross et al. have rence-free survival21. Therefore, these data suggest that demonstrated that the upregulation of CYP24A1 and

low levels of 25(OH)D3 are an important risk factor for downregulation of CYP27B1 can occur in high-grade cancer incidence. colon carcinomas. The chromosomal region 20q13.2, containing the CYP24A1 gene, is amplified in human Synthesis and catabolism of vitamin D breast tumours36, and CYP24A1 mRNA expression

1α,25(OH)2D3 is synthesized from vitamin D in a is upregulated in samples from human lung, colon Secosteroid hormones highly regulated multistep process (FIG. 1). The first and ovarian tumours, suggesting that 1α,25(OH) D Molecules that are very similar 2 3 35,37,38 (TABLE 3) in structure to steroids but with step in vitamin D synthesis is the formation of vita- levels would be reduced in these cases . a ‘broken’ ring; two of the min D3 in the skin through the action of ultraviolet This suggests that inhibition of CYP24A1 expression B‑ring carbon atoms (C-9 and irradiation; vitamin D3 can also be taken in the diet and activity is essential for prevention to be effective. 10) of the four steroid rings are but in North America and Europe dietary vitamin D3 Small-molecule inhibitors with varying specificity for not joined. 39–42 intake is a minor component of vitamin D3 acquisi- 24-OHase render tumour cells more sensitive to the

Autocrine tion because dairy products, eggs, fish and fortified action of 1α,25(OH)2D3 and its analogues. Consistent A substance secreted by a cell foods contain only small quantities of vitamin D22. with the epidemiological studies discussed above, these that acts on the surface Decreased sun exposure further limits vitamin D findings indicate that 1α,25(OH)2D3 catabolism could receptors of the same cell. synthesis. modulate tumour growth in some tissues, indicat-

Paracrine Vitamin D3 (cholecalciferol) is hydroxylated by ing the potential for the development of 24-OHase A substance secreted by a cell liver mitochondrial and microsomal 25-hydroxylases inhibitors as cancer preventative and/or anticancer that acts on adjacent cells. (25-OHase)23, encoded by the gene CYP27A1. The therapeutic agents.

REVIEWS

Table 1 | Mouse models of vitamin D metabolic enzymes and receptor signalling Genetic modification Mouse phenotype Cancer phenotype Refs Cyp27b1–/– (which Pseudo-vitamin D deficiency rickets (PDDR); decreased serum Ca2+ and None 2,3

encodes 1α-OHase) Pi; secondary hyperthyroidism; undetectable 1α,25(OH)2D3 (calcitriol) levels; disorganized growth plate structure and osteomalacia. In addition: infertile females; uterine hypoplasia; decreased ovarian size; compromised folliculogenesis; reduction in CD4+ and CD8+ peripheral T lymphocyte Cyp24a1–/– (24-OHase) Lethal hypercalcemia; impaired intramembranous bone mineralization None 11 Vdr–/– (VDR signalling) Vitamin D deficiency rickets type II (VDDR II) and osteomalacia; alopecia, Hyperproliferation of descending 4,5,6,7 hypocalcemia, hyperparathyroidism; impaired bone formation, growth colon (increased PCNA positivity retardation; female infertility, uterine hypoplasia, impaired folliculogenesis and cyclin D1 expression) PCNA, proliferating cell nuclear antigen.

48 1α,25(OH)2D3-mediated transcription of target genes. in a transcriptionally repressed state . Conformational

1α,25(OH)2D3 exerts transcriptional activation and change also repositions the VDR activation function repression of target genes by binding to the VDR (BOX 1). 2 (AF2) domains to bind to stimulatory coactivators, The VDR is a member of the steroid hormone receptor consisting of the steroid receptor coactivators (SRCs), superfamily and regulates gene expression in a ligand- nuclear coactivator 62 kDa–SKI-interacting protein dependent manner43 (FIG. 2). Interestingly, N‑terminal VDR (NCoA62–SKIP) and the chromatin modifiers, CREB variants show tissue-specific expression44,45 that might also binding protein (CBP)–p300 and PBAF (polybromo- and

contribute to the differential specificity of 1α,25(OH)2D3- SWI‑2-related gene 1 associated factor), which acetylates

mediated regulation. 1α,25(OH)2D3–VDR-dependent histones in the nucleosomes to unravel DNA for transcrip- transcriptional activity is modulated through synergistic tion49. Once the chromatin is de-repressed, the vitamin ligand-binding and dimerization with retinoic X receptor D receptor-interacting proteins (DRIPs) form a complex

(RXR). The activated 1α,25(OH)2D3–VDR–RXR com- that binds to the AF2 domain of VDR and interacts with plex specifically binds to vitamin D response elements the transcription machinery, such as TF2B (transcription (VDREs), composed of two hexanucleotide repeats inter- factor 2B) and RNA polymerase II, and initiates transcrip- spaced by varying numbers of nucleotides (for example, tion (for a review see REF. 23). Recently, it has been shown GGTCCA-NNN-GGTCCA, where N is any nucleotide; that epigenetic regulation of VDR through increased this is denoted DR3), in the promoter regions of target expression of NCoR1 and SMRT repress VDR-mediated genes46. For transcriptional activation, VDR occupies the 3′ signalling in prostate50 and breast cancer51 cell lines, and half-site whereas RXR binds the 5′ half-site of VDRE47. may have a role in mediating the antiproliferative effects

Co-factor proteins also have the ability to modulate of 1α,25(OH)2D3 in these tissues.

VDR-mediated gene expression; these proteins possess The mechanism by which 1α,25(OH)2D3 represses intrinsic chromatin-modifying enzymatic activities, act gene expression through the binding of VDR to negative as a platform for the recruitment of chromatin-modifying VDREs (DR3-type), placing VDR on the 5′ half-site of proteins and recruit basal transcription factors to the pro- the VDRE52,53, such as is the case with human PTH, may 23 moters . 1α,25(OH)2D3 binding induces phosphoryla- involve interference with transcriptional machinery but tion and conformational changes in VDR, which causes is less understood. Recently, transcriptional repression

the release of co-repressors (such as nuclear receptor by 1α,25(OH)2D3 has been further elucidated for the co-repressors (NCoRs) and the silencing mediator for human CYP27B1 (REFS 27,54,55) and PTH56 genes. The retinoid and thyroid hormone receptors (SMRT)–histone VDR–RXR heterodimer represses gene transcription in

deacetylase (HDAC) complex) that maintain chromatin a 1α,25(OH)2D3-dependent manner through E‑box-type

Table 2 | Vdr knockout mice and carcinogenesis Oncogene/ Tissue Cancer phenotype Ref carcinogen MPA plus DMBA Skin 40% sebaceous, 25% squamous and 15% follicular papillomas 9 carcinogens compared with WT littermates; other infrequent lesions include basal cell carcinoma and haemangioma Mammary Higher incidence of alveolar and ductal hyperplasias in 8 Vdr–/– mice compared with WT mice; development of palpable mammary tumours was not altered by Vdr ablation Lymph nodes and/ Lymphoblastic and thymic lymphoma higher in Vdr–/– (27%) 8 or thymus compared with WT mice (11%) Vdr–/– ;neu oncogene Mammary Decreased survival of Vdr–/–; neu mice compared with their 10 Vdr+/+; neu and Vdr+/–; neu littermates; increased development of mammary tumours driven by the neu oncogene DMBA, 7,12-dimethylbenzanthracene; MPA, medroxyprogesterone acetate; WT, wild-type.

REVIEWS

Table 3 | Expression of molecules that function in vitamin D metabolism and signalling in human cancers. Protein Altered expression observed Prognostic or histological Types of cancer (gene) in human cancer tissues observations Vitamin D metabolic enzymes 25-OHase Increased mRNA NA Breast34, cervical34 and ovarian cancer34, HCC172 (CYP27A1) 1α-OHase Increased mRNA NA Basal cell carcinoma173, breast34,174, cervical34 (CYP27B1) and ovarian cancer34 Increased mRNA Moderately differentiated Colon cancer31,35,175 Decreased mRNA Poorly differentiated Colon cancer35 Splice variants (Hyd‑V5, ‑V6, ‑V7 NA Glioblastoma multiforme176, melanoma177, and ‑V8) cervical cancer177 Immunoreactivity NA Pancreatic132, breast132 and colon cancer 33, renal cell carcinoma132 Increased immunoreactivity Moderately differentiated Colon cancer33,35 Decreased immunoreactivity Poorly differentiated Colon cancer33,35 24-OHase Amplified at 20q13.2 locus NA Gastric adenocarcinoma37, breast cancer36 (CYP24A1) Increased mRNA NA Basal cell carcinoma173, SCC (cutaneous)178, lung38,42, breast34,36, colon35,38, cervical34 and ovarian cancer34,38 Increased mRNA Poor prognosis Oesophageal cancer179 Decreased mRNA NA Breast cancer38 Increased mRNA and activity Poorly differentiated Colon cancer32 Increased protein NA Lung cancer (NSCLC)42 Vitamin D receptor VDR Increased mRNA NA Basal cell carcinoma173, SCC (cutaneous)178, (VDR) colon cancer31 Decreased mRNA Poorly differentiated Colon cancer31 Increased immunoreactivity NA Breast34, cervical34 and ovarian cancer34 Increased, predominantly Well differentiated Colon cancer166 cytoplasmic Decreased immunoreactivity Moderately and poorly Colon cancer166 differentiated Poorly differentiated Colon cancer35 HCC, hepatocellular carcinoma; NA, non applicable; NSCLC, non-small cell lung carcinoma; SCC, squamous cell carcinoma.

55 negative VDREs, comprised of a CANNTG-like motif control of 1α,25(OH)2D3 biosynthesis , as well as negative in the promoter regions of the CYP27B1 (REFS 27,54,55) regulation of other genes with nVDREs in their promot- and PTH56 genes, which are distinct from the DR3- ers. Furthermore, Kim et al.57 demonstrated that not only type response elements. VDR-interacting repressor is histone deacetylation crucial for chromatin structure (VDIR), when bound to E‑box-type elements, induces remodelling in suppression of the CYP27B1 gene, but the transcriptional activation of CYP27B154. However, that transrepression by VDR requires DNA methylation

the binding of 1α,25(OH)2D3 to VDR causes VDR to of the CYP27B1 gene promoter, suggesting complicated

interact with VDIR. 1α,25(OH)2D3-induced association epigenetic modifications for transcriptional regulation between VDR and VDIR induces dissociation of the of the CYP27B1 gene. Epigenetic regulation of CYP27B1 histone acetyltransferase (HAT) co-activator and recruit- and CYP24A1 has been previously reported for the PNT‑2

ment of HDAC co-repressor for 1α,25(OH)2D3-induced human normal prostate cells and DU‑145 prostate cancer transrepression of CYP27B1 gene expression54. In addi- cell line58. Histone methylation and demethylation are cru- tion, Williams syndrome transcription factor (WSTF) cial events that impose ligand- and signal-dependent gene

potentiates 1α,25(OH)2D3-induced transrepression by activation by nuclear receptors and prevent the recruit- VDR of the CYP27B1 gene promoter by facilitating the ment of unliganded nuclear receptors and transcription association between WINAC, a multifunctional, ATP- factors from binding to their target promoters and causing dependent chromatin-remodelling complex, and chro- constitutive gene activation59. matin55. This transrepression mechanism is an important Examples of genes with DR3-type response elements

REVIEWS

UV-B

D3 DBP

7-dehydrocholesterol D3 D3 Circulation

Pre-D3

D3 Skin Intestine Liver P , Ca2+ and 25-OHase i D other factors 3 24-OHase 25(OH)D3 +/– 24,25(OH)2D3

Parathyroid PTH Kidney + Excretion glands 1α,24,25(OH)2D3

1 -OHase α Dietary sources 24-OHase of vitamin D 1α,25(OH)2D3 Intestine Tumour Increases microenvironment absorption • Inhibits proliferation of Ca2+ Bone Immune cells • Induces differentiation and Pi Increases bone Induces • Inhibits angiogenesis mineralization differentiation

Figure 1 | Vitamin D metabolism. Photochemical synthesis of vitamin D3 (cholecalciferol, D3) occurs cutaneously

where pro-vitamin D3 (7-dehydrocholesterol) is converted to pre-vitamin D3 (pre‑D3) in response to ultraviolet B Nature Reviews | Cancer (sunlight) exposure. Vitamin D3, obtained from the isomerization of pre-vitamin D3 in the epidermal basal layers or intestinal absorption of natural and fortified foods and supplements, binds to vitamin D‑binding protein (DBP) in the

bloodstream, and is transported to the liver. D3 is hydroxylated by liver 25-hydroxylases (25-OHase). The resultant

25‑hydroxycholecalciferol (25(OH)D3) is 1α-hydroxylated in the kidney by 25-hydroxyvitamin D3‑1α-hydroxylase

(1α‑OHase). This yields the active secosteroid 1α,25(OH)2D3 (calcitriol), which has different effects on various target 23 tissues . The synthesis of 1α,25(OH)2D3 from 25(OH)D3 is stimulated by parathyroid hormone (PTH) and suppressed by 2+ Ca , Pi and 1α,25(OH)2D3 itself. The rate-limiting step in catabolism is the degradation of 25(OH)D3 and 1α,25(OH)2D3

to 24,25(OH)D3 and 1α,24,25(OH)2D3, respectively, which occurs through 24-hydroxylation by 25-hydroxyvitamin D 24-

hydroxylase (24-OHase), encoded by the CYP24A1 gene. 24,25(OH)D3 and 1α,24,25(OH)2D3 are consequently excreted.

The main effects of 1α,25(OH)2D3 on various target tissues are highlighted above.

(REF. 60) consist of CYP24A1 (encoding 24-OHase), Nongenomic action of 1α,25(OH)2D3. Nongenomic 61 BGLAP (osteocalcin; expressed in bone osteoblasts), actions mediated by 1α,25(OH)2D3 are rapid and not and CDKN1A62 (which encodes the cyclin depend- dependent on transcription. However, nongenomic ent kinase (CDK) inhibitor p21). Those repressed signalling may indirectly affect transcription through 53 68,69 by 1α,25(OH)2D3 include PTH . Although VDREs cross-talk with other signalling pathways . Although are traditionally thought to occur in the promoter there is no agreement on how the nongenomic actions regions of the target genes, a DR3-type VDRE was are initiated, data suggest that these effects begin at the recently identified in exon 4 of the growth arrest plasma membrane and involve a non-classical membrane and DNA-damage-inducible (GADD45) gene63. receptor (memVDR; FIG. 2) described in intestinal caveo- 70 1α,25(OH)2D3-mediated repression or activation of lae , and a 1α,25(OH)2D3-membrane-associated rapid-

many proto-oncogenes or tumour-suppressor genes response steroid binding protein (1α,25D3-MARRS) is described in normal and tumour tissues62,64–67; isolated from chick intestinal basal-lateral membrane71. however, only a few such genes contain VDREs in the The most well-described nongenomic effect of 2+ promoter regions and are under the direct transcrip- 1α,25(OH)2D3 is the rapid intestinal absorption of Ca 62 (REF. 72) tional control of 1α,25(OH)2D3, such as CDKN1A . Binding of 1α,25(OH)2D3 to the proposed mem- and CCNC (which encodes cyclin C, containing a brane receptor can result in the activation of numerous 65 68,69 (FIG. 2) DR4-type VDRE) . This suggests that 1α,25(OH)2D3 signalling cascades . Activation of these signalling exerts many of its effects indirectly by modulating cascades, such as protein kinase C (PKC), can result in the signalling cascades or by unknown nongenomic rapid opening of voltage-gated Ca2+ channels and an increase mechanisms (FIGS 2,3). in intracellular Ca2+ (REF. 73), which may subsequently

REVIEWS

Box 1 | The vitamin D receptor The human VDR gene (which encodes the vitamin D receptor), located on chromosome 12q, is composed of promoter and regulatory regions (1a–1f) and exons 2–9, which encode 6 domains (A – F) of the full length VDR protein (see figure)23. VDR nuclear localization signals (blue) direct the receptor into the nucleus155,156 along microtubule tracts to the nuclear pores157.

Upon 1α,25(OH)2D3 binding to the hormone ligand-binding domain (red), VDR is stabilized by the phosphorylation of serine 51 in the DNA-binding domain (green) by protein kinase C76, and serine 208 in the hinge region by casein kinase II158.

VDR associates with the retinoic acid receptor (RXR) through the dimerization domains (yellow). The 1α,25(OH)2D3–VDR– RXR complex binds to the vitamin D response elements (VDREs) through the DNA-binding domain in the promoters of target genes. Conformational change in the VDR results in the dissociation of the co-repressor, silencing mediator for retinoid and thyroid hormone receptors (SMRT), and allows interaction of the VDR activation function 2 (AF2) transactivation domain (light grey) with stimulatory coactivators, such as steroid receptor coactivators (SRCs), vitamin D receptor-interacting proteins complex and nuclear coactivator‑62 kDa–Ski-interacting protein (NCoA62–SKIP)23 that mediate transcriptional activation. Non-synonymous (FokI) and synonymous (BsmI, ApaI, TaqI and Tru9I) single-nucleotide polymorphisms (SNPs) have been identified in VDR (defined by restriction enzymes, polymorphisms are indicated in parentheses). FokI polymorphism at translation initiation codon results in a smaller VDR that interacts with transcription factor 2B (TF2B) more efficiently and has greater transcriptional activity than the full length VDR159. Although the functional effects of these SNPs remain unknown, they have been reported to be associated with increased susceptibility to primary and metastatic breast cancer17, squamous cell carcinoma160, colorectal cancer161,162, and prostate cancer163,164, but may be protective against head and neck cancer165.

The expression of VDR is an important determinant of the tumour cell response to 1α,25(OH)2D3. The VDR is overexpressed or repressed in several histological types of cancer (TABLE 3), demonstrating tissue-type variations in

1α,25(OH)2D3 signalling (supplemental information S1 (table)). VDR expression increases in hyperplastic polyps and in the early stages of tumorigenesis, but declines in late-stage poorly differentiated tumours and is absent in associated

metastases. Tumours of the colon with the highest expression of VDR were most responsive to 1α,25(OH)2D3 treatment85,166. However, downregulation of the VDR in colon cancer cells through the transcription factor SNA1L167 reduces the anticancer effect of the vitamin D analogue EB1089. VDR gene Chromosome 12 1f 1e 1a 1d 1b 1c 2 3 4 5 6 7 8 9 q13–14 ~75 kb Bsml (A60890G) Fokl (C27823T) Tru9l (G61050A) Apal (G61888T) Taql (T61938C) DNA binding (aa 24–90, 91–115) VDR protein Nuclear localization S51 S208 P P (aa 49–55, 79–105) AF-2 Hormone ligand binding N Hinge region 48 kDa (aa 227–244, 268–316, 396–422) 1 24 49 91 115 227244268 317 396 422 427 aa Dimerization (aa 37, 91–92, 244–263, 317–395) A/B C D E/F Transactivation (aa 246, 416–422)

Nature Reviews | Cancer activate the Raf–mitogen-activated protein kinase extracel- Anti-tumour effects of 1α,25(OH)2D3 signalling

lular signal-regulated kinase kinase (MEK)–mitogen-acti- 1α,25(OH)2D3 has been examined preclinically for its vated protein kinase (MAPK)–extracellular signal-regulated therapeutic efficacy in chemopreventive and anticancer kinase (ERK) cascade in skeletal muscle cells74. Activation activity. A chemoprevention study used Nkx3‑1;Pten of the Raf–MEK–MAPK–ERK cascade, which mediates mutant mice to recapitulate prostate carcinogenesis,

proliferative cellular effects, may be a response to increased and showed that 1α,25(OH)2D3 administration delayed Ca2+ in normal colon73 and skeletal muscle cells74, and may the onset of prostate intraepithelial neoplasias (PIN) and not have a direct role in the antiproliferative activities of had better anti-tumour activity when administered to

1α,25(OH)2D3 in tumour cells (discussed below). In addi- mice with early-stage (PIN) rather than advanced-stage tion, ERK can also increase the transcriptional activity of prostate disease77. Furthermore, studies using model the VDR75, and nongenomic activation of PKC may stabi- systems of squamous cell carcinoma (SCC)78, prostate lize VDR (through phosphorylation)23,76, thereby affecting adenocarcinoma79, cancers of the ovary80, breast81 and 82 the transcriptional activity of the receptor. Therefore, the lung showed that the administration of 1α,25(OH)2D3 nongenomic activation of these pathways may cooperate or vitamin D analogues had significant anticancer effects.

with the classical genomic pathway to transactivate VDR The effects of 1α,25(OH)2D3 and its derivatives have

and elicit the antiproliferative effects of 1α,25(OH)2D3, but been shown to function through the VDR to regulate this remains to be elucidated. proliferation, apoptosis and angiogenesis62,83–87.

REVIEWS

SOC 1α,25(OH)2D3 GPCR 1α,25(OH)2D3 channels

P VDR 1 ,25(OH) D α 2 3 PLCγ d AC Caveolae Ca2+ PI3K

[cAMP] mem ? PKC VDR Ras 9cRA PKA RXR P VDR Raf isoforms

MEK1/2 PBAF a SWI/SNF CBP/ ERK–MAPK1/2 P300 Chromatin NCoA62– remodeling 1α,25(OH)2D3 SKIP 9cRA SRC-1 (histone P Nucleus RXR VDR acetylation) Transcriptional 9cRA P Cross-talk activation RXR VDR Transcriptional repression 5′ 3′ VDREs NCoA62– b Gene SKIP transcription IP205 DR DRIPs RNA c 9cRA TF2B CDKN1A WINAC P Pol II CYP24A1 HDAC Gene RXR VDR SPP1 complexes NCoR– SMRT WSTF repression CYP27B1 Chromatin 9cRA 5′ 3′ P PTH remodelling RXR VDR VDREs (histone VDIR deacetylation) nVDREs

Nature Reviews | Cancer Figure 2 | 1α,25(OH)2D3-mediated transcriptional regulation. Classical action of 1α,25(OH)2D3 is mediated by binding of the vitamin D receptor (VDR)−9-cis-retinoic acid receptor (RXR) complex at the vitamin‑D response elements (VDREs). a | Transcriptional activation involves the co-activators, steroid receptor coactivators (SRCs), nuclear coactivator‑62 kDa–Ski-interacting protein (NCoA62–SKIP) and histone acetyltransferases (HATs), CREB binding protein (CBP)–p300 and polybromo- and SWI‑2-related gene 1 associated factor (PBAF–SNF) to acetylate histones to derepress chromatin. b | Binding of the vitamin D receptor-interacting protein 205 (DRIP205) to the activation function 2 (AF2) of VDR (and RXR) attracts a mediator complex containing other vitamin D receptor- interacting proteins (DRIPs) that bridge the VDR–RXR–NCoA62–SKIP–DRIP205 complex with transcription factor 2B (TF2B) and RNA polymerase II (RNA Pol II) for transcription initiation. The presence of the multiprotein complex facilitates increased transcription of genes, such as CDKN1A (which encodes the cyclin-dependent kinase inhibitor p21), CYP24A1 (which encodes 24-OHase) and SPP1 (which encodes osteopontin)23.

c | 1α,25(OH)2D3-mediated transcriptional repression involves VDR–RXR heterodimer association with VDR- interacting repressor (VDIR) bound to E‑box-type negative VDREs (nVDREs), dissociation of the HAT co-activator and recruitment of histone deacetylase (HDAC) co-repressor54. Williams syndrome transcription factor (WSTF) potentiates transrepression by interacting with a multifunctional, ATP-dependent chromatin-remodelling complex (WINAC) and chromatin55. This leads to the repression of genes, such as CYP27B1 (which encodes 1α-

OHase) and PTH (which encodes parathyroid hormone). d | Non-genomic, rapid actions of 1α,25(OH)2D3 are

hypothesized to involve 1α,25(OH)2D3 binding to cytosolic (VDR) and membrane VDR (memVDR), also found in caveolae, and speculated to activate the mitogen-activated protein kinase (MAPK)–extracellular signal-regulated kinase (ERK) 1 and 2 cascade68 through the phosphorylation (P) and activation of Raf by protein kinase C (PKC) by 2+ 2+ 2+ Ca influx through store-operated Ca (SOC) channels. 1α,25(OH)2D3 stimulates SOC Ca influx (in muscle cells) by trafficking of the classic VDR to the plasma membrane, where the VDR interacts with the SOC channel. Ca2+ 2+ influx activates Ca messenger systems, such as PKC. Activated PKC can phosphorylate VDR. 1α,25(OH)2D3 binding to G‑protein coupled receptors (GPCRs) activates phospholipase Cγ (PLCγ), Ras, phosphatidylinositol 3‑kinase (PI3K) and protein kinase A (PKA) pathways, and induces MAPK–ERK1 and 2 signalling. Activated Raf– MAPK–ERK may engage in cross-talk with the classical VDR pathway to modulate gene expression. AC, adenylate cyclase; cAMP, cyclic adenosine monophosphate.

REVIEWS

a b c d e f g

1α,25(OH) D IGF1 2 3

TGFβ ↑E-cadherin Wnt R1 R2 β β F F GF1R EGFR T TG TG Frizzled β-catenin

Ras P13K Cytosol BCL2 BAX SMADs β-catenin Raf isoforms BCL-X BCL-X APC Akt L S β-catenin VDR MEK1/2 Degradation

Telomerase Effector caspases ERK–MAPK1/2

β-catenin MYC p15 p21 p27 SKP2 Apoptosis TCF4 MYC TCF1 Differentiation CD44 Apoptosis CDK4/6 CDK2 Degradation PARG Growth inhibition Cyclin D1,2,3 Cyclin E

p107/p130 Nucleus

P P P E2F4/5 pRB pRB + DP1 Cell cycle E2F1,2,3

Growth arrest Figure 3 | Key cancer-related signalling pathways targeted by 1α,25(OH) D . 1α,25(OH) D inhibits mitogen-activated 2 3 2 3 Nature Reviews | Cancer protein kinase (MAPK)–extracellular signal-regulated kinase (ERK) 1 and 2 signalling through suppression of epidermal

growth factor (EGFR; a) and insulin-like growth factor 1 (IGF1; b), which both target Ras. 1α,25(OH)2D3 induces apoptosis through the IGFR1−phosphatidylinositol 3‑kinase (PI3K)−Akt-dependent signalling pathway (b), inhibiting telomerase (c), downregulating BCL2, inducing BAX and activating caspase cleavage (d). Cell-cycle progression is perturbed by

1α,25(OH)2D3 through S‑phase kinase-associated protein ubiquitin ligase (SKP2; targeting p27 for degradation; e), and MYC, which results in pRB dephosphorylation; and transforming growth factor-β (TGFβ; f) cross-talk. Cell-cycle

perturbation by 1α,25(OH)2D3 ultimately affects the association of retinoblastoma pocket proteins (pRB and p107/p130) and the E2F family of transcription factors and DP polypepitide (DP1) heterodimers that mediate the transcription of cell- cycle genes. Association of E2F1, 2 and 3 with pRB in its hypophosphorylated state and interaction of the E2F4 and 5 transcriptional repressors and DP1 with p107/p130 prevent transcription of cell-cycle genes and restrain cell-cycle

progression. Activation of VDR by 1α,25(OH)2D3 induces the expression of E‑cadherin (g), thereby promoting the translocation of β‑catenin from the nucleus to the plasma membrane and competing with T-cell transcription factor 4 (TCF4) for β‑catenin binding; thus inhibiting the Wnt–β-catenin–TCF4 signalling pathway, which leads to the induction of MYC, TCF1 (transcription factor 1), CD44 and PPARG (peroxisome proliferator-activated receptor-γ). APC, adenomatosis polyposis coli; CDK, cyclin-dependent kinase; pRB, phosphorylated retinoblastoma; Wnt, wingless-related MMTV integration site.

Antiproliferative effects of 1α,25(OH)2D3. Cell-cycle contain VDREs, and their transcriptional activation

perturbation is central to 1α,25(OH)2D3-mediated or repression may not be directly mediated by VDR.

antiproliferative activity in tumour cells (supplemen- 1α,25(OH)2D3–VDR transcriptional activation of tal information S1 (table)). Progression through the CDKN1A induces cell-cycle exit (differentiation) and cell cycle is regulated by cyclins, and their association cell-cycle arrest in human U937 myelomonocytic cells62. with CDKs and CDK inhibitors (CKIs). Expression Treatment of human breast cancer MCF7 cells with

of the CKIs p21 and p27 inhibits proliferation, in part 1α,25(OH)2D3 also increases the expression of CDKN1A by inducing G1 cell-cycle arrest and withdrawal from and CDKN1B, (which encodes p27) and represses the cell cycle (G0). CDKN1A and GADD45A contain CCND1 (encoding cyclin D1), CCND3 (encoding cyc- a functional VDRE and are direct transcriptional tar- lin D3), CCNA1 (which encodes cyclin A1) and CCNE1

gets of 1α,25(OH)2D3–VDR. However, many genes are (which encodes cyclin E1), and hence leads to the inhibi- 88,89 transcriptionally affected by 1α,25(OH)2D3 but do not tion of CDK activity and pRb hypophosphorylation . nature reviews | cancer volume 7 | SEPTEMBER 2007 | 691 © 2007 Nature Publishing Group

REVIEWS

102 Similarly, the treatment of SCC cells with 1α,25(OH)2D3 of 1α,25(OH)2D3 . Recent findings reported by Palmer 103 induces G0/G1 cell-cycle arrest owing to the tran- et al. indicate that 1α,25(OH)2D3 promotes differentia- scriptional activation of CDKN1B and consequent tion through the induction of CDH1 (which encodes E pRb hypophosphorylation90. However, in this context cadherin) in adenomatosis polyposis coli (APC)-mutated CDKN1A expression was repressed, indicating that the human colorectal cancer SW480 cells. CDH1 activation

cell-cycle arrest is an indirect effect of 1α,25(OH)2D3 consequently restrained cell growth by facilitating the treatment or that cell-type specificity might determine translocation of β‑catenin from the nucleus to the plasma

the ability of activated 1α,25(OH)2D3–VDR to induce membrane, thus inhibiting β‑catenin-mediated tran- CDKN1A expression90. Other genes have been shown scription and allowing activated VDR to compete with

to be transcriptionally affected by 1α,25(OH)2D3 in β‑catenin for transcription factor binding. Again, there colon cancer, ovarian carcinoma and leukaemia cells, appears to be no specific mechanism regarding the ability (REF 63) such as activation of GADD45 , which is involved of 1α,25(OH)2D3 to induce differentiation in tumour cells in DNA damage responses, repression of TYMS (supplemental information S1 (table)). (which encodes thymidylate synthetase)91 and TK1 (which encodes thymidine kinase)91, which are involved Apoptosis. In addition to the antiproliferative effects

in DNA replication, and activation of the INK4 fam- of 1α,25(OH)2D3, there is increasing evidence that 92 ily of cyclin D‑dependent kinase inhibitors, which 1α,25(OH)2D3 exerts anti-tumour effects by regulat- mediate G1 cell-cycle arrest; whereas cyclin E–CDK2 ing key mediators of apoptosis, such as repressing the and the SKP2 (S-phase kinase-associated protein 2) expression of the anti-apoptotic, pro-survival proteins

ubiquitin ligase, which targets CKIs to the proteasome, BCL2 and BCL-XL, or inducing the expression of pro- 93 are downregulated by 1α,25(OH)2D3. 1α,25(OH)2D3 apoptotic proteins (such as BAX, BAK and BAD). It

treatment also results in the repression of the proto- has been reported that 1α,25(OH)2D3 downregulates oncogene MYC89,94, which significantly contributes to BCL2 expression in MCF‑7 breast tumour and HL‑60

the antiproliferative effects of 1α,25(OH)2D3. leukaemia cells and upregulates BAX and BAK expres-

1α,25(OH)2D3 can have many indirect effects on sion in prostate cancer, colorectal adenoma and carci- cell-cycle regulation owing to cross-talk with other noma cells84. In addition to regulating the expression

pathways; for example, 1α,25(OH)2D3 treatment can of the BCL2 family, 1α,25(OH)2D3 might also directly result in the upregulation of IGFBP3 (which encodes activate caspase effector molecules, although it is

insulin growth factor binding protein 3) and trans- unclear whether 1α,25(OH)2D3-induced apoptosis is forming growth factor‑β (TGFβ)–SMAD3 signalling caspase-dependent84. In support of this idea, the treat-

cascades and by downregulating the epidermal growth ment of mouse SCC tumour cells with 1α,25(OH)2D3 factor receptor (EGFR) signalling pathway67,95,96 (FIG. 3). increased VDR expression and concomitantly inhib- Although there appears to be an overall inhibition of ited the phosphorylation of ERK104. Upstream of ERK, cell-cycle progression in tumour cells treated with the growth-promoting and pro-survival signalling

1α,25(OH)2D3, the precise molecular basis for such molecule MEK is cleaved and inactivated in a caspase- an effect differs from one tumour cell type to another dependent manner in cells that undergo apoptosis

such that a unifying hypothesis with regard to the after treatment with 1α,25(OH)2D3. Recently, a novel

exact mechanism of 1α,25(OH)2D3-mediated cell- mechanism of 1α,25(OH)2D3-mediated apoptosis in cycle perturbation has not been possible (supplemental epithelial ovarian cancer cells was proposed by Jiang 105 information S1 (table)). et al. , wherein they showed that 1α,25(OH)2D3

Activation of the VDR by 1α,25(OH)2D3 can also destabilizes telomerase reverse transcriptase (TERT) inhibit tumour cell proliferation by inducing differentia- mRNA, therefore inducing apoptosis through tel- tion in various myeloid leukaemia cell lines and freshly omere attrition resulting from the down-regulation isolated leukaemia cells62,83, which is dependent on the of telomerase activity. The diverse effects observed

formation of activated VDR and phosphatidylinositol 3- for 1α,25(OH)2D3-mediated apoptosis suggest that kinase (PI3K) complexes97. However, in haematopoeitic although anti-proliferative effects directed against

progenitor cells, 1α,25(OH)2D3 inhibits differentiation the tumour are clear in vitro and in vivo (supple- through VDR-independent suppression of interleukin mental information S1 (table)), dissecting the exact 12 (IL12) protein secretion and down-regulation of other mechanism(s) central to these activities remains a co-stimulatory molecules (CD40, CD80 and CD86)98. In challenge. cell lines of head and neck, colon and prostate tumours,

administration of 1α,25(OH)2D3 or vitamin D analogues Angiogenesis. 1α,25(OH)2D3 inhibits the proliferation induces the expression of genes that are associated with the of endothelial cells in vitro and reduces angiogenesis differentiated cell of origin91,99,100. In various colon cancer in vivo106–108. Vascular endothelial growth factor (VEGF)-

cells, treatment with 1α,25(OH)2D3 induces differentia- induced endothelial cell tube formation and tumour

tion either by increasing PKC- and JNK-dependent JUN growth are inhibited in vivo by 1α,25(OH)2D3 admin- activation101 or by differentially regulating the expression istration to mice with VEGF-overexpressing MCF‑7 86 of inhibitor of DNA binding 1 and 2 (ID1 and ID2), which xenografts . 1α,25(OH)2D3 can increase VEGF mRNA encode proteins that are transcriptional regulators of epi- levels in vascular smooth muscle cells109 and upregu- thelial cell proliferation (ID2) and differentiation (ID1); late mRNA levels of the potent anti-angiogenic factor the repression of ID2 mediated the antiproliferative effects thrombospondin 1 (THBS1) in SW480-ADH human

REVIEWS

102 colon tumour cells . In SCC cells, 1α,25(OH)2D3 prostaglandin activity, the induction of their degradation induces the angiogenic factor interleukin 8 (IL8)110, but through the upregulation of 15-hydroxyprostaglandin

in prostate cancer cells 1α,25(OH)2D3 interrupts IL8 dehydrogenase, and reduction of prostaglandin recep- signalling leading to the inhibition of endothelial cell tors120. These findings support the rationale for clinical 111 migration and tube formation . A significant inhibition evaluation of a combination of 1α,25(OH)2D3 and non- of metastasis is observed in prostate and lung murine steroidal anti-inflammatory drugs (NSAIDs) for prostate 120 models treated with 1α,25(OH)2D3, and these effects cancer therapy . Increased anti-tumour effects with

may be based, at least in part, on the anti-angiogenic 1α,25(OH)2D3 combination therapy offers the opportu- 79,82 effects described . Interestingly, in tumour-derived nity for the clinical use of 1α,25(OH)2D3 across several

Platinum analogues endothelial cells (TDECs), 1α,25(OH)2D3 induces apop- tumour types where modest effects are observed with Platinum-based tosis and cell-cycle arrest; however, these effects are not chemotherapy alone. chemotherapeutics that seen in endothelial cells isolated from normal tissues crosslink DNA and therefore or from Matrigel plugs (Matrigel-derived endothelial Clinical trials of 1α,25(OH)2D3 impair the progression of DNA 106 112 replication machinery. cells) . Recently, Chung et al. demonstrated that With the recognition of the preclinical antiproliferative TDECs may be more sensitive to 1α,25(OH)2D3 owing and pro-differentiating effects of vitamin D in the 1970s Taxanes to the epigenetic silencing of CYP24A1. Therefore, and 1980s, a number of attempts were made to trans- Drugs that inhibit microtubule direct effects of 1α,25(OH)2D3 on endothelial cells may late these findings into the clinic. Several investigators dynamics by stabilizing GDP- α α bound tubulin. Microtubules have a primary role in the 1 ,25(OH)2D3-mediated attempted to administer 1 ,25(OH)2D3 as a differenti- 121–123 form the mitotic spindle and so anti-tumour activity that is observed in animal models ating agent in myelodysplasia and acute leukaemia . taxanes prevent the of cancer. Although some patients seemed to respond to the ther- completional of mitosis. apy, these improvements were not enough to encour- Preclinical combination studies age further trials, as 20–30% of patients who received Myelodysplasia hypercalcemia Any of a group of bone marrow In vitro and in vivo analyses indicate that 1α,25(OH)2D3 a daily dose of 1α,25(OH)2D3 developed . disorders that have markedly acts synergistically with chemotherapeutic agents. Such findings have reinforced the conviction that less abnormal reduction in one or 1α,25(OH)2D3 potentiates the anticancer activity of hypercalcemic analogues of vitamin D, with modified more types of circulating blood agents such as platinum analogues113–115, taxanes116,117 and chemical structures to make them less prone to degra- cells owing to defective growth 117,118 (FIG. 4a) and maturation of blood- DNA-intercalating agents . Optimal potentiation dation by 24-OHase , must be developed if the forming cells in the bone is seen when 1α,25(OH)2D3 is administered before or therapeutic advantages of vitamin D biological effects are marrow. simultaneously with chemotherapy treatment; admin- to be realized124,125. It is important to note that the early

istration of 1α,25(OH)2D3 after the cytotoxic agent anticancer studies of 1α,25(OH)2D3 were conducted Hypercalcemia does not provide potentiation114,116. The combination of using dosing schedules optimized for the treatment Excess of Ca2+ in the blood. osteodystrophy osteoporosis Chronic elevated serum levels 1α,25(OH)2D3 and cisplatin in SCC cells in vitro induced of renal and , and the doses 2+ of Ca (12.0 mg dL) can result tumour cell apoptosis characteristic of 1α,25(OH)2D3 important for anticancer effects were not investigated in urinary calculi (renal or alone. The pro-apoptotic signalling molecule MEKK1 separately. Had the administration of 1α,25(OH)2D3 bladder stones) and abnormal (mitogen-activated protein kinase kinase kinase 1), is been developed from an anticancer standpoint, the heart rhythms. Severe hypercalcemia (above 15–16 up-regulated in both apoptotic and pre-apoptotic SCC following considerations would have been determined: (REF 104) mg dL) can result in coma and cells treated with 1α,25(OH)2D3 . This up-regu- first, optimal biologically-effective dose and maximum cardiac arrest. lation of MEKK1 was potentiated in combination with tolerated dose (MTD) across several cancers; second,

cisplatin treatment, suggesting that 1α,25(OH)2D3 pre- the most effective dosing schedules to achieve antican- Osteodystrophy treatment commits cells to undergo apoptosis through cer activity; third, 1α,25(OH) D -dependent signalling Defective bone ossification that 2 3 occurs when the kidney fails to specific molecular pathways (probably the MEK signal- targets and molecular end-points; fourth, 1α,25(OH)2D3 maintain proper levels of Pi and ling pathway), and that this effect is increased when cells interactions with other cytotoxic or other anticancer Ca2+. This results in slowed are treated with an additional genotoxic stimulus113. drugs that may be therapeutically advantageous; and bone growth and causes bone Similar effects are seen in MCF‑7 cells treated with the finally, design of clinical trials that mirror, as much as deformities in children. In adults, renal osteodystrophy vitamin D analogue ILX 23‑7553 in combination with possible, the exposures active in preclinical models to 118 results in thin and weak bones, doxorubicin or ionizing radiation . In these studies, ILX determine whether biological effects can be achieved in bone and joint pain and 23‑7553 increased doxorubicin cytotoxicity and blocked human tumours in clinical therapy (FIG. 4b). vulnerability to osteoporosis. the induction of p53 expression. Increased anti-tumour Several studies have attempted to define a safe activity with 1α,25(OH) D and the taxane paclitaxel is and effective clinical treatment regimen126–129. These Osteoporosis 2 3 A condition that is associated with a significant decrease in p21 expression, investigations were based on the recognition that most characterized by a decrease in which sensitizes cells to both DNA-damaging agents positive preclinical studies used high-dose, intermit- bone mass with decreased (such as cisplatin and doxorubicin) and microtubule- tent 1α,25(OH)2D3. Although it is clear that 20–30% density and enlargement of 116 disrupting agents (such as paclitaxel and docetaxel) . of patients receiving 1α,25(OH)2D3 at a dose of 1.5–2.0 bone spaces producing µ 130 porosity and brittleness of the In SCC and PC‑3 (prostate cancer) xenografts, pre- g a day develop hypercalcemia , there have been few bone. treatment with 1α,25(OH)2D3 resulted in an increased studies that have compared continuous and intermittent anti-tumour effect in combination with paclitaxel116. dosing regimens in cancer patients. Muindi and col- Pharmacokinetics Similar results have also been observed in vivo with leagues131 have determined the pharmacokinetic profile of The characteristic interactions MCF‑7 xenografts in which mice were treated with a 1α,25(OH)2D3 regimen that is active in a preclinical of a drug and the body in 119 terms of its absorption, vitamin D analogues and paclitaxel . 1α,25(OH)2D3- animal model. High-dose 1α,25(OH)2D3 (daily for 3 distribution, metabolism and mediated downregulation of cyclooxygenase 2 (COX2) days 0.125 µg per mouse ~6.25 µg per kg (body weight)) excretion. expression in prostate cancer cells leads to decreased resulted in growth inhibition of the syngeneic mouse SCC

REVIEWS

a 21 Side chain 20 22 18 24 12 17 23 11 13 25 OH 16 CD14 OH 9 8 15 7 6 Seco-B-ring 5 19 4 10 A Vitamin D3 25(OH)D3 1α,25(OH)2D3 HO 3 1 Cholecalciferol 25-Hydroxycholecalciferol Calcitriol HO 2 HO OH Vitamin D analogues Vitamin D receptor modulators

OH S OH HO O S O LY2108491 O

Paricalcitol EB1089 HO OH S HO OH HO OH O O LY2109866 O OH OH

O O HO OH ILX23-7553 O LG190119 O HO OH OCT b 1 ,25(OH)2D3 or new analogues Epidemiology/risk factors Prediction of • Nutritional composition and drug combinations 1 ,25(OH) D response • UV exposure α 2 3 • VDR polymorphisms In vitro systems • Genetic background • Tumour cells Mechanisms of action • Stromal cells • Genomic (VDR) Vitamin D levels • Progenitor cancer stem cells • Non-genomic (memVDR) • Apoptosis (serum and cancer tissue) 1α,25(OH)2D3 1α,25(OH)2D3 or new associated toxicities: • Cell cycle • Effectors of 1α,25(OH) D 2 3 analogues Improve drug dosing • Angiogenesis metabolism: expression In vivo systems • Cell signaling cross-talk and activity of 25-OHase, and drug Pharmacokinetics combinations Pharmacodynamics • Preclinical animal models: • Cell–cell interaction 1α-OHase, 24-OHase, VDR syngeneic, xenografts and genetically-modified (VDR–/–) Clinical assessment Cancer • Therapeutic impact, Clinical trials patient response and biomarkers • Prevention Clinical dosing evaluation • Anti-tumour therapy schedules

Figure 4 | Development of 1 ,25(OH) D and vitamin D analogues as anticancer drugs. a | Cholecalciferol (vitamin α 2 3 Nature Reviews | Cancer D3) is 25-hydroxylated at C‑25 (denoted by carbon atom number on the structure of cholecalciferol) to form 25-hydroxyc-

holecalciferol (25(OH)D3). This is 1α-hydroxylated at C‑1 by 1α-OHase to yield 1α,25(OH)2D3 (calcitriol). 1α,25(OH)2D3 is a secosteroid that is similar in structure to steroids but with a ‘broken’ B‑ring (denoted seco‑B-ring) where two of the carbon atoms (C‑9 and C‑10) of the four steroid rings are not joined. Many vitamin D analogues (left) retain the secosteroid structure with modified side chain structures around the C‑24 position, which makes them less hypercalcemic and less prone to degradation by 24-OHase170,171. Several structures of vitamin D analogues referred to in the text are shown:

paricalcitol (19-nor‑1α(OH)2D2), ILX23‑7553 (16-ene‑23-yne‑1α,25(OH)2D3), OCT (Maxacalcitol, 22-oxa‑1α,25(OH)2D3)

and EB1089 (Seocalcitol, 1α-dihydroxy‑22,24-diene‑24,26,27-trihomo-vitamin D3). Vitamin D receptor modulators (VDRMs, right) are non-secosteroidal in structure. Some of the representative compounds described are LY2108491,

LY2109866 and LG190119 (REFs 146,147). b | Paradigm for development and clinical translation of 1α,25(OH)2D3 as an

anticancer agent. Establishment of in vitro and in vivo experimental systems is crucial to developing 1α,25(OH)2D3 or vitamin D analogues that target vitamin D metabolism and signalling. These systems allow the mechanisms of action of

1α,25(OH)2D3 to be studied along with novel analogues (also in combination with cytotoxic drugs) in multiple transformed cell types and their biological effects (tumour and normal tissues) in animals. Importantly, studies on the pharmacokinetics and pharmacodynamics of drug action will enable the development of better designed clinical dosing schedules for

clinical trials that will mirror the exposures active in preclinical models where optimal biological effects of 1α,25(OH)2D3 are demonstrated and are achievable in human tumours in clinical therapy.

REVIEWS

have been completed (supplemental information S2 Box 2 | The pharmacology of 1α,25(OH)2D3 (table)). Most of these studies are based on persuasive In developing an agent for human use, it is crucial to understand the preclinical data, but confront a problem not usually pharmacokinetics and, when possible, the pharmacodynamic parameters encountered in single-agent or combination phase I associated with drug administration. Although the pharmacodynamics of studies in cancer: dose, toxicity and pharmacokinetic 1α,25(OH) D have not been well studied, the pharmacokinetics of 1α,25(OH) D 2 3 2 3 α have been extensively investigated using a standard, commercial formulation of data for 1 ,25(OH)2D3 as a single agent are limited. Although investigators working with the Novacea 1α,25(OH)2D3 (Rocaltrol, Hoffman–LaRoche). Importantly, these pharmacokinetic studies revealed that dose escalation did not result in the escalation of systemic formulation of 1α,25(OH)2D3 (DN‑101) have defined exposure. Both groups found that the desirable linear relationship between dose suitable pharmacokinetics and have shown the feasibil-

administered and systemic exposure (area under the curve (AUC) and Cmax) was ity of very high doses of 1α,25(OH)2D3, an aggressive lost at doses >16 µg127,129,131. In addition, there was marked variation (5–10×) in MTD in cancer patients has not been determined. 139 AUC and Cmax among patients receiving the same dose of 1α,25(OH)2D3. Two Trump and colleagues completed a 43-patient study approaches indicate that these pharmacokinetic findings are a product of the of 1α,25(OH)2D3 in escalating oral doses to a maximum pharmaceutical characteristics of Rocaltrol when administered at high dose of 12 µg 1α,25(OH) D three times a week together rather than reduced absorption or increased catabolism of 1α,25(OH) D within 2 3 2 3 with dexamethasone. Minimal hypercalcemia was the patients studied. Novocea, Inc.168 has developed a new formulation of observed and high-dose intermittent 1α,25(OH) D 1α,25(OH) D (DN‑101, Ascentar) for high-dose applications (15 µg and 45 µg 2 3 2 3 139 caplets) and pharmacokinetic studies indicate a linear relationship between dose plus dexamethasone was safe and feasible . At present, and exposure at oral doses up to 168 µg, indicating that there is no ‘barrier’ to dexamethasone in combination with weekly intra- 168 α gastrointestinal absorption of 1α,25(OH)2D3 at the doses studied . venous 1 ,25(OH)2D3 plus gefitinib is being explored

High-dose intravenous 1α,25(OH)2D3 (Calcijex, Abbott Pharmaceuticals) has been in a phase I clinical trial to determine whether an aggres- investigated in a phase I clinical trial136 (supplemental information S2 (table)). A linear sive MTD can be achieved (supplemental information relationship between dose and exposure was observed across a wide dose range S2 (table)). Glucocorticoids are used clinically to amelio- (10–125 g), indicating that 1 ,25(OH) D administration is not associated with rapid µ α 2 3 rate hypercalcemia in a number of situations including induction of 1α,25(OH)2D3 catabolism. The same patients monitored on multiple 140 1α,25(OH)2D3 intoxication . Dexamethasone also sig- occasions had no convincing evidence that the pharmacokinetics of 1α,25(OH)2D3 on nificantly improves 1α,25(OH)2D3 anti-tumour efficacy, day 1 of a once a day for 3 days a week schedule are different from the 141 pharmacokinetics on day 28; neither does the administration of either paclitaxel or in vitro and in vivo, through direct effects on the VDR . In studies of tumour-bearing animals, dexamethasone dexamethasone modify 1α,25(OH)2D3 pharmacokinetics. Although formal increases VDR receptor number without changing the bioavailability studies of 1α,25(OH)2D3 have not been done, inspection of

pharmacokinetic curves in our studies of intravenous 1α,25(OH)2D3 and those of the ligand affinity (Kd) in SCC tumour tissue xenograft DN‑101 study suggest that oral absorption of a suitable formulation is very efficient and the kidney, but not in gastrointestinal mucosa141. (80–90%) even at a high dose136,169. The ability of dexamethasone to increase anti-tumour activity in certain tissues and decrease toxicity is mediated through the modulation of VDR expression. The area under the curve (AUC) cells. At these doses, the 0–24 h combined use of a glucocorticoid and 1α,25(OH)2D3 is C (37 ± 2.5 ng•hr ml) and max (22 nM) of 1α,25(OH)2D3 a viable approach to reducing side-effects experienced Area under the curve α (AUC). In pharmacokinetics, the were within the concentration range that does not by patients treated with 1 ,25(OH)2D3. 131 area under the curve is a plot cause toxicity in patients . Although SCC is a sensitive Another approach to reducing potential toxicity and of concentration of drug in model, concentrations and exposure to 1α,25(OH)2D3 increasing anti-tumour activity is the development of serum over time that that inhibit SCC tumour growth are also active in many vitamin D analogues and vitamin D receptor modu- represents the measure of an human tumour xenograft models41,78,80–82,132,133. lators (VDRMs) (FIG. 4a) that are less prone to cause individual’s exposure to the drug. In developing an anticancer agent, more aggressive hypercalcemia. However, considerable data indicate management of toxicity and use of supportive care that when 1α,25(OH)2D3 is given at an intermittent Bioavailability approaches often allow one to overcome mild to mod- schedule, clinical use is not limited by hypercalcemia Measurement of an erate side effects. Beer and colleagues129 conducted a or hypercalciuria126–129,137,138,142. Vitamin D analogues have administered dose of a α 143 therapeutically active drug that standard phase I dose-escalation trial of 1 ,25(OH)2D3 been synthesized and their properties examined . reaches the systemic administered orally once a week, and found that 2.8 µg Although many appear to be less hypercalcemogenic circulation and depends on the per kg (body weight) can be safely administered with- than 1α,25(OH)2D3, the complexities of in vitro–in vivo mode of administration. out any side effects. Dose escalation was not continued data and dose extrapolation limit the conclusions that because at doses higher than 2.4 µg per kg (body weight) analogues which cause less hypercalcemia are equipo- Cmax Maximum or ‘peak’ oral absorption was found to be incomplete and unreli- tent in terms of anticancer effects. For example, most concentration of a drug able (BOX 2). Several phase I trials have been conducted in analogues that cause less hypercalcemia bind less tightly observed after its 127,134,135 which an MTD of 1α,25(OH)2D3 has been sought . to the VDR, a property that probably reduces their administration. 143 Dose-limiting toxicity of oral 1α,25(OH)2D3 has not anti-tumour effects . As non-steroidal tissue-selective Glucocorticoids been observed in these studies. oestrogen-receptor (ER) modulators (SERMs) such Corticosteroids are involved in As discussed above, combinations of 1α,25(OH)2D3 as tamoxifen have proven clinically successful for the carbohydrate, protein and fat with other anticancer agents demonstrate synergistic prevention of breast cancer in high-risk women, and metabolism to regulate liver 144 interactions. Phase I studies of 1α,25(OH)2D3 plus for the treatment of ER‑positive breast cancer , recent glycogen and blood sugar by 127 136 paclitaxel and 1α,25(OH)2D3 plus gefitinib for the development of non-secosteroidal VDRMs have shown increasing gluconeogenesis; 145 clinically used for anti- treatment of advanced malignancies, and phase II stud- potential in anticancer therapy . The novel non-secos- (FIG. 4a) inflammatory and ies of 1α,25(OH)2D3 plus carboplatin and 1α,25(OH)2D3 teroidal VDRMs LY2108491 and LY2109866 immunosuppressive effects. plus docetaxel for the treatment of prostate cancer137,138 were identified as potent tissue-selective agonists in

REVIEWS

keratinocytes, human peripheral blood mononuclear mon, at least in some parts of the US and Europe, and

cells and osteoblasts, but have weak potency in intestinal that inadequate levels of 25(OH)D3 are associated with cells146. Furthermore, the non-secosteroidal, tissue-selec- an increased risk and poor prognosis of several types tive VDRMs were less calcemic in vivo compared with of cancer16. In view of the numerous other potential

1α,25(OH)2D3, and show efficacy in an animal model of consequences of vitamin D deficiency to human health, psoriasis146; however, their potential in anticancer therapy such as rickets and osteomalacia, one could easily rec-

has not been determined. Polek and collegues reported ommend more aggressive monitoring of 25(OH)D3 that a novel VDRM, LG190119 (FIG. 4a), inhibited LNCaP levels as part of a health maintenance programme. xenograft tumour growth without hypercalcemia147. Non- Meta-analysis and cancer-prevention trials indicate

secosteroidal VDRMs represent promising therapeutic that vitamin D3 supplementation to achieve a level of

agents for the treatment of cancers with tissue-selectivity >82 nmol per L 25(OH)D3 can lower the incidence of and potential evasion of hypercalcemia. colorectal cancer by 50%20. To achieve serum levels in Clinical studies of vitamin D analogues have this range, individuals require a daily 4,000 IU (inter- (REF. 20) focused primarily on continuous daily administration national unit) supplement of vitamin D3 , which of EB1089 (FIG. 4a; seocalcitol, 1α-dihydroxy‑22,24- is achievable with the current formulations that range

diene‑24,26,27-trihomovitamin D3) to patients with from 200–2,000 IU and a liquid formulation of 2,000 breast cancer, colorectal cancer or hepatocellular IU per drop. Formal randomized studies to optimize 148–150 carcinoma . EB1089 failed to show evidence of replacement strategies and to evaluate vitamin D3 as a anti-tumour activity in these studies, and potentially cancer-preventative approach should be considered. problematic hypercalcemia was seen but was not dose Changes in the expression of proteins important FIG. 4a limiting. Paricalcitol ( ; 19-nor‑1α,25-(OH)2D2, in vitamin D synthesis and catabolism (25-OHase, Zemplar), an analogue developed by Abbott, appears 1α-OHase, 24-OHase) and those crucial for mediat-

to be more effective than 1α,25(OH)2D3 in the man- ing the biological effects of 1α,25(OH)2D3 (VDR) have agement of renal osteodystrophy and chronic renal been shown to be associated with poor differentiation disease151, and preclinical data indicate that paricalcitol status and prognosis of several types of cancer, such as has anti-tumour effects in prostate, pancreas, lung and colon cancer31–33,35. The overexpression of vitamin D breast cancers, as well as multiple myeloma152. Current catabolic enzymes in cancer suggests that low cellular

phase I clinical trials have been initiated for paricalcitol 1α,25(OH)2D3 is also associated with poor prognosis, plus gemcitabine and paricalcitol plus zoledronic acid but this has not yet been addressed convincingly. In

(a bisphosphonate) in patients with advanced solid addition, the steady-state level of cellular 1α,25(OH)2D3 tumours and multiple myeloma, respectively, to estab- in tumour tissue is difficult to measure. Assessment

lish whether very high doses of these analogues can be of the catabolic enzyme that degrades 1α,25(OH)2D3, safely administered intravenously when an intermit- such as 24-OHase (encoded by CYP24A1), may have tent schedule is used (supplemental information S2 merit in the development of prognostic models. (table)). Indeed, CYP24A1 is overexpressed in many cancers In patients with prostate cancer that progresses (TABLE 3). Of interest, epigenetic silencing of CYP24A1 despite castration (so-called androgen-independent in tumour-derived endothelial cells renders the tumour

prostate cancer or AIPC), 1α,25(OH)2D3 has been stud- sensitive to the anti-angiogenic effects of 1α,25(OH)2D3 ied in ‘standard’ dose and schedule and at a high dose. (REF. 112). Various molecules can inhibit 24-OHase 1α,25(OH) D intoxication 2 3 The most striking indication of 1α,25(OH) D anti- (such as azoles and vitamin D analogues). These merit The symptoms of 2 3 hypervitaminosis D (excessive tumour effects in AIPC is the randomized trial reported exploration and further development as specific small- 138 doses of vitamin D) are a result by Beer et al. that used docetaxel (36 µg once a week) molecule 24-OHase inhibitors, especially in combi- of hypercalcemia caused by and 1α,25(OH)2D3 (DN‑101, 45 µg one day before nation with high-dose intermittent 1α,25(OH)2D3 increased intestinal Ca2+ docetaxel). The survival in the DN‑101 plus docetaxel- or other vitamin D analogues. These may maximize absorption. Gastrointestinal α symptoms include anorexia, treated patients was improved, but further confirmation intracellular 1 ,25(OH)2D3 content and exert optimal nausea and vomiting. is required because survival was not the primary end antiproliferative effects. point of this phase II study. Novacea is currently con- The growth restraining, differentiation and apop- Hypercalciuria ducting a 1,000 patient phase III trial to further evaluate tosis-inducing effects of 1α,25(OH) D in different 2+ 2 3 Excessive urinary Ca this survival difference. It is also striking that severe or tumour cell types is well documented (supplemental excretion. The morbidity thromboembolic associated with hypercalciuria life-threatening side effects, including information S1 (table)). Across various tumour cell 153 is related to kidney stone complications, were reduced in the DN‑101 arm . The lines, different molecular markers of cell cycle, differ- disease and bone results of this trial point to potentially clinically relevant entiation and apoptosis can be observed with no clear demineralization leading to anti-tumour effects of 1α,25(OH) D in combination pattern of modulation by 1α,25(OH) D ; perhaps these osteopaenia (decrease in bone 2 3 2 3 density) and osteoporosis. with docetaxel. studies demonstrate the importance of heterogeneity to the 1α,25(OH)2D3 response, even among similar Thromboembolic Conclusions and future perspectives tumour cell types. The antiproliferative actions of complications The data described above support the continued 1α,25(OH)2D3 may depend on the differentiation status Associated with blockage of a exploration of vitamin D supplementation and of the tumour cells and VDR expression level, as well blood vessel by a particle that has dislodged from a blood 1α,25(OH)2D3 as approaches to cancer prevention as genomic or post-translational modifications of co- clot at its primary formation and treatment, respectively. The epidemiological data activator proteins that are essential for the assembly of site. indicate that vitamin D deficiency is relatively com- the transcriptionally active VDR complex154.

REVIEWS

1α,25(OH)2D3 has prominent antiproliferative, 1α,25(OH)2D3 is required so that more extensive tri-

anti-angiogenic and pro-differentiative effects in a als of 1α,25(OH)2D3, alone and in combination with broad range of cancers (supplemental information S1 cytotoxic and other anticancer agents, can be con- (table)). These effects are mediated through perturba- ducted. Initial phase II studies and a single phase III

tion of several important signalling pathways medi- study suggest that 1α,25(OH)2D3 is an active agent ated through genomic and non-genomic mechanisms. in cancer therapy for prostate cancer (supplemental

1α,25(OH)2D3 potentiates the anti-tumour effects of information S2 (table)). The ease of administration,

many anticancer therapeutic compounds, and sev- broad biological effects of 1α,25(OH)2D3 and the eral clinical trials indicate that the administration of potentiation of several anticancer drugs strongly sup-

high-dose 1α,25(OH)2D3 and vitamin D analogues port the continued development of 1α,25(OH)2D3 as is safe and feasible. Clear delineation of the MTD of an anticancer drug.

1. Holick, M. F. Vitamin D and bone health. J. Nutr. 126, 25(OH)D3 level and risk factor in tumour incidence An important study that shows the importance of

1159S–1164S (1996). and mortality in men. 1α,25(OH)2D3 and its regulated synthesis in colon 2. Dardenne, O., Prud’homme, J., Arabian, A., 17. Cui, Y. & Rohan, T. E. Vitamin D, calcium, and breast mucosa; expression of 1α-OHase is lost in Glorieux, F. H. & St-Arnaud, R. Targeted inactivation of cancer risk: a review. Cancer Epidemiol. Biomarkers undifferentiated, malignant human colon cancer. the 25-hydroxyvitamin D(3)‑1α-hydroxylase gene Prev. 15, 1427–1437 (2006). 34. Friedrich, M. et al. Analysis of the vitamin D system in (CYP27B1) creates an animal model of pseudovitamin 18. Schwartz, G. G. & Skinner, H. G. Vitamin D status and cervical carcinomas, breast cancer and ovarian cancer. D‑deficiency rickets. Endocrinology 142, 3135–3141 cancer: new insights. Curr. Opin. Clin. Nutr. Metab. Recent Results Cancer Res. 164, 239–246 (2003). (2001). Care 10, 6–11 (2007). 35. Cross, H. S., Bises, G., Lechner, D., Manhardt, T. & 3. Panda, D. K. et al. Targeted ablation of the 25- 19. Schwartz, G. G. Vitamin D and the epidemiology of Kallay, E. The Vitamin D endocrine system of the gut— hydroxyvitamin D 1α-hydroxylase enzyme: evidence for prostate cancer. Semin. Dial. 18, 276–289 (2005). its possible role in colorectal cancer prevention. skeletal, reproductive, and immune dysfunction. Proc. 20. Gorham, E. D. et al. Vitamin D and prevention of J. Steroid Biochem. Mol. Biol. 97, 121–128 (2005). Natl Acad. Sci. USA 98, 7498–7503 (2001). colorectal cancer. J. Steroid Biochem. Mol. Biol. 97, 36. Albertson, D. G. et al. Quantitative mapping of amplicon 4. Erben, R. G. et al. Deletion of deoxyribonucleic acid 179–194 (2005). structure by array CGH identifies CYP24 as a candidate binding domain of the vitamin D receptor abrogates 21. Zhou, W. et al. Circulating 25-hydroxyvitamin d levels oncogene. Nature Genet. 25, 144–146 (2000). genomic and nongenomic functions of vitamin D. Mol. predict survival in early-stage non‑small‑cell lung Pioneering work which identified the vitamin D Endocrinol. 16, 1524–1537 (2002). cancer patients. J. Clin. Oncol. 25, 479–485 (2007). catabolic enzyme, CYP24, as a potential oncogene. 5. Yoshizawa, T. et al. Mice lacking the vitamin D receptor 22. Hollis, B. W. Circulating 25-hydroxyvitamin D levels Subsequent studies show increased expression of exhibit impaired bone formation, uterine hypoplasia indicative of vitamin D sufficiency: implications for CYP24 in cancers, giving the tumour increased

and growth retardation after weaning. Nature Genet. establishing a new effective dietary intake ability to degrade 1α,25(OH)2D3, which therefore 16, 391–396 (1997). recommendation for vitamin D. J. Nutr. 135, 317–322 prevents its growth inhibitory and differentiation 6. Li, Y. C. et al. Targeted ablation of the vitamin D (2005). effects. receptor: an animal model of vitamin D‑dependent 23. Haussler, M. R. et al. The nuclear vitamin D receptor: 37. Weiss, M. M. et al. Determination of amplicon rickets type II with alopecia. Proc. Natl Acad. Sci. USA biological and molecular regulatory properties boundaries at 20q13.2 in tissue samples of human 94, 9831–9835 (1997). revealed. J. Bone Miner. Res. 13, 325–349 (1998). gastric adenocarcinomas by high-resolution microarray 7. Kallay, E. et al. Characterization of a vitamin D receptor A thorough review of vitamin D metabolism, VDR comparative genomic hybridization. J. Pathol. 200, knockout mouse as a model of colorectal biology and structure and its interaction with co- 320–326 (2003). hyperproliferation and DNA damage. Carcinogenesis regulators and its transcriptional regulation of 38. Anderson, M. G., Nakane, M., Ruan, X., Kroeger, P. E. 22, 1429–1435 (2001). target genes. & Wu-Wong, J. R. Expression of VDR and CYP24A1 8. Zinser, G. M., Suckow, M. & Welsh, J. Vitamin D 24. Takeyama, K. et al. 25-Hydroxyvitamin D3 1α- mRNA in human tumors. Cancer Chemother. receptor (VDR) ablation alters carcinogen-induced hydroxylase and vitamin D synthesis. Science 277, Pharmacol. 57, 234–240 (2006). tumorigenesis in mammary gland, epidermis and 1827–1830 (1997). 39. Zhao, J., Tan, B. K., Marcelis, S., Verstuyf, A. & lymphoid tissues. J. Steroid Biochem. Mol. Biol. 97, 25. Brenza, H. L. & DeLuca, H. F. Regulation of 25- Bouillon, R. Enhancement of antiproliferative activity of 153–164 (2005). hydroxyvitamin D3 1α-hydroxylase gene expression by 1α, 25-dihydroxyvitamin D3 (analogs) by cytochrome This paper shows the importance of VDR in cancer parathyroid hormone and 1, 25-dihydroxyvitamin D3. P450 enzyme inhibitors is compound- and cell-type and that optimal VDR signalling may be required to Arch. Biochem. Biophys. 381, 143–152 (2000). specific. J. Steroid Biochem. Mol. Biol. 57, 197–202 suppress tumorigenesis. 26. Hewison, M., Zehnder, D., Bland, R. & Stewart, P. M. (1996). 9. Zinser, G. M., Sundberg, J. P. & Welsh, J. Vitamin D(3) 1α-Hydroxylase and the action of vitamin D. J. Mol. 40. Ly, L. H., Zhao, X. Y., Holloway, L. & Feldman, D. receptor ablation sensitizes skin to chemically induced Endocrinol. 25, 141–148 (2000). Liarozole acts synergistically with 1α, 25- tumorigenesis. Carcinogenesis 23, 2103–2109 (2002). 27. Murayama, A. et al. The promoter of the human 25- dihydroxyvitamin D3 to inhibit growth of DU 145 10. Zinser, G. M. & Welsh, J. Vitamin D receptor status hydroxyvitamin D3 1 α-hydroxylase gene confers human prostate cancer cells by blocking 24- alters mammary gland morphology and tumorigenesis positive and negative responsiveness to PTH, hydroxylase activity. Endocrinology 140, 2071–2076 in MMTV-neu mice. Carcinogenesis 25, 2361–2372 calcitonin, and 1 α, 25(OH)2D3. Biochem. Biophys. (1999). (2004). Res. Commun. 249, 11–16 (1998). 41. Peehl, D. M., Seto, E., Hsu, J. Y. & Feldman, D. 11. St-Arnaud, R. et al. Deficient mineralization of 28. Zehnder, D. et al. Extrarenal expression of 25- Preclinical activity of ketoconazole in combination with intramembranous bone in vitamin D‑24‑hydroxylase- hydroxyvitamin d(3)‑1 α-hydroxylase. J. Clin. calcitriol or the vitamin D analogue EB 1089 in ablated mice is due to elevated 1, 25-dihydroxyvitamin Endocrinol. Metab. 86, 888–894 (2001). prostate cancer cells. J. Urol. 168, 1583–1588 D and not to the absence of 24, 25-dihydroxyvitamin 29. Townsend, K. et al. Autocrine metabolism of vitamin D (2002). D. Endocrinology 141, 2658–2666 (2000). in normal and malignant breast tissue. Clin. Cancer 42. Parise, R. A. et al. CYP24, the enzyme that catabolizes 12. Garland, C. F. & Garland, F. C. Do sunlight and Res. 11, 3579–3586 (2005). the antiproliferative agent vitamin D, is increased in vitamin D reduce the likelihood of colon cancer? Int. J. 30. Schwartz, G. G., Whitlatch, L. W., Chen, T. C., lung cancer. Int. J. Cancer 119, 1819–1828 (2006). Epidemiol. 9, 227–231 (1980). Lokeshwar, B. L. & Holick, M. F. Human prostate cells 43. Evans, R. M. The steroid and thyroid hormone receptor A seminal finding that higher mortality rates from synthesize 1, 25-dihydroxyvitamin D3 from 25- superfamily. Science 240, 889–895 (1988). colon cancer in the northeast and lower rates in the hydroxyvitamin D3. Cancer Epidemiol. Biomarkers 44. Crofts, L. A., Hancock, M. S., Morrison, N. A. & south, southwest and west US led to the proposed Prev. 7, 391–395 (1998). Eisman, J. A. Multiple promoters direct the tissue- concept that vitamin D can reduce the risk of 31. Cross, H. S. et al. 25-Hydroxyvitamin specific expression of novel N‑terminal variant human colorectal cancer. D(3)‑1α‑hydroxylase and vitamin D receptor gene vitamin D receptor gene transcripts. Proc. Natl Acad. 13. Garland, C. F. et al. Serum 25-hydroxyvitamin D and expression in human colonic mucosa is elevated during Sci. USA 95, 10529–10534 (1998). colon cancer: eight-year prospective study. Lancet 2, early cancerogenesis. Steroids 66, 287–292 (2001). 45. Sunn, K. L., Cock, T. A., Crofts, L. A., Eisman, J. A. & 1176–1178 (1989). Early finding that the expression of CYP27B1 (1α- Gardiner, E. M. Novel N‑terminal variant of human 14. Bertone-Johnson, E. R. et al. Plasma 25- OHase) and VDR are increased in early colon VDR. Mol. Endocrinol. 15, 1599–1609 (2001). hydroxyvitamin D and 1, 25-dihydroxyvitamin D and tumorigenesis; supports the notion for colorectal 46. Carlberg, C. et al. Two nuclear signalling pathways for

risk of breast cancer. Cancer Epidemiol. Biomarkers cancer chemoprevention with 25(OH)D3. vitamin D. Nature 361, 657–660 (1993). Prev. 14, 1991–1997 (2005). 32. Bareis, P., Bises, G., Bischof, M. G., Cross, H. S. & 47. Kurokawa, R. et al. Differential orientations of the 15. Ahonen, M. H., Tenkanen, L., Teppo, L., Hakama, M. & Peterlik, M. 25‑hydroxy‑vitamin d metabolism in DNA-binding domain and carboxy-terminal Tuohimaa, P. Prostate cancer risk and prediagnostic human colon cancer cells during tumor progression. dimerization interface regulate binding site selection serum 25-hydroxyvitamin D levels (Finland). Cancer Biochem. Biophys. Res. Commun. 285, 1012–1017 by nuclear receptor heterodimers. Genes Dev. 7, Causes Control 11, 847–852 (2000). (2001). 1423–1435 (1993). 16. Giovannucci, E. et al. Prospective study of predictors of 33. Bises, G. et al. 25-hydroxyvitamin 48. Tagami, T., Lutz, W. H., Kumar, R. & Jameson, J. L. The vitamin D status and cancer incidence and mortality in D3‑1α‑hydroxylase expression in normal and interaction of the vitamin D receptor with nuclear men. J. Natl Cancer Inst. 98, 451–459 (2006). malignant human colon. J. Histochem. Cytochem. receptor corepressors and coactivators. Biochem. A paper presenting a predictive model for serum 52, 985–989 (2004). Biophys. Res. Commun. 253, 358–363 (1998).

REVIEWS

49. Li, B., Carey, M. & Workman, J. L. The role of One of the first papers to demonstrate that VDR 88. Verlinden, L. et al. Action of 1α, 25(OH)2D3 on the cell chromatin during transcription. Cell 128, 707–719 signalling cross-talks with other signalling pathways cycle genes, cyclin D1, p21 and p27 in MCF‑7 cells. (2007). (TGFβ signalling). Mol. Cell Endocrinol. 142, 57–65 (1998). 50. Khanim, F. L. et al. Altered SMRT levels disrupt vitamin 68. Losel, R. & Wehling, M. Nongenomic actions of steroid 89. Jensen, S. S., Madsen, M. W., Lukas, J., Binderup, L. & D3 receptor signalling in prostate cancer cells. hormones. Nature Rev. Mol. Cell Biol. 4, 46–56 (2003). Bartek, J. Inhibitory effects of 1α, 25-dihydroxyvitamin

Oncogene 23, 6712–6725 (2004). A good review on the effects of 1α,25(OH)2D3 on D(3) on the G(1)‑S phase-controlling machinery. Mol. 51. Banwell, C. M. et al. Altered nuclear receptor the nongenomic signalling pathways. Endocrinol. 15, 1370–1380 (2001). corepressor expression attenuates vitamin D receptor 69. Norman, A. W. Vitamin D receptor: new assignments 90. Hershberger, P. A. et al. 1α, 25-Dihydroxycholecalciferol signaling in breast cancer cells. Clin. Cancer Res. 12, for an already busy receptor. Endocrinology 147, (1, 25‑D3) inhibits the growth of squamous cell 2004–2013 (2006). 5542–5548 (2006). carcinoma and down-modulates p21(Waf1/Cip1) 52. Mackey, S. L., Heymont, J. L., Kronenberg, H. M. & This minireview provides a historical account of the in vitro and in vivo. Cancer Res. 59, 2644–2649 Demay, M. B. Vitamin D receptor binding to the role of VDR in genomic and nongenomic rapid (1999).

negative human parathyroid hormone vitamin D responses to 1α,25(OH)2D3. 91. Palmer, H. G. et al. Genetic signatures of differentiation response element does not require the retinoid x 70. Huhtakangas, J. A., Olivera, C. J., Bishop, J. E., induced by 1α, 25-dihydroxyvitamin D3 in human receptor. Mol. Endocrinol. 10, 298–305 (1996). Zanello, L. P. & Norman, A. W. The vitamin D receptor colon cancer cells. Cancer Res 63, 7799–7806 53. Demay, M. B., Kiernan, M. S., DeLuca, H. F. & is present in caveolae-enriched plasma membranes (2003). Kronenberg, H. M. Sequences in the human and binds 1 α, 25(OH)2-vitamin D3 in vivo and 92. Wang, Q. M., Jones, J. B. & Studzinski, G. P. Cyclin- parathyroid hormone gene that bind the 1, 25- in vitro. Mol. Endocrinol. 18, 2660–2671 (2004). dependent kinase inhibitor p27 as a mediator of the dihydroxyvitamin D3 receptor and mediate 71. Nemere, I. et al. Ribozyme knockdown functionally G1‑S phase block induced by 1α, 25-dihydroxyvitamin transcriptional repression in response to 1, 25- links a 1α, 25(OH)2D3 membrane binding protein (1, D3 in HL60 cells. Cancer Res. 56, 264–267 (1996). dihydroxyvitamin D3. Proc. Natl Acad. Sci. USA 89, 25D3-MARRS) and phosphate uptake in intestinal 93. Li, P. et al. p27(Kip1) stabilization and G(1) arrest by 1α, 8097–8101 (1992). cells. Proc. Natl Acad. Sci. USA 101, 7392–7397 25-dihydroxyvitamin D(3) in ovarian cancer cells 54. Murayama, A., Kim, M. S., Yanagisawa, J., Takeyama, K. (2004). mediated through down-regulation of cyclin E/cyclin- & Kato, S. Transrepression by a liganded nuclear 72. Nemere, I., Yoshimoto, Y. & Norman, A. W. Calcium dependent kinase 2 and Skp1‑Cullin‑F-box protein/Skp2 receptor via a bHLH activator through co-regulator transport in perfused duodena from normal chicks: ubiquitin ligase. J. Biol. Chem. 279, 25260–25267 switching. EMBO J. 23, 1598–1608 (2004). enhancement within fourteen minutes of exposure to (2004). The authors proposed a novel mechanism of 1α, 25-dihydroxyvitamin D3. Endocrinology 115, 94. Caligo, M. A., Cipollini, G., Petrini, M., Valentini, P. &

1α,25(OH)2D3-induced transrepression of 1α- 1476–1483 (1984). Bevilacqua, G. Down regulation of NM23.H1, NM23. OHase by VDR. VDR and/or RXR interaction with 73. Wali, R. K., Baum, C. L., Sitrin, M. D. & Brasitus, T. A. H2 and c‑myc genes during differentiation induced by the VDIR, bound to the E‑box-type negative VDRE 1α 25(OH)2 vitamin D3 stimulates membrane 1α, 25 dihydroxyvitamin D3. Leukemia Res. 20, in the promoter region of the CYP27B1 gene, phosphoinositide turnover, activates protein kinase C, 161–167 (1996). causes the dissociation of p300 co-activators from and increases cytosolic calcium in rat colonic 95. Huynh, H., Pollak, M. & Zhang, J. C. Regulation of VDIR and the association of HDAC co-repressor epithelium. J. Clin. Invest. 85, 1296–1303 (1990). insulin-like growth factor (IGF) II and IGF binding complex components which results in ligand- 74. Morelli, S., Buitrago, C., Boland, R. & de Boland, A. R. protein 3 autocrine loop in human PC‑3 prostate induced transrepression. The stimulation of MAP kinase by 1α, 25(OH)(2)- cancer cells by vitamin D metabolite 1α, 25(OH)2D3 55. Fujiki, R. et al. Ligand-induced transrepression by VDR vitamin D(3) in skeletal muscle cells is mediated by and its analog EB1089. Int. J. Oncol. 13, 137–143 through association of WSTF with acetylated histones. protein kinase C and calcium. Mol. Cell Endocrinol. (1998). EMBO J. 24, 3881–3894 (2005). 173, 41–52 (2001). 96. Tong, W. M., Hofer, H., Ellinger, A., Peterlik, M. & 56. Kim, M. S. et al. 1α, 25(OH)2D3-induced 75. Narayanan, R., Sepulveda, V. A., Falzon, M. & Cross, H. S. Mechanism of antimitogenic action of transrepression by vitamin D receptor through Weigel, N. L. The functional consequences of cross- vitamin D in human colon carcinoma cells: relevance E‑box‑type elements in the human parathyroid talk between the vitamin D receptor and ERK signaling for suppression of epidermal growth factor-stimulated hormone gene promoter. Mol. Endocrinol. 21, pathways are cell-specific. J. Biol. Chem. 279, cell growth. Oncol. Res. 11, 77–84 (1999). 334–342 (2007). 47298–47310 (2004). 97. Hmama, Z. et al. 1α, 25-dihydroxyvitamin D(3)- 57. Kim, M. S., Fujiki, R., Kitagawa, H. & Kato, S. 1α, 76. Hsieh, J. C. et al. Human vitamin D receptor is induced myeloid cell differentiation is regulated by a 25(OH)2D3-induced DNA methylation suppresses selectively phosphorylated by protein kinase C on vitamin D receptor-phosphatidylinositol 3‑kinase the human CYP27B1 gene. Mol. Cell Endocrinol. serine 51, a residue crucial to its trans-activation signaling complex. J. Exp. Med. 190, 1583–1594 265–266, 168–173 (2007). function. Proc. Natl Acad. Sci. USA 88, 9315–9319 (1999). 58. Khorchide, M., Lechner, D. & Cross, H. S. Epigenetic (1991). 98. Penna, G. & Adorini, L. 1α, 25-dihydroxyvitamin D3 regulation of vitamin D hydroxylase expression and 77. Banach-Petrosky, W. et al. Vitamin D inhibits the inhibits differentiation, maturation, activation, and activity in normal and malignant human prostate cells. formation of prostatic intraepithelial neoplasia in survival of dendritic cells leading to impaired J. Steroid Biochem. Mol. Biol. 93, 167–172 (2005). Nkx3.1;Pten mutant mice. Clin. Cancer Res. 12, alloreactive T cell activation. J. Immunol. 164, 59. Garcia-Bassets, I. et al. Histone methylation-dependent 5895–5901 (2006). 2405–2411 (2000). mechanisms impose ligand dependency for gene 78. McElwain, M. C. et al. Antiproliferative effects in vitro 99. Akutsu, N. et al. Regulation of gene Expression by 1α, activation by nuclear receptors. Cell 128, 505–518 and in vivo of 1α, 25-dihydroxyvitamin D3 and a 25-dihydroxyvitamin D3 and Its analog EB1089 under (2007). vitamin D3 analog in a squamous cell carcinoma model growth-inhibitory conditions in squamous carcinoma 60. Chen, K. S. & DeLuca, H. F. Cloning of the human 1 system. Mol. Cell Differ. 3, 31–50 (1995). cells. Mol. Endocrinol. 15, 1127–1139 (2001). alpha, 25-dihydroxyvitamin D‑3 24-hydroxylase gene 79. Getzenberg, R. H. et al. Vitamin D inhibition of 100. Guzey, M., Luo, J. & Getzenberg, R. H. Vitamin D3 promoter and identification of two vitamin prostate adenocarcinoma growth and metastasis in modulated gene expression patterns in human D‑responsive elements. Biochim. Biophysica Acta the Dunning rat prostate model system. Urology 50, primary normal and cancer prostate cells. J. Cell 1263, 1–9 (1995). 999–1006 (1997). Biochem. 93, 271–285 (2004). 61. Kerner, S. A., Scott, R. A. & Pike, J. W. Sequence 80. Zhang, X. et al. Growth suppression of ovarian cancer 101. Chen, A., Davis, B. H., Bissonnette, M., Scaglione- elements in the human osteocalcin gene confer basal xenografts in nude mice by vitamin D analogue Sewell, B. & Brasitus, T. A. 1α, 25-dihydroxyvitamin activation and inducible response to hormonal vitamin EB1089. Clin. Cancer Res. 11, 323–328 (2005). D(3) stimulates activator protein‑1‑dependent caco‑2 D3. Proc. Natl Acad. Sci. USA 86, 4455–4459 (1989). 81. Colston, K. W., Chander, S. K., Mackay, A. G. & cell differentiation. J. Biol. Chem. 274, 35505–35513 62. Liu, M., Lee, M. H., Cohen, M., Bommakanti, M. & Coombes, R. C. Effects of synthetic vitamin D (1999). Freedman, L. P. Transcriptional activation of the Cdk analogues on breast cancer cell proliferation in vivo 102. Fernandez-Garcia, N. I. et al. 1α, 25-dihydroxyvitamin inhibitor p21 by vitamin D3 leads to the induced and in vitro. Biochem. Pharmacol. 44, 693–702 D3 regulates the expression of Id1 and Id2 genes and differentiation of the myelomonocytic cell line U937. (1992). the angiogenic phenotype of human colon carcinoma Genes Dev. 10, 142–153 (1996). 82. Nakagawa, K., Kawaura, A., Kato, S., Takeda, E. & cells. Oncogene 24, 6533–6544 (2005). 63. Jiang, F., Li, P., Fornace, A. J. Jr., Nicosia, S. V. & Bai, Okano, T. 1 α, 25-Dihydroxyvitamin D(3) is a 103. Palmer, H. G. et al. Vitamin D(3) promotes the W. G2/M arrest by 1, 25-dihydroxyvitamin D3 in preventive factor in the metastasis of lung cancer. differentiation of colon carcinoma cells by the induction ovarian cancer cells mediated through the induction of Carcinogenesis 26, 429–440 (2005). of E‑cadherin and the inhibition of β-catenin signaling. GADD45 via an exonic enhancer. J. Biol. Chem. 278, 83. Wang, X. & Studzinski, G. P. Activation of extracellular J. Cell Biol. 154, 369–387 (2001). 48030–48040 (2003). signal-regulated kinases (ERKs) defines the first phase A prominent paper presenting evidence for the

64. Muto, A. et al. 1, 25-Dihydroxyvitamin D3 induces of 1α, 25-dihydroxyvitamin D3-induced differentiation involvement of 1α,25(OH)2D3 in suppressing differentiation of a retinoic acid-resistant acute of HL60 cells. J. Cell Biochem. 80, 471–482 (2001). β‑catenin activation, a dysregulated pathway in promyelocytic leukemia cell line (UF‑1) associated with 84. Ylikomi, T. et al. Antiproliferative action of vitamin D. colon cancer, and for the induction of E‑cadherin to expression of p21(WAF1/CIP1) and p27(KIP1). Blood Vitam. Horm. 64, 357–406 (2002). promote differentiation in colon carcinoma cells. 93, 2225–2233 (1999). 85. Shabahang, M. et al. 1α, 25-Dihydroxyvitamin D3 104. McGuire, T. F., Trump, D. L. & Johnson, C. S. Vitamin 65. Sinkkonen, L., Malinen, M., Saavalainen, K., Vaisanen, S. receptor as a marker of human colon carcinoma cell D(3)-induced apoptosis of murine squamous cell & Carlberg, C. Regulation of the human cyclin C gene line differentiation and growth inhibition. Cancer Res. carcinoma cells. Selective induction of caspase- via multiple vitamin D3-responsive regions in its 53, 3712–3718 (1993). dependent MEK cleavage and up-regulation of promoter. Nucleic Acids Res. 33, 2440–2451 (2005). 86. Mantell, D. J., Owens, P. E., Bundred, N. J., Mawer, MEKK‑1. J. Biol. Chem. 276, 26365–26373 (2001). 66. Tu-Yu, A. H., Morris, R. C. & Ives, H. E. Differential E. B. & Canfield, A. E. 1α, 25-dihydroxyvitamin D(3) 105. Jiang, F., Bao, J., Li, P., Nicosia, S. V. & Bai, W. modulation of fos and jun gene expression by 1, 25- inhibits angiogenesis in vitro and in vivo. Circ. Res. 87, Induction of ovarian cancer cell apoptosis by 1α, 25- dihydroxyvitamin D3. Biochem. Biophys. Res. 214–220 (2000). dihydroxyvitamin D3 through the down-regulation of Commun. 193, 161–166 (1993). 87. Simboli-Campbell, M., Narvaez, C. J., Tenniswood, M. telomerase. J. Biol. Chem. 279, 53213–53221 67. Yanagisawa, J. et al. Convergence of transforming & Welsh, J. 1α, 25-Dihydroxyvitamin D3 induces (2004). growth factor-β and vitamin D signaling pathways on morphological and biochemical markers of apoptosis 106. Chung, I. et al. Differential antiproliferative effects of SMAD transcriptional coactivators. Science 283, in MCF‑7 breast cancer cells. J. Steroid Biochem. Mol. calcitriol on tumor-derived and matrigel-derived 1317–1321 (1999). Biol. 58, 367–376 (1996). endothelial cells. Cancer Res. 66, 8565–8573 (2006).

REVIEWS

107. Merke, J. et al. Identification and regulation of 1α, 25- 127. Muindi, J. R. et al. Pharmacokinetics of high-dose oral 146. Ma, Y. et al. Identification and characterization of dihydroxyvitamin D3 receptor activity and biosynthesis calcitriol: results from a phase 1 trial of calcitriol and noncalcemic, tissue-selective, nonsecosteroidal vitamin of 1α, 25-dihydroxyvitamin D3. Studies in cultured paclitaxel. Clin. Pharmacol. Ther. 72, 648–659 D receptor modulators. J. Clin. Invest. 116, 892–904 bovine aortic endothelial cells and human dermal (2002). (2006). capillaries. J. Clin. Invest. 83, 1903–1915 (1989). 128. Smith, D. C. et al. A Phase I trial of calcitriol (1α, 25- 147. Polek, T. C. et al. Novel nonsecosteroidal vitamin D One of the first papers to show that endothelial cells dihydroxycholecalciferol) in patients with advanced receptor modulator inhibits the growth of LNCaP

are responsive to 1α,25(OH)2D3. This led to studies malignancy. Clin. Cancer Res. 5, 1339–1345 (1999). xenograft tumors in athymic mice without increased

that target tumour angiogenesis with 1α,25(OH)2D3. 129. Beer, T. M., Munar, M. & Henner, W. D. A Phase I trial serum calcium. Prostate 49, 224–233 (2001). 108. Iseki, K. et al. Inhibition of angiogenesis as a of pulse calcitriol in patients with refractory 148. Dalhoff, K. et al. A phase II study of the vitamin D mechanism for inhibition by 1α-hydroxyvitamin D3 and malignancies: pulse dosing permits substantial dose analogue Seocalcitol in patients with inoperable 1, 25-dihydroxyvitamin D3 of colon carcinogenesis escalation. Cancer 91, 2431–2439 (2001). hepatocellular carcinoma. Br. J. Cancer 89, 252–257 induced by azoxymethane in Wistar rats. Int J Cancer An important clinical finding that high-dose (2003).

81, 730–733 (1999). 1α,25(OH)2D3 is achievable in cancer patients. 149. Evans, T. R. et al. A phase II trial of the vitamin D 109. Cardus, A. et al. 1, 25-Dihydroxyvitamin D3 stimulates 130. Osborn, J. L. et al. Phase II trial of oral 1, 25- analogue Seocalcitol (EB1089) in patients with vascular smooth muscle cell proliferation through a dihydroxyvitamin D (calcitriol) in hormone refractory inoperable pancreatic cancer. Br. J. Cancer 86, VEGF-mediated pathway. Kidney Int 69, 1377–1384 prostate cancer. Urol. Oncol. 1, 195–198 (1995). 680–685 (2002). (2006). 131. Muindi, J. R., Modzelewski, R. A., Peng, Y., Trump, 150. Gulliford, T. et al. A phase I study of the vitamin D 110. Lin, R. et al. Expression profiling in squamous D. L. & Johnson, C. S. Pharmacokinetics of 1α, 25- analogue EB 1089 in patients with advanced breast carcinoma cells reveals pleiotropic effects of vitamin dihydroxyvitamin D3 in normal mice after systemic and colorectal cancer. Br. J. Cancer 78, 6–13 (1998). D3 analog EB1089 signaling on cell proliferation, exposure to effective and safe antitumor doses. 151. Teng, M. et al. Survival of patients undergoing differentiation, and immune system regulation. Mol. Oncology 66, 62–66 (2004). hemodialysis with paricalcitol or calcitriol therapy. Endocrinol. 16, 1243–1256 (2002). This paper establishes the use of preclinical animal N. Engl. J. Med. 349, 446–456 (2003).

111. Bao, B. Y., Yao, J. & Lee, Y. F. 1α, 25-dihydroxyvitamin model systems to determine the 1α,25(OH)2D3 152. Kumagai, T., O’Kelly, J., Said, J. W. & Koeffler, H. P. D3 suppresses interleukin‑8‑mediated prostate cancer dosing regimen that has a growth-inhibitory Vitamin D2 analog 19‑nor‑1, 25-dihydroxyvitamin D2: cell angiogenesis. Carcinogenesis 27, 1883–1893 biological response without toxicity, and shows that antitumor activity against leukemia, myeloma, and (2006). high dosing is achievable in patients without colon cancer cells. J. Natl Cancer Inst. 95, 896–905 112. Chung, I. et al. Epigenetic silencing of CYP24 in tumor- toxicity. (2003). derived endothelial cells contributes to selective growth 132. Schwartz, G. G. et al. Pancreatic cancer cells express 153. Beer, T. M. et al. High dose calcitriol may reduce inhibition by calcitriol. J. Biol. Chem. 282, 8704–8714 25-hydroxyvitamin D‑1 alpha-hydroxylase and their thrombosis in cancer patients. Br. J. Haematol. 135, (2007). proliferation is inhibited by the prohormone 25- 392–394 (2006). 113. Hershberger, P. A. et al. Cisplatin potentiates 1α, 25- hydroxyvitamin D3. Carcinogenesis 25, 1015–1026 154. Abedin, S. A., Banwell, C. M., Colston, K. W., dihydroxyvitamin D3-induced apoptosis in association (2004). Carlberg, C. & Campbell, M. J. Epigenetic corruption with increased mitogen-activated protein kinase kinase 133. Albert, D. M. et al. Responsiveness of human of VDR signalling in malignancy. Anticancer Res. 26, kinase 1 (MEKK‑1) expression. Mol. Cancer Ther. 1, retinoblastoma and neuroblastoma models to a non- 2557–2566 (2006). 821–829 (2002). calcemic 19-nor Vitamin D analog. J. Steroid Biochem. 155. Hsieh, J. C. et al. Novel nuclear localization signal 114. Light, B. W. et al. Potentiation of cisplatin antitumor Mol. Biol. 97, 165–172 (2005). between the two DNA-binding zinc fingers in the human activity using a vitamin D analogue in a murine 134. Beer, T. M. et al. Phase I study of weekly DN‑101, a vitamin D receptor. J. Cell Biochem. 70, 94–109 squamous cell carcinoma model system. Cancer Res. new formulation of calcitriol, in patients with cancer. (1998). 57, 3759–3764 (1997). Cancer Chemother. Pharmacol. 59, 581–587 (2007). 156. Michigami, T. et al. Identification of amino acid 115. Moffatt, K. A., Johannes, W. U. & Miller, G. J. 1α, 25 135. Beer, T. M. et al. Pharmacokinetics and tolerability of a sequence in the hinge region of human vitamin D dihydroxyvitamin D3 and platinum drugs act single dose of DN‑101, a new formulation of calcitriol, in receptor that transfers a cytosolic protein to the synergistically to inhibit the growth of prostate cancer patients with cancer. Clin. Cancer Res. 11, 7794–7799 nucleus. J. Biol. Chem. 274, 33531–33538 (1999). cell lines. Clin. Cancer Res. 5, 695–703 (1999). (2005). 157. Yasmin, R., Williams, R. M., Xu, M. & Noy, N. Nuclear 116. Hershberger, P. A. et al. Calcitriol (1α, 25- 136. Fakih, M. G. et al. A phase I pharmacokinetic and import of the retinoid X receptor, the vitamin D dihydroxycholecalciferol) enhances paclitaxel pharmacodynamic study of intravenous calcitriol in receptor, and their mutual heterodimer. J. Biol. Chem. antitumor activity in vitro and in vivo and accelerates combination with oral gefitinib in patients with 280, 40152–40160 (2005). paclitaxel-induced apoptosis. Clin. Cancer Res. 7, advanced solid tumors. Clin. Cancer Res. 13, 158. Jurutka, P. W. et al. Phosphorylation of serine 208 in 1043–1051 (2001). 1216–1223 (2007). the human vitamin D receptor. The predominant amino 117. Wang, Q., Yang, W., Uytingco, M. S., Christakos, S. & 137. Beer, T. M., Garzotto, M. & Katovic, N. M. High-dose acid phosphorylated by casein kinase II, in vitro, and Wieder, R. 1α, 25-Dihydroxyvitamin D3 and calcitriol and carboplatin in metastatic androgen- identification as a significant phosphorylation site in all‑trans‑retinoic acid sensitize breast cancer cells to independent prostate cancer. Am. J. Clin. Oncol. 27, intact cells. J. Biol. Chem. 268, 6791–6799 (1993). chemotherapy-induced cell death. Cancer Res. 60, 535–541 (2004). 159. Jurutka, P. W. et al. The polymorphic N terminus in 2040–2048 (2000). 138. Beer, T. M. et al. Double-blinded randomized study of human vitamin D receptor isoforms influences 118. Chaudhry, M., Sundaram, S., Gennings, C., Carter, H. high-dose calcitriol plus docetaxel compared with transcriptional activity by modulating interaction with & Gewirtz, D. A. The vitamin D3 analog, ILX‑23‑7553, placebo plus docetaxel in androgen-independent transcription factor IIB. Mol. Endocrinol. 14, 401–420 enhances the response to adriamycin and irradiation in prostate cancer: a report from the ASCENT (2000). MCF‑7 breast tumor cells. Cancer Chemother. Investigators. J. Clin. Oncol. 25, 669–674 (2007). 160. Han, J., Colditz, G. A. & Hunter, D. J. Polymorphisms Pharmacol. 47, 429–436 (2001). A recent phase II clinical study on the combination in the MTHFR and VDR genes and skin cancer risk.

119. Koshizuka, K. et al. Combined effect of vitamin D3 of 1α,25(OH)2D3 (DN‑101) and docetaxel Carcinogenesis 28, 390–397 (2007). analogs and paclitaxel on the growth of MCF‑7 breast treatment. Patients with androgen-independent 161. Park, K., Woo, M., Nam, J. & Kim, J. C. Start codon cancer cells in vivo. Breast Cancer Res. Treat. 53, prostate cancer who were treated with polymorphisms in the vitamin D receptor and

113–120 (1999). 1α,25(OH)2D3and docetaxel had improved survival colorectal cancer risk. Cancer Lett. 237, 199–206 120. Krishnan, A. V. et al. Novel pathways that contribute to compared with the placebo and docetaxel-only (2006). the anti-proliferative and chemopreventive activities of groups. 162. Sweeney, C. et al. Haplotype analysis of common calcitriol in prostate cancer. J. Steroid Biochem. Mol. 139. Trump, D. L., Potter, D. M., Muindi, J., Brufsky, A. & vitamin D receptor variants and colon and rectal Biol. 103, 694–702 (2007). Johnson, C. S. Phase II trial of high-dose, intermittent cancers. Cancer Epidemiol. Biomarkers Prev. 15, 121. Ferrero, D. et al. Combined differentiating therapy for calcitriol (1α, 25 dihydroxyvitamin D3) and 744–749 (2006). myelodysplastic syndromes: a phase II study. Leukemia dexamethasone in androgen-independent prostate 163. Taylor, J. A. et al. Association of prostate cancer with Res. 20, 867–876 (1996). cancer. Cancer 106, 2136–2142 (2006). vitamin D receptor gene polymorphism. Cancer Res. 122. Slapak, C. A., Desforges, J. F., Fogaren, T. & 140. Haynes, R. C. in The Pharmacological Basis of 56, 4108–4110 (1996). Miller, K. B. Treatment of acute myeloid leukemia in Therapeutics (eds A. G. Gilman, T. W. Rall, A. S. Nies & 164. Ingles, S. A. et al. Association of prostate cancer risk the elderly with low-dose cytarabine, hydroxyurea, and P. Taylor) 1496–1522 (Pergamon Press, New York; with genetic polymorphisms in vitamin D receptor and calcitriol. Am. J. Hematol. 41, 178–183 (1992). 1990). androgen receptor. J. Natl Cancer Inst. 89, 166–170 123. Hellstrom, E. et al. Treatment of myelodysplastic 141. Yu, W. D. et al. Enhancement of 1α, 25- (1997). syndromes with retinoic acid and 1 dihydroxyvitamin D3-mediated antitumor activity with First paper to show genetic polymorphism in VDR is alpha‑hydroxy‑vitamin D3 in combination with low- dexamethasone. J. Natl Cancer Inst. 90, 134–141 associated with cancer risk. dose ara‑C is not superior to ara‑C alone. Results from (1998). 165. Liu, Z. et al. Polymorphisms of vitamin D receptor gene a randomized study. The Scandinavian Myelodysplasia 142. Beer, T. M., Hough, K. M., Garzotto, M., Lowe, B. A. & protect against the risk of head and neck cancer. Group (SMG). Eur. J. Haemat. 45, 255–261 (1990). Henner, W. D. Weekly high-dose calcitriol and Pharmacogenet. Genomics 15, 159–165 (2005). 124. Hansen, C. M., Hamberg, K. J., Binderup, E. & docetaxel in advanced prostate cancer. Semin. Oncol. 166. Matusiak, D., Murillo, G., Carroll, R. E., Mehta, R. G. & Binderup, L. Seocalcitol (EB 1089): a vitamin D 28, 49–55 (2001). Benya, R. V. Expression of vitamin D receptor and 25- analogue of anti-cancer potential. Background, 143. Bouillon, R., Okamura, W. H. & Norman, A. W. hydroxyvitamin D3–1α-hydroxylase in normal and design, synthesis, pre-clinical and clinical evaluation. Structure-function relationships in the vitamin D malignant human colon. Cancer Epidemiol. Biomarkers Curr. Pharm. Des. 6, 803–828 (2000). endocrine system. Endocr. Rev. 16, 200–257 (1995). Prev. 14, 2370–2376 (2005). 125. Masuda, S. & Jones, G. Vitamin D analogs—drug 144. Lewis, J. S. & Jordan, V. C. Selective estrogen receptor 167. Palmer, H. G. et al. The transcription factor SNAIL design based on proteins involved in vitamin D signal modulators (SERMs): mechanisms of represses vitamin D receptor expression and transduction. Curr. Drug Targets 3, 43–66 (2003). anticarcinogenesis and drug resistance. Mutat. Res. responsiveness in human colon cancer. Nature Med. 126. Muindi, J. R. et al. A limited sampling method for the 591, 247–263 (2005). 10, 917–919 (2004). estimation of serum calcitriol area under the curve in 145. Yee, Y. K., Chintalacharuvu, S. R., Lu, J. & Nagpal, S. 168. Beer, T. M. & Myrthue, A. Calcitriol in the treatment of cancer patients. J. Clin. Pharmacol. 43, 894–900 Vitamin D receptor modulators for inflammation and prostate cancer. Anticancer Res. 26, 2647–2651 (2003). cancer. Mini Rev. Med. Chem. 5, 761–778 (2005). (2006).

REVIEWS

169. Trump, D. L., Muindi, J., Fakih, M., Yu, W. D. & individuals with colorectal cancer. Lancet 359, Competing interests statement Johnson, C. S. Vitamin D compounds: clinical 1831–1832 (2002). The authors declare competing financial interests, see web development as cancer therapy and prevention agents. 176. Diesel, B. et al. Vitamin D(3) metabolism in human version for details. Anticancer Res. 26, 2551–2556 (2006). glioblastoma multiforme: functionality of CYP27B1 170. Lohnes, D. & Jones, G. Further metabolism of 1α, 25- splice variants, metabolism of calcidiol, and effect of dihydroxyvitamin D3 in target cells. J. Nutr. Sci. calcitriol. Clin. Cancer Res. 11, 5370–5380 (2005). DATABASES Vitaminol. (Tokyo) Spec No, 75–78 (1992). 177. Diesel, B. et al. Towards a complete picture of splice Entrez Gene: http://www.ncbi.nlm.nih.gov/entrez/query. 171. Posner, G. H. et al. Potent, low-calcemic, selective variants of the gene for 25-hydroxyvitamin D31α- fcgi?db=gene inhibitors of CYP24 hydroxylase: 24-sulfone analogs of hydroxylase in brain and skin cancer. J. Steroid BCL2 | BGLAP | CBP | CCNA1 | CCNC | CCND1 | CCND3 | the hormone 1alpha, 25-dihydroxyvitamin D3. J. Biochem. Mol. Biol. 89–90, 527–532 (2004). CCNE1 | CDH1 | CDKN1A | CDKN1B | COX2 | CYP24A1 | Steroid Biochem. Mol. Biol. 89–90, 5–12 (2004). 178. Reichrath, J. et al. Analysis of the vitamin D system in CYP27A1 | CYP27B1 | EGFR | GADD45 | HDAC | ID1 | ID2 | 172. Tsunedomi, R. et al. Patterns of expression of cutaneous squamous cell carcinomas. J. Cutan. Pathol. IGFBP3 | MEKK1 | MYC | PI3K | PTH | SKIP | SKP2 | SMAD3 | cytochrome P450 genes in progression of hepatitis C 31, 224–231 (2004). SMRT | TGFβ | THBS1 | TF2B | TK1 | TYMS | VDR | VEGF | WSTF virus-associated hepatocellular carcinoma. Int. J. 179. Mimori, K. et al. Clinical significance of the National Cancer Institute: http://www.cancer.gov Oncol. 27, 661–667 (2005). overexpression of the candidate oncogene CYP24 in breast cancer | colorectal cancer | prostate cancer 173. et al. Ann. Oncol. 15 Mitschele, T. Analysis of the vitamin D system esophageal cancer. , 236–241 (2004). FURTHER INFORMATION in basal cell carcinomas (BCCs). Lab. Invest. 84, Authors’ homepage: http://roswellpark.org 693–702 (2004). Acknowledgements 174. Friedrich, M. et al. Analysis of 25-hydroxyvitamin D.L.T. and C.S.J. are supported by grants from the US SUPPLEMENTARY INFORMATION D3‑1alpha‑hydroxylase in normal and malignant breast National Cancer Institute, Department of Defense, American See online article: S1 (table) | S2 (table) tissue. Anticancer Res. 26, 2615–2620 (2006). Cancer Society and The Roswell Park Alliance Foundation. 175. Ogunkolade, B. W. et al. Expression of 25- C.S.J. is also supported by the Robert, Lew and Ann Wallace All links are active in the online pdf hydroxyvitamin D‑1‑α-hydroxylase mRNA in Endowment Fund.