Article

1,25-(OH)2D-24 Hydroxylase (CYP24A1) Deficiency as a Cause of Nephrolithiasis

Galina Nesterova,* May Christine Malicdan,* Kaori Yasuda,† Toshiyuki Sakaki,† Thierry Vilboux,* Carla Ciccone,* | | | Ronald Horst,‡ Yan Huang,§ Gretchen Golas, Wendy Introne, Marjan Huizing,* David Adams,§ Cornelius F. Boerkoel,§ | Michael T. Collins,¶ and William A. Gahl*§

Summary Background and objectives Elevated serum D with hypercalciuria can result in nephrocalcinosis and *Medical Genetics a Branch, National nephrolithiasis. This study evaluated the cause of excess 1,25-dihydroxycholecalciferol (1 ,25(OH)2D3)inthe development of those disorders in two individuals. Research Institute, National Institutes of Design, setting, participants, & measurements Two patients with elevated levels and nephrocalcinosis Health, Bethesda, Maryland; or nephrolithiasis were investigated at the National Institutes of Health (NIH) Clinical Center and the NIH † Department of Undiagnosed Diseases Program, by measuring calcium, phosphate, and vitamin D metabolites, and by Biotechnology, performing CYP24A1 mutation analysis. Faculty of Engineering, Toyama Results Both patients exhibited hypercalciuria, hypercalcemia, low , elevated vitamin D Prefectural University, a Toyama, Japan; (1 ,25(OH)2D3), normal 25-OHD3, decreased 24,25(OH)2D, and undetectable activity of 1,25(OH)2D-24- ‡Heartland Assays, a hydroxylase (CYP24A1), the that inactivates 1 ,25(OH)2D3. Both patients had bi-allelic mutations in Ames, Iowa; CYP24A1 leading to loss of function of this enzyme. On the basis of dbSNP data, the frequency of predicted §Undiagnosed deleterious bi-allelic CYP24A1 variants in the general population is estimated to be as high as 4%–20%. Diseases Program, Office of Rare fi Diseases Research, Conclusions The results of this study show that 1,25(OH)2D-24-hydroxylase de ciency due to bi-allelic mutations National Human in CYP24A1 causes elevated serum vitamin D, hypercalciuria, nephrocalcinosis, and renal stones. Genome Research Clin J Am Soc Nephrol 8: 649–657, 2013. doi: 10.2215/CJN.05360512 Institute, National Institutes of Health, Bethesda, Maryland; | Office of the Clinical Introduction parathyroid hormone (PTH), calcitonin, and 1a,25- Director, National a Human Genome Although nephrocalcinosis and nephrolithiasis are dihydroxyvitamin D3 (1 ,25(OH)2D3), the active Research Institute, distinct entities, evidence suggests a common under- form of vitamin D. Today, hypercalciuria is con- National Institutes of lying mechanism (1). Nephrocalcinosis refers to the sidered a complex trait, governed by several genetic Health, Bethesda, Maryland; and diffuse precipitation of calcium salts within the tu- mechanisms. ¶ a Skeletal Clinical bules, tubular epithelium, and/or interstitial tissue One of those mechanisms involves 1 ,25(OH)2D3, Studies Unit, of the kidney (2). It involves the medulla in 98% of whichactsoncalciumandphosphatetostimulate Craniofacial and cases, and is readily detected by ultrasonography or mobilization from bone, reabsorption by the kidney, Skeletal Diseases a Branch, National computed tomography (CT). Nephrolithiasis refers to and absorption by the intestine, 1 ,25(OH)2D3 is Institute of Dental and renal stones, generally visible on plain radiographs, formed through 1-hydroxylation of 25(OH)D3 (chole- CYP27B1, Craniofacial Research, ultrasounds, or CT scans; the majority of renal stones calciferol) by the product of which is National Institutes of are composed of calcium salts (3). Because calcium present in many tissues; 25(OH)D3 is formed by 25- Health, Bethesda, precipitation is common to both nephrocalcinosis hydroxylation of vitamin D in the liver (Figure 1) (8). Maryland a and nephrolithiasis, the two entities often occur Excessive 1 ,25(OH)2D3 leads to absorptive hypercal- together (4). In addition, both nephrolithiasis and cemia and/or hypercalciuria (9), as first described in Correspondence: nephrocalcinosis frequently result from metabolic ab- Dr. Galina Nesterova, 1928 by Kreitmair and Moll (10). Medical Genetics normalities such as hyperphosphaturia, hyperoxaluria, Both major vitamin D metabolites, 25(OH)D3 and Branch, National a hypocitruria, hyperuricosuria, defective urinary 1 ,25(OH)2D3, are inactivated by 24-hydroxylation, a Human Genome fi Research Institute, acidi cation, and especially hypercalciuria (5). Ap- process catalyzed by 1,25(OH)2D-24-hydroxylase proximately 50% of patients with nephrolithiasis (6) (CYP24A1), a mitochondrial P-450 National Institutes of – Health, 10 Center and 14% 27% of non-stone formers (7) have hypercal- mixed-function oxidase present largely in intestine Drive, Building 10, ciuria. and kidney. Bi-allelic mutations in CYP24A1 and re- Room 10C 107, Previous classifications of hypercalciuria have been duced activity of the enzyme have been associated Bethesda, MD 20892. a Email: nesterovag@ revised in view of recent evidence of intrinsic alter- with elevated levels of 1 ,25(OH)2D3 in individuals mail.nih.gov ations of calcium associated with this given large amounts of vitamin 25(OH)D3 (11). In phenomenon. is regulated by addition, a recent report demonstrated reduced 1,25 www.cjasn.org Vol 8 April, 2013 Copyright © 2013 by the American Society of Nephrology 649 650 Clinical Journal of the American Society of Nephrology

Figure 1. | Vitamin D metabolism. The first step of activation, 25-hydroxylation by CYP2R1 and CYP27A1, occurs in the liver. The second step, 1a- a hydroxylation by CYP27B1 to yield active vitamin D (i.e., 1 ,25(OH)2D3), occurs in the kidney. Inactivation of vitamin D occurs via the C-23 and C-24 oxidation pathways, catalyzed by CYP24A1 in the kidney. has no biologic activity. Larger type emphasizes important metabolites.

(OH)2D-24-hydroxylase activity in patients with mono- for Biotechnology Information (NCBI) (http://www.ncbi. allelic CYP24A1 mutations (12). nlm.nih.gov/mapview/), and the UCSC Genome Browser We present a young boy with nephrocalcinosis and an (http://genome.ucsc.edu/), which were used to design adult male patient with nephrolithiasis, both of whom have primers (available upon request). molecular and biochemical defects in CYP24A1, leading to a elevated levels of 1 ,25(OH)2D3 and hypercalciuria. Estimated Frequency of CYP24A1 Mutations The dbSNP database (http://www.ncbi.nlm.nih.gov/ projects/SNP/) was searched (March 2012) for nonsynon- Materials and Methods ymous variants within the CYP24A1 gene and their minor Patients allele frequencies (MAFs). The effect of missense variations Patients were enrolled in clinical protocol 76-HG-0238, on protein function was evaluated using pathogenicity “ Diagnosis and Treatment of Patients with Inborn Errors prediction programs, including POLYPHEN (Polymor- ” of Metabolism and Other Genetic Disorders and were phism Phenotyping; http://genetics.bwh.harvard.edu/ evaluated at the National Institutes of Health (NIH) Clin- pph/), PANTHER (Protein Analysis Through Evolution- ical Center. Written, informed consent was obtained. Pa- ary Relationships; http://www.pantherdb.org/), pMut tient 2 was enrolled in the NIH Undiagnosed Diseases (http://mmb2.pcb.ub.es:8080/PMut/), SIFT (Sorting In- Program (UDP) (13,14), dedicated to investigating the ba- tolerant from Tolerant; http://sift.jcvi.org/), Blosum62 sic defects underlying undiagnosed disorders. (Blocks of Amino Acid Substitution Matrix; ftp://ftp. ncbi.nih.gov/blast/matrices/BLOSUM62), and SNAP Cell Culture (Screening for Non-Acceptable Polymorphisms; http:// Primary fibroblasts from the patients were grown from cubic.bioc.columbia.edu/services/snap/) (Supplemental forearm punch skin biopsies. Normal control fibroblasts Methods). Missense variants were tested on 100 ethnically were racially matched and randomly selected from our matched control individuals (i.e., a Caucasian DNA panel) fibroblast bank. The cells were maintained in high-glucose (Coriell Institute of Medical Research). (4.5 g/L) DMEM supplemented with 15% FBS, 2 mM L- glutamine, nonessential amino acid solution, and penicillin- Western Blot Analyses streptomycin, as previously described (15,16). Fibroblasts were washed twice with cold phosphate- buffered saline (PBS) and lysed with radioimmunoprecipi- Mutation Screening and Genotyping tation assay buffer (50 mM Tris HCl pH 8 150 mM NaCl, Genomic DNA was isolated from PBMCs using the 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS) contain- Gentra Puregene Blood Kit (Qiagen, Valencia, CA). All ing protease inhibitors (Complete Mini, EDTA-Free; Roche CYP24A1 exons and flanking introns were PCR-amplified Diagnostics). Supernatants (10 mg) obtained after 30 min- and directly sequenced using standard protocols (Applied utes of centrifugation (15,000 rpm at 4°C) were boiled in Biosystems, Foster City, CA). Data regarding position, reducing SDS buffer and electrophoresed on 4%–20% gra- coding region sequence, and exon-intron boundaries of dient polyacrylamide gels (Invitrogen). After transfer to CYP24A1 were obtained using the Ensembl Genome nitrocellulose membranes, the proteins were blocked Browser (http://www.ensembl.org/), the National Center with 4% BSA and probed with rabbit anti-CYP24A1 Clin J Am Soc Nephrol 8: 649–657, April, 2013 Vitamin D 24-Hydroxylase Deficiency: Nephrocalcinosis and Nephrolithiasis, Nesterova et al. 651

b antibody (sc-66851; Santa Cruz) and rabbit anti- actin decreased 24,25-(OH)2D (Table 2). Renal ultrasonography (clone EP1123Y; Chemicon). Appropriate IRDye 800 CW- performed at the NIH showed extensive medullary nephro- conjugated secondary antibodies (Li-Cor Biosciences) were calcinosis. A dual energy x-ray absorptiometry scan yielded used. The antigen-antibody complexes were detected with Z-scores of 21.4 for the spine, 0.2 for the femoral neck, and the Li-Cor Odyssey infrared imaging system. 21.3 for the forearm. The patient’sunaffecteddizygotic twin brother did not have kidney stones or nephro- Biochemical Studies calcinosis, and neither did any other family member. Serum 24,25-(OH)2D levels were measured by RIA after Patient 2 (family II, patient II-2.2), a 38 year-old man, had extraction and purification by high-performance liq- recurrent kidney stones since age 25 years, hypercalcemia, a uid chromatography (HPLC). The intra-assay and interas- hypercalciuria, and marked elevation of 1 ,25(OH)2D3.His say coefficients of variation (CVs) were 9.3% and 11.8% , 24-hour urine evaluations revealed hypercalciuria without respectively (Heartland Assays LLC) (17). 1,25(OH)2 vita- hyperuricosuria or oxaluria (Table 1), high serum calcium min D testing was performed by immunoextraction and levels and suppressed intact PTH, normal serum 25(OH) a liquid chromatography–tandem mass spectrometry, simi- D3,high1 ,25(OH)2D3, and low 24,25-(OH)2D(Table2). lar to published methods (18). 1,25(OH)2D2 interassay im- He had an elevated fractional excretion of phosphate (Ta- precision was 8.5% and 8% at levels of 67 and 219 pg/ml, ble 1). An abdominal CT scan performed at age 34 years 3 respectively. Intact human fibroblast growth factor-23 demonstrated a large renal calculus (25 34 mm) in the left (FGF-23) was measured using a commercially available kidney. Treatments included extracorporeal shockwave ELISA kit (Kainos, Tokyo, Japan). The CV for the FGF-23 lithotripsy, percutaneous nephrolithotomy with subse- assay was ,5%, as reported by the manufacturer. quent residual stone fragments, Harrington rod placement for scoliosis, thiazide diuretics, sodium cellulose phos- phate, and, after his admission to the NIH Clinical Center, 1,25(OH)2D-24-Hydroxylase Assay Fibroblasts were seeded into a 35-mm dish at 23105 cells ketoconazole. Stone analysis of patient 2, performed at the per well and incubated at 37°C. After 3 hours, 1 mM 1a,25 Mayo Clinic, demonstrated 100% calcium phosphate (brushite). A gallium scan showed no evidence of granu- (OH)2D3 in ethanol (Wako Pure Chemicals, Osaka, Japan) was added to a final concentration of 1 mM. After further lomatous disease. The dual energy x-ray absorptiometry Z-scores and T-scores were 22.8 and 22.8 in the spine, incubation for 24 hours, culture medium and cells were 2 2 2 2 recovered separately. Each metabolite was analyzed as de- 1.3 and 1.8 in the femoral neck, and 2.8 and 3.0 a in the forearm, respectively,. The plasma FGF-23 level, a scribed below. To obtain the metabolites of 1 ,25(OH)2D3 possible contributing factor to the patient’s hyperphospha- using recombinant CYP24A1, a reconstituted system con- 6 taining the membrane fraction of recombinant Escherichia turia, was 64 RU/ml, within the normal range (63 45 SD). coli cells expressing human CYP24A1 was utilized (19,20). The patient had decreased urine calcium after administra- — m a tion of sodium cellulose phosphate, confirming the pres- The reaction mixture containing 10 M1 ,25(OH)2D3, 2.0 mM adrenodoxin, 0.2 mM adrenodoxin reductase, 20 ence of increased enteric calcium absorption (data not shown). On ketoconazole therapy (800 mg per day in di- nM CYP24A1, 1 mM dinucleotide ’ phosphate, 100 mM Tris HCl (pH 7.4), and 1 mM EDTA— vided doses), the patient s serum 1,25(OH)2D3, serum cal- a cium, and urine calcium decreased, and the PTH returned was incubated at 37°C for 30 minutes. 1 ,25(OH)2D3 and its metabolites were extracted using chloroform/methanol to normal (Figure 2). His male sibling (II-2.1) had a ques- (3:1, v/v). The organic phase was dried under reduced tionable history of renal stones. pressure. The resultant residue was dissolved in acetoni- Both probands adhered to low calcium and oxalate diets fl trile and applied to an HPLC column under the following with no excess sun exposure and copious uid intake. Mul- m 3 tivitamins, vitamin D supplementations, and vitamin D conditions: column, YMC-Pack ODS-AM (5 m, 4.6 300 fi mm; YMC Co., Kyoto, Japan); ultraviolet detection, 265 nm; forti ed food had been largely eliminated in both patients. 2 flow rate, 1.0 ml/min 1; column temperature, 40°C; and Biochemical Studies mobile phase, linear gradient of 20%–100% acetonitrile Fibroblast 1,25(OH) D-24-hydroxylase activity was as- aqueous solution per 25 minutes followed by 100% aceto- 2 sayedbyexposingculturedcellsto1a,25(OH) D and nitrile for 20 minutes. 2 3 measuring the 24-hydroxylated derivative. HPLC profiles of both patients’ fibroblasts showed no identifiable metab- a Results olites of 1 ,25(OH)2D3, indicating negligible1,25(OH)2D- Patients 24-hydroxylase activity compared with non-stone forming Patient 1 (family I, patient I-2.2) is a 9-year-old boy with controls (Figure 3A) (20). Moreover, CYP24A1 protein ex- fi nephrocalcinosis diagnosed by routine renal ultrasonogra- pression was reduced in the broblasts of both patient 1 phy performed because of a urinary tract infection at 3 and patient 2 (Figure 3, B and C) compared with a normal years of age. Initial investigations outside the NIH showed control. hypercalciuria up to 8 mg/kg per day (normal ,4.0 mg/kg a per day) and increased plasma 1 ,25(OH)2D3 to 86 pg/ml Molecular Analyses (normal 18–64 pg/ml). Repeated 24-hour urine evalua- We focused our molecular analysis on candidate tions revealed significant hypercalciuria, and no hyper- involved in vitamin D metabolism. CYP27B1 sequence uricosuria or oxaluria (Table 1), elevated serum calcium analysis revealed no mutations in either patient, eliminat- levels and appropriately suppressed intact PTH, as well ing the possibility of constitutively upregulated produc- a a as normal serum 25(OH)D2,elevated1 ,25(OH)2D3,and tion of 1 ,25(OH)2D3. 652 Clinical Journal of the American Society of Nephrology

Table 1. Patients’ 24-hour urine values

a Calcium/ Phosphate Excretion Fractional Excretion Urine pH Calcium Excretion a Creatinine Ratio (g/24 h) of PO4 (%) Patient 1 6.5 9.3 mg/kg per day 0.42 0.51 12 Patient 2 6.0 160–405 mg/d 0.33 0.97 34

aThese laboratory values were elevated compared with normal values.

Table 2. Patients’ blood laboratory values

Plasma 1a,25 24,25 Ionized Ca2+ Parathyroid 25-OHD Creatinine 3 (OH) D (OH) D (mmol/L) Hormone (pg/ml) (ng/ml) 2 3 2 (mg/dl) (pg/ml) (ng/ml)

Patient 1 1.23–1.34 3 0.4 71 79–115 0.64 Patient 2 1.32–1.41 3–10 1.2–1.3 39–59 83–160 0.33 Normal 1.12–1.32 16–87 0.5–1.2 10–80 18–64 1.2–2.6b reference rangea

aIn the authors’ laboratories. bObtained for this study.

Figure 2. | Effect of ketoconazole treatment in patient 2 on 1,25 vitamin D levels, hypercalcemia, hypercalciuria, and parathyroid hormone. Patient 2 was treated with ketoconazole and the dosage was escalated to 800 mg per day in divided doses. The effects monitored before the initiation of therapy, approximately 1 month after the initiation of therapy, and approximately 1 month after discontinuation. The following a levels are displayed: (A) 1 ,25(OH)2D3, (B) serum ionized calcium, (C) 24-hour urine calcium, and (D) serum parathyroid hormone.

However, sequencing of CYP24A1 in patient 1 revealed I-2.2). Patient 1 also had a heterozygous transition of a heterozygous 3-bp deletion (c.428_430del; p.E143del) in- T to C (c.443T.C), inherited from his father (I-1.1) and herited from his mother (I-1.2) and predicted to remove altering a moderately conserved amino acid (p.L148P). a highly conserved glutamate at position 143 (Figure 4; His unaffected twin (I-2.1) harbored only the p.E143del Clin J Am Soc Nephrol 8: 649–657, April, 2013 Vitamin D 24-Hydroxylase Deficiency: Nephrocalcinosis and Nephrolithiasis, Nesterova et al. 653

Figure 3. | CYP24A1 activity in patient 1 and patient 2. CYP24A1 activity was measured in fibroblasts from patients 1 and 2 and compared with values for normal non-stone formers. (A) CYP24A1 activity was measured in fibroblasts from patients 1 and 2, and compared with that of normal non-stone formers (all fibroblasts are the same passage number 3). Fibroblasts from patient 1 (P1) and patient 2 (P2) produced no metabolites of a 1 ,25(OH)2D3. In contrast, normal fibroblasts produced metabolites that have the same retention time (18.4 minutes) of metabolites using a a recombinant CYP24A1 (as indicated by the arrow). Our previous study revealed that this peak contained 1- ,24R,25(OH)2D3 and 1- ,23S,25- (OH)2D3 in the ratio of 4:1 (19). Although we found no other metabolites in normal fibroblasts, further metabolites may be observed with a increasing reaction time. The shoulder on the main substrate peak is the 6-s-cis form of 1- ,25(OH)2D3 generated by rotation around the 6,7 carbon bond. The interconversion between the 6-s-trans- and 6-s-cis-forms has a low energy barrier and therefore occurs rapidly in solution at room temperature. (B) CYP24A1 protein amount is reduced in patient 1 and patient 2. Western blot of whole fibroblast lysates probed with CYP24A1 antibody shows reduced protein amount in patients compared with normal. (C) b-actin serves as loading control quantification by densitometry. mutation. Patient 2 exhibited bi-allelic mutations in also included in Supplemental Table 1; they were not CYP24A1, including the p.E143del deletion that he inherited reported in dbSNP as of March 2012. from his father (Figure 4; II-1.1). The second mutation was a heterozygous transition in exon 9 (c.1226T.C), changing a moderately conserved amino acid (p.L409S); Discussion his mother (II-1.2) and brother (II-2.1) were heterozygous Nephrolithiasis represents a major global health problem for this variant. with a lifetime prevalence estimated at 10%–15%, depend- ing upon age, sex, race, and geographic location (21–23). Frequency of CYP24A1 Mutations Family history of stones has been reported in close to 40% A search of dbSNP for CYP24A1 variants identified 37 of patients with nephrolithiasis (24,25). In contrast, the fre- nonsynonymous single nucleotide polymorphisms (SNPs). quency of nephrocalcinosis is not available because most Reported MAFs within this group varied between 0.001 patients are asymptomatic, but detection of nephrocalcino- and 0.075 (19 SNPs); 16 SNPs had not reported MAFs, sis is increasing due to the performance of routine diag- (Table 3 and Supplemental Table). We applied six patho- nostic renal ultrasonography (26,27). Prevention of genicity prediction programs (Blossum 62, POLYPHEN, nephrocalcinosis and nephrolithiasis could have wide- SIFT, Pmut, Panther, and SNAP) to estimate the deleteri- spread beneficial effects, but prophylaxis should be based ousness of each variant (Supplemental Table 1). on understanding the etiologies. Thirteen of the SNPs were considered likely deleterious Both nephrocalcinosis and nephrolithiasis exhibit hyper- (i.e., evaluated by POLYPHEN to have an 80%–90% calciuria as a manifestation of abnormal calcium handling, a chance of causing a functional defect in enzyme activity) which can be related to the level of 1 ,25(OH)2D3,orcal- s (Table 3). The cumulative allele frequency of these var- citriol. This hormone stimulates the synthesis of epithelial iants was 0.140. If the variant with unusually high MAF, calcium channels, calbindin 9, and the calcium adenosine p.M374T, was discarded, then the cumulative frequency triphosphatase pump to upregulate calcium absorption by a was 0.065. These estimates do not include the 16 variants the kidney and intestine. Elevated 1 ,25(OH)2D3 can also in dbSNP considered neutral or possibly deleterious by adversely affect the kidney, because calcium ions alter mi- POLYPHEN, or those listed without frequencies (Supple- tochondrial structure and metabolism, causing damage to mental Table 1). Published disease-causing mutations were renal epithelial cells, tubular necrosis, and calcium 654 Clinical Journal of the American Society of Nephrology

Figure 4. | Family pedigree and molecular analysis. The pedigrees of the two families are illustrated. Affected members are shown in black squares, whereas unaffected members are denoted in white solid objects. Patient 1 (I-2.2) harbored a 3-bp deletion, p.E143del, inherited from his mother (I-1.2), and a p.L148P missense mutation that was inherited from his father (I-1.1). Patient 2 (II-2.2) also had the same p.E143del inherited from his father (II-1.1), and a p.L409S inherited from his mother (II-1.2).

a deposition (10). In the intestine, excess 1 ,25(OH)2D3 leads (32,33). The p.E143del mutation was reported by to hyperabsorptive hypercalcemia and hypercalciuria. Schlingmann et al. in patients with increased sensitivity Because vitamin D 24-hydroxylase is the key regulator in to vitamin D supplementation (11), and the same mutation a preventing the development of high levels of 1 ,25(OH)2D3 has been found in patients with nephrolithiasis (34). (28), we pursued a defect in this enzyme as the cause of Our pathogenicity and frequency assessment of the non- nephrocalcinosis and nephrolithiasis in our patients. Sev- synonymous CYP24A1 variants reported in dbSNP predicts eral lines of evidence supported this hypothesis. First, both that bi-allelic pathogenic defects in CYP24A1 may account a fi of our patients had increased 1 ,25(OH)2D3 and very low for a signi cant portion of all calcium-containing renal levels of 24,25-(OH) 2D in the blood (Table 2). Low PTH stone patients. detected in our patients is not commonly seen in patients Human CYP24A1 contains 514 amino acids and pos- with high urine calcium and high serum calcium. Second, sesses both 23- and 24-hydroxylating activity (Figure 1) the cultured fibroblasts of both patients showed dimin- (8). The purified CYP24A1 has an absorption spectrum ished 1,25(OH)2D-24-hydroxylase activity (Figure 3A) characteristic of P450 (35). The 1,25(OH)2D-24- and reduced amounts of Cyp24A1 protein (Figure 3, B hydroxylase molecule resides within the mitochondria of re- and C). The occurrence of small concentrations of 24,25 nal tubular cells of normal kidney (36) and is expressed by a (OH)2D in serum could be explained by the presence of most 1 25(OH)2D3-responsive tissues (37). After the intes- another enzyme, Cyp27A1, which catalyzes multiple hy- tine and kidney, the skin has the highest CYP24A1 expres- droxylation steps involving vitamin D metabolites (29). sion (38). The abundance of this enzyme in tissues plays a fi The naturally occurring 24,25(OH)2D has the 24(R) con g- critical role in the removal of vitamin D metabolites (39). uration (30)(Figure 1). Studies of the Cyp24a1-null mouse also support a catabolic a Finally, our patients had bi-allelic mutations in the role for CYP24A1, because the clearance of 1 25(OH)2D3 is CYP24A1 t gene. Patient 1 has a p.E143del, a known dele- dramatically reduced in these mice; the plasma 1/2 in- terious change, and a second variant, p.L148P. Residue 148 creased from 6 to 60 h when CYP24A1 was absent. (31,37) directly interacts with the enzyme’s substrate, and the In addition, mutant rat Cyp24a1 had less hydroxylating ac- L148P change decreases enzyme activity by 25%–50% tivity than the wild-type enzyme (40). (31,32). Patient 2 has the p.E143del as well as the previously Despite the fact that CYP24A1 has equivalent Km values a reported p.L409S mutation (31). p.L409S weakens the binding for 25(OH)D3 and 1 ,25(OH)2D3 (41), our patients had a of 1,25-dihydroxyvitamin D to 1,25(OH)2D-24-hydroxylase normal levels of 25(OH)D3 in the face of elevated 1 ,25 Clin J Am Soc Nephrol 8: 649–657, April, 2013 Vitamin D 24-Hydroxylase Deficiency: Nephrocalcinosis and Nephrolithiasis, Nesterova et al. 655

hypercalciuria. Ketoconazole, an inhibitor of 25-hydroxyvitamin Table 3. Reported CYP24A1 mutations and likely deleterious D-1a-hydroxylase, normalized calcium, vitamin D, and nonsynonymous variants in dbSNP PTH levels and may be an effective treatment for patients with 24-hydroxylase deficiency (12). However, additional Reported Minor fi Disease-Causing dbSNP ID Allele studies are needed to assess the ef cacy and safety of this Mutations Frequencya regimen (44) because ketoconazole also inhibits other P450 enzymes, including the steroidogenic pathways producing p.L148Pb rs139763321 — testosterone, , and (45). Extended use of c fi p.R396W rs114368325 0.001 ketoconazole in 1,25(OH)2D-24-hydroxylase de ciency p.E143delb,c,d — 0.002d b,c may be problematic, although its long-term safety and ef- p.L409S rs6068812 0.003 ficacy in Cushing’s syndrome (up to 83 months) appears p.E1513 —— c ——encouraging (46,47). p.R159Q CYP24A1 p.E322Kc ——Previous reports have shown mutations in p.A475fsX490c ——children whose high vitamin D intake led to idiopathic c.732+1G.A (splice site ——infantile hypercalcemia (IIH) (11,48), and others hypothe- mutation)d sized that CYP24A1 dysregulation causes hypercalcemia c.733–2A.G (splice site ——and nephrolithiasis (31). We have shown that CYP24A1 mutation)d mutations are, in fact, associated with both nephrocalcino- Deleterious variants sis and nephrolithiasis, and have extended the phenotypic in dbSNP spectrum of CYP24A1 defects. Our findings are derived p.R120H rs114476330 0.009 from investigations into patients with suspected novel p.P126S rs148084028 — metabolic disorders; we did not specifically target patients p.R157W rs35873579 0.035 p.P375L rs189801930 — with renal disorders or disorders of vitamin D metabolism. p.M374T rs6022990 0.075 Genome-wide association studies have demonstrated fl CYP24A1 p.C380Y rs150006710 0.001 the in uence of variants on vitamin D concen- p.R396Qc rs143934667 — trations (49). In our patients, bi-allelic mutations in p.Y407N rs140189382 — CYP24A1 (50) represent an autosomal recessive disorder p.R439H rs141152573 0.001 involving enzyme deficiency. The sum of the MAFs of p.V457I rs112596218 0.002 published mutations plus mutations reported as deleteri- — p.R481C rs143523685 ous in dbSNP is 0.140. According to the Hardy-Weinberg p.A510V rs116065115 0.011 principal [(p2)+(2pq)+(q2) = 1], the frequency of a reces- Total deleterious minor 0.140 sive disorder with this allele frequency will be (0.140 (2)), allele frequency or 1960 per 100,000 individuals in the general population. aMinor allele frequency, as reported in dbSNP (March 2012). This includes the p.M374T mutation with an MAF of 0.075 Dashes indicate no minor allele frequency reported in dbSNP or (Table 3). If this most common variant is disregarded, the fi not computed in Exome Sequencing Project cohort populations frequency of 1,25(OH)2D-24-hydroxylase de ciency is es- (4548 control , National Heart Lung and Blood timated at 420 per 100,000. The lifetime risk of renal stones Institute Exome Sequencing Project). in the general population is 10%, or 10,000 per 100,000, so b Decreased CYP24A1 enzyme activity; reported in this study, the estimated frequency of kidney stones due to 1,25 fi measured in broblasts. (OH) D-24-hydroxylase deficiency will be between 420 c in 2 Enzyme activity decreased, measured in mutants expressed and 1960 per 10,000, or 4%–20%. This may be an overes- vitro. timate; however, some CYP24A1 polymorphisms may be dThese variants are not reported in dbSNP, but are reported in publications (12,45). associated with mild and more severe diseases, including IIH (31). Nevertheless, recognition of 1,25(OH)2D-24- hydroxylase has important implications (51), particularly in patients who have high urine calcium, high serum cal- (OH)2D3. This could be explained by the fact that 25(OH) cium, and low PTH, a pattern that is not commonly seen D3 can undergo hydroxylation by two different enzymes, by experienced clinicians. Further studies involving a a CYP27B1 and CYP24A1, whereas 1 ,25(OH)2D3 under- larger number of kidney stone formers are needed to de- goes catabolic hydroxylation only by CYP24A1. termine if the hydroxylated forms of vitamin D (not 25 Besides nephrocalcinosis and nephrolithiasis, osteo- (OH)D3 levels) should be routinely measured in patients penia may be a clinical manifestation of 1,25(OH)2D-24- with nephrocalcinosis, nephrolithiasis, and hypercalciuria hydroxylase deficiency. Our pediatric patient manifested of undetermined etiology. osteopenia of the spine and radius, and our adult patient In summary, we demonstrate that one cause of nephro- had osteoporosis of the spine and forearm as well as osteo- calcinosis and nephrolithiasis is elevated vitamin D due to fi penia of the femoral neck. This phenomenon may be be- 1,25(OH)2D-24-hydroxylase de ciency. Our analysis of a cause 1 ,25(OH)2D3 stimulates osteoclastic resorption of whole exome sequencing data suggests that between 4% bone (42), a finding long recognized to occur in vitamin D and 20% of all calcium-containing kidney stone patients intoxication (43). may have this enzyme deficiency. Identification of patients fi Our adult patient with nephrolithiasis attempted several with vitamin 1,25(OH)2D-24-hydroxylase de ciency could different therapies. Thiazide diuretics and sodium cellu- prompt salutary avoidance of vitamin D–supplemented lose phosphate apparently had beneficial effects on the dietary products. Our study was limited by having clinical 656 Clinical Journal of the American Society of Nephrology

and molecular data on only two patients, but the findings Huizing M, Gahl WA: NBEAL2 is mutated in gray platelet syn- a provide a basis for future investigations into the mecha- drome and is required for biogenesis of platelet -granules. Nat Genet 43: 732–734, 2011 nism of nephrolithiasis related to vitamin D metabolism. 17. Horst RL, Littledike ET, Gray RW, Napoli JL: Impaired 24,25- dihydroxyvitamin D production in anephric human and pig. Acknowledgments J Clin Invest 67: 274–280, 1981 The authors thank Genia Dubrovsky, Patra Yeetong, and Dimitre 18. Strathmann FG, Laha TJ, Hoofnagle AN: Quantification of 1a,25-dihydroxy vitamin D by immunoextraction and liquid Simeonov for their technical assistance in sequencing analysis. chromatography-tandem mass spectrometry. Clin Chem 57: This work was supported by the Intramural Research Program, 1279–1285, 2011 National Human Genome Research Institute, National Institute of 19. Abe D, Sakaki T, Kusudo T, Kittaka A, Saito N, Suhara Y,Fujishima Dental and Craniofacial Diseases, and Undiagnosed Diseases Pro- T, Takayama H, Hamamoto H, Kamakura M, Ohta M, Inouye K: gram, and Office of Rare Disorders Research, National Institutes of Metabolism of 2 alpha-propoxy-1 alpha,25-dihydroxyvitamin D3 and 2 alpha-(3-hydroxypropoxy)-1 alpha,25- Health, Bethesda, Maryland. dihydroxyvitamin D3 by human CYP27A1 and CYP24A1. Drug Metab Dispos 33: 778–784, 2005 Disclosures 20. Kusudo T, Sakaki T, Abe D, Fujishima T, Kittaka A, Takayama H, None. Hatakeyama S, Ohta M, Inouye K: Metabolism of A-ring dia- stereomers of 1alpha,25-dihydroxyvitamin D3 by CYP24A1. Biochem Biophys Res Commun 321: 774–782, 2004 References 21. Trinchieri A: Epidemiology of urolithiasis: An update. Clin Cases Miner Bone Metab 5: 101–106, 2008 1. Sayer JA, Carr G, Simmons NL: Nephrocalcinosis: Molecular 22. Curhan GC: Epidemiology of stone disease. Urol Clin North Am insights into calcium precipitation within the kidney. Clin Sci 34: 287–293, 2007 (Lond) 106: 549–561, 2004 23. Romero V, Akpinar H, Assimos DG: Kidney stones: A global 2. Vervaet BA, Verhulst A, D’Haese PC, De Broe ME: Nephro- picture of prevalence, incidence, and associated risk factors. Rev calcinosis: New insights into mechanisms and consequences. Nephrol Dial Transplant 24: 2030–2035, 2009 Urol 12: e86–e96, 2010 24. Trinchieri A, Mandressi A, Luongo P, Coppi F, Pisani E: Familial 3. Knoll T, Schubert AB, Fahlenkamp D, Leusmann DB, Wendt- Nordahl G, Schubert G: Urolithiasis through the ages: Data on aggregation of renal calcium stone disease. JUrol139: 478–481, more than 200,000 urinary stone analyses. JUrol185: 1304– 1988 1311, 2011 25. Curhan GC, Willett WC, Rimm EB, Stampfer MJ: Family history 4. Miller NL, Humphreys MR, Coe FL, Evan AP, Bledsoe SB, Handa and risk of kidney stones. J Am Soc Nephrol 8: 1568–1573, 1997 SE, Lingeman JE: Nephrocalcinosis: Re-defined in the era of en- 26. Sas DJ: An update on the changing epidemiology and metabolic dourology. Urol Res 38: 421–427, 2010 risk factors in pediatric kidney stone disease. Clin J Am Soc 5. Stechman MJ, Loh NY, Thakker RV: Genetics of hypercalciuric Nephrol 6: 2062–2068, 2011 nephrolithiasis: Renal stone disease. Ann N Y Acad Sci 1116: 27. Habbig S, Beck BB, Hoppe B: Nephrocalcinosis and urolithiasis 461–484, 2007 in children. Kidney Int 80: 1278–1291, 2011 6. Cioppi F, Taddei L, Brandi ML, Croppi E: Idiopathic hyper- 28. Masuda S, Prosser DE, Guo YD, Kaufmann M, Jones G: Gener- calciuria and calcium renal stone disease: Our cases. Clin Cases ation of a homology model for the human , Miner Bone Metab 6: 251–253, 2009 CYP24A1, and the testing of putative substrate binding residues 7. Curhan GC, Willett WC, Speizer FE, Stampfer MJ: Twenty-four- by site-directed mutagenesis and enzyme activity studies. Arch hour urine chemistries and the risk of kidney stones among Biochem Biophys 460: 177–191, 2007 women and men. Kidney Int 59: 2290–2298, 2001 29. Sawada N, Sakaki T, Ohta M, Inouye K: Metabolism of vitamin D 8. Sakaki T, Kagawa N, Yamamoto K, Inouye K: Metabolism of vi- (3) by human CYP27A1. Biochem Biophys Res Commun 273: tamin D3 by P450. Front Biosci 10: 119–134, 2005 977–984, 2000 9. DeLuca HF: Metabolism of vitamin D: Current status. Am J Clin 30. Tanaka Y, DeLuca HF, Ikekawa N, Morisaki M, Koizumi N: De- Nutr 29: 1258–1270, 1976 termination of stereochemical configuration of the 24-hydroxyl 10. Scarpelli DG, Tremblay G, Pearse AGE: A comparative cyto- group of 24,25-dihydroxyvitamin D3 and its biological impor- chemical and cytologic study of vitamin D induced nephro- tance. Arch Biochem Biophys 170: 620–626, 1975 calcinosis. Am J Pathol 36: 331–353, 1960 31. Jones G, Prosser DE, Kaufmann M: 25-Hydroxyvitamin D-24- 11. Schlingmann KP, Kaufmann M, Weber S, Irwin A, Goos C, John hydroxylase (CYP24A1): Its important role in the degradation of U, Misselwitz J, Klaus G, Kuwertz-Bro¨king E, Fehrenbach H, vitamin D. Arch Biochem Biophys 523: 9–18, 2012 Wingen AM, Gu¨ran T, Hoenderop JG, Bindels RJ, Prosser DE, 32. Kaufmann M, Prosser DE, Jones G: Bioengineering anabolic vi- Jones G, Konrad M: Mutations in CYP24A1 and idiopathic in- tamin D-25-hydroxylase activity into the human vitamin D cat- fantile hypercalcemia. N Engl J Med 365: 410–421, 2011 abolic enzyme, cytochrome P450 CYP24A1, by a V391L 12. Tebben PJ, Milliner DS, Horst RL, Harris PC, Singh RJ, Wu Y, mutation. J Biol Chem 286: 28729–28737, 2011 Foreman JW, Chelminski PR, Kumar R: Hypercalcemia, hyper- 33. Ji HF, Shen L: CYP24A1 mutations in idiopathic infantile hyper- calciuria, and elevated concentrations with autosomal calcemia. N Engl J Med 365: 1741; author reply 1742–1743, dominant transmission due to CYP24A1 mutations: Effects of 2011 ketoconazole therapy. J Clin Endocrinol Metab 97: E423–E427, 34. Streeten EA, Zarbalian K, Damcott CM: CYP24A1 mutations in 2012 idiopathic infantile hypercalcemia. N Engl J Med 365: 1741– 13. Gahl WA, Tifft CJ: The NIH Undiagnosed Diseases Program: 1742, author reply 1742–1743, 2011 Lessons learned. JAMA 305: 1904–1905, 2011 35. Okuda K, Usui E, Ohyama Y: Recent progress in enzymology and 14. Gahl WA, Markello TC, Toro C, Fajardo KF, Sincan M, Gill F, molecular biology of enzymes involved in vitamin D metabo- Carlson-Donohoe H, Gropman A, Pierson TM, Golas G, Wolfe L, lism. J Lipid Res 36: 1641–1652, 1995 Groden C, Godfrey R, Nehrebecky M, Wahl C, Landis DM, Yang 36. Iwata K, Yamamoto A, Satoh S, Ohyama Y, Tashiro Y, Setoguchi T: S, Madeo A, Mullikin JC, Boerkoel CF, Tifft CJ, Adams D: The Quantitative immunoelectron microscopic analysis of the locali- National Institutes of Health Undiagnosed Diseases Program: zation and induction of 25-hydroxyvitamin D3 24-hydroxylase Insights into rare diseases. Genet Med 14: 51–59, 2012 in rat kidney. J Histochem Cytochem 43: 255–262, 1995 15. Takashima A: Establishment of fibroblast cultures. Curr Protoc 37. Akeno N, Saikatsu S, Kawane T, Horiuchi N: Mouse vitamin D- Cell Biol Chapter 2: Unit 2.1, 2001 24-hydroxylase: Molecular cloning, tissue distribution, and 16. Gunay-Aygun M, Falik-Zaccai TC, Vilboux T, Zivony-Elboum Y, transcriptional regulation by 1alpha,25-dihydroxyvitamin D3. Gumruk F, Cetin M, Khayat M, Boerkoel CF, Kfir N, Huang Y, Endocrinology 138: 2233–2240, 1997 Maynard D, Dorward H, Berger K, Kleta R, Anikster Y, Arat M, 38. Akeno N, Matsunuma A, Maeda T, Kawane T, Horiuchi N: Reg- Freiberg AS, Kehrel BE, Jurk K, Cruz P, Mullikin JC, White JG, ulation of vitamin D-1alpha-hydroxylase and -24-hydroxylase Clin J Am Soc Nephrol 8: 649–657, April, 2013 Vitamin D 24-Hydroxylase Deficiency: Nephrocalcinosis and Nephrolithiasis, Nesterova et al. 657

expression by dexamethasone in mouse kidney. J Endocrinol 48. Dauber A, Nguyen TT, Sochett E, Cole DE, Horst R, Abrams SA, 164: 339–348, 2000 Carpenter TO, Hirschhorn JN: Genetic defect in CYP24A1, the 39. Anderson PH, May BK, Morris HA: Vitamin D metabolism: New vitamin D 24-hydroxylase gene, in a patient with severe infantile concepts and clinical implications. Clin Biochem Rev 24: 13–26, hypercalcemia. J Clin Endocrinol Metab 97: E268–E274, 2012 2003 49. Wang TJ, Zhang F, Richards JB, Kestenbaum B, van Meurs JB, 40. Annalora AJ, Bobrovnikov-Marjon E, Serda R, Pastuszyn A, Berry D, Kiel DP, Streeten EA, Ohlsson C, Koller DL, Peltonen L, Graham SE, Marcus CB, Omdahl JL: Hybrid homology modeling Cooper JD, O’Reilly PF, Houston DK, Glazer NL, Vandenput L, and mutational analysis of cytochrome P450C24A1 (CYP24A1) Peacock M, Shi J, Rivadeneira F, McCarthy MI, Anneli P, de Boer of the vitamin D pathway: Insights into substrate specificity and IH, Mangino M, Kato B, Smyth DJ, Booth SL, Jacques PF, Burke membrane bound structure-function. Arch Biochem Biophys GL, Goodarzi M, Cheung CL, Wolf M, Rice K, Goltzman D, 460: 262–273, 2007 Hidiroglou N, Ladouceur M, Wareham NJ, Hocking LJ, Hart D, 41. Annalora A, Bobrovnikova-Marjon E, Serda R, Lansing L, Chiu ML, Arden NK, Cooper C, Malik S, Fraser WD, Hartikainen AL, Zhai Pastuszyn A, Iyer S, Marcus CB, Omdahl JL: Rat cytochrome G, Macdonald HM, Forouhi NG, Loos RJ, Reid DM, Hakim A, P450C24 (CYP24A1) and the role of F249 in substrate binding and Dennison E, Liu Y, Power C, Stevens HE, Jaana L, Vasan RS, catalytic activity. Arch Biochem Biophys 425: 133–146, 2004 Soranzo N, Bojunga J, Psaty BM, Lorentzon M, Foroud T, Harris 42. St-Arnaud R, Arabian A, Travers R, Barletta F, Raval-Pandya M, TB, Hofman A, Jansson JO, Cauley JA, Uitterlinden AG, Gibson Chapin K, Depovere J, Mathieu C, Christakos S, Demay MB, Q, Ja¨rvelin MR, Karasik D, Siscovick DS, Econs MJ, Kritchevsky Glorieux FH: Deficient mineralization of intramembranous bone SB, Florez JC, Todd JA, Dupuis J, Hyppo¨nen E, Spector TD: in vitamin D-24-hydroxylase-ablated mice is due to elevated Common genetic determinants of vitamin D insufficiency: A 1,25-dihydroxyvitamin D and not to the absence of 24,25- genome-wide association study. Lancet 376: 180–188, 2010 dihydroxyvitamin D. Endocrinology 141: 2658–2666, 2000 50. Sherry ST, Ward MH, Kholodov M, Baker J, Phan L, Smigielski 43. Davies M, Mawer EB, Freemont AJ: The osteodystrophy of hy- EM, Sirotkin K: dbSNP: The NCBI database of genetic variation. pervitaminosis D: A metabolic study. Q J Med 61: 911–919, 1986 Nucleic Acids Res 29: 308–311, 2001 44. Nguyen M, Boutignon H, Mallet E, Linglart A, Guillozo H, Jehan 51. Carpenter TO: Take another CYP: Confirming a novel mechanism F, Garabedian M: Infantile hypercalcemia and hypercalciuria: for “idiopathic” hypercalcemia. J Clin Endocrinol Metab 97: New insights into a vitamin D-dependent mechanism and re- 768–771, 2012 sponse to ketoconazole treatment. JPediatr157: 296–302, 2010 45. Peehl DM, Seto E, Hsu JY, Feldman D: Preclinical activity of keto- Received: May 25, 2012 Accepted: December 3, 2012 conazole in combination with calcitriol or the vitamin D analogue EB 1089 in prostate cancer cells. JUrol168: 1583–1588, 2002 Published online ahead of print. Publication date available at www. 46. Sharma S, Nieman L. Prolonged remission after long-term treat- ment with steroidogenesis inhibitors in Cushing’s syndrome caused cjasn.org. by ectopic ACTH secretion. Eur J Endocrinol 166: 531–536, 2012 47. Chou SC, Lin JD: Long-term effects of ketoconazole in the This article contains supplemental material online at http://cjasn. treatment of residual or recurrent Cushing’s disease. Endocr J 47: asnjournals.org/lookup/suppl/doi:10.2215/CJN.05360512/-/ 401–406, 2000 DCSupplemental. Supplementary Table

dbSNP ID MAF1 Blossum Polyphen SIFT Panther Pmut SNAP Reported disease causing mutations and nonsynomymous SNPs with MAF in dbSNP p.R120H rs114476330 0.009 - 1 0 -4.11417 0.5381 N p.E143del*, # - 0.002˟ ------p.L148P# rs139763321 - -3 0.999 0.01 -4.59230 0.0733 NN p.R159Q# - - 1 1 0 -2.67823 0.4965 NN p.R157W rs35873579 0.035 -3 1 0 -4.63087 0.8827 NN p.R157Q rs35051736 0.012 1 0.617 0.19 -1.64275 0.4156 NN p.K178R rs146404747 0.001 2 0.025 0.45 -1.33609 0.0816 N p.D202H rs114579367 0.007 -1 0.981 0.03 -2.71613 0.0847 N p.E206K rs115260488 0.001 1 1 0.20 -1.83365 0.4336 N p.M245I rs114930663 0.002 1 0.92 0 -2.54376 0.1471 N p.T248R rs16999131 0.015 -1 1 0.09 -2.38001 0.4019 N p.C303S rs76747058 0.025 -1 0.128 0.36 -1.45217 0.0616 N p.E322K# - - 1 1 0 -2.19785 0.2993 NN p.A332T rs116804918 0.009 - 1 0.03 -3.10125 0.1251 N p.R344H rs116548533 0.007 -1 0.128 0.03 -1.45217 0.0616 N p.R367Q rs142282494 0.001 1 0.027 0.60 -1.72945 0.5621 N p.M374T% rs6022990 0.075% -1 0.607 0 -3.09497 0.5060 NN p.C380Y rs150006710 0.001 -2 1 0.01 -4.89778 0.9316 NN p.R396W# rs114368325 0.001 -3 1 0 -5.49258 0.9135 NN p.L409S*, #* rs6068812 0.003 -2 0.999 0.01 -3.98574 0.1234 NN p.R439H rs141152573 0.001 - 1 0 -3.94494 0.6497 NN p.V457I rs112596218 0.002 3 0.003 1 0.60917 0.0367 N p.A510V rs116065115 0.011 - 0.838 0.14 -1.54737 0.4924 N Total MAF: 0.212 Total deleterious MAF: 0.140 Other nonsynonymous SNPs in dbSNP p.E105K rs147642444 -2 1 0.01 0.71 -1.15838 0.2905 N p.L129M rs149806586 - 2 0.976 0.09 -1.97344 0.0766 N p.L148P rs139763321 - -3 0.999 0.01 -4.5923 0.0733 NN p.V158A rs139655790 - - 0.022 0.02 -2.72874 0.1611 N p.L207M rs149235939 - 2 0.999 0.30 -2.45364 0.0825 N p.K209R rs138489641 - 2 1 0.43 -1.97704 0.0353 N p.R396Q# rs143934667 - 1 1 0 -3.50209 0.5003 NN p.Y407N rs140189382 - -2 1 0 -4.26812 0.2150 NN p.R481C rs143523685 - -3 1 0 -4.12556 0.7764 NN p.R505Q rs146980218 - 1 1 0.14 -2.88802 0.5593 N p.P25A rs140851407 - -1 0.049 0.09 0.21306 0.0828 N p.P126S rs148084028 - -1 1 0 -3.76463 0.0430 N p.E153K rs185120393 - 1 - 0.92 -0.73103 0.5131 N p.E258D rs190860407 - 2 0.994 0.15 -2.71834 0.0436 N p.P375L rs189801930 - -3 1 0 -4.40742 0.5923 NN p.M495V rs77167734 - 1 0.001 1 -0.45058 0.4063 N Reported artificial CYP24A1 mutations p.L148F# Artificial - 0 0.989 0.2 -2.84807 0.0922 NN p.I131F# Artificial - 0 1 0.15 -3.42230 0.1951 N p.A326G# Artificial - 0 0 0.23 -0.98296 0.1583 N Reported splice-site variants BDGP score NatGene2 NN score c.732+1G>A 0.94>0 0.742>0 c.733-2A>G 0.9>0 0.852>0

1MAF = Minor Allele Frequency, as reported in dbSNP (March 2012).

2MAF= - ; not reported MAF in dbSNP. % SNP p.M374T has an unusual high MAF (see text) ˟Not reported in dbSNP, but in publication (45) *Decreased CYP24A1 enzyme activity; reported in this paper, measured in fibroblasts. # Enzyme activity in vitro decreased, measured in expressed mutants. Gray Highlight: Likely deleterious variant predicted by at least 4 out of 6 prediction programs and/or decreased CYP24A1 enzyme activity. Deleterious ranking per program: BLOSSUM: Negative score is deleterious; Polyphen: > 0.85 is deleterious; SIFT: <0.01 is deleterious; Panther: < -3 is deleterious; Pmut: >0.5 is deleterious; SNAP: NN, Non-neutral, deleterious (N=neutral).

*This paper, measured in fibroblasts

a In vitro activity, measured in expressed mutant Supplementary Methods

SNP arrays

For SNP genotyping, genomic DNA was run on a Human 1M-Duo DNA Analysis BeadChip and the data analyzed using the GenomeStudio software (both from Illumina, San Diego, CA).

Missense Variant Prediction Tools

The effect of missense variations on protein function was evaluated using the mutation prediction programs POLYPHEN, PANTHER and PMUT.

POLYPHEN

(http://genetics.bwh.harvard.edu/pph/; POLYmorphism PHENotyping) predicts the effect of an amino acid substitution on the structure and function of a protein. POLYPHEN predictions are based on empirical rules that are applied to the sequence, as well as phylogenetic and known structural information that characterize the substitution. The Position-Specific Independent Counts (PSIC) is calculated for the two different alleles and the score for wild type and variant mapping to the known 3D structure.1

PANTHER

(http://www.pantherdb.org/; Protein ANalysis THrough Evolutionary Relationships) estimates the likelihood of a non-synonymous variant to cause loss of function of the protein. The output, the subPSEC (substitution position-specific evolutionary conservation), is the negative logarithm of the probability ratio of the wild-type and mutant amino acids at a particular position based on a library. This library contains over 5,000 protein families and 30,000 subfamilies, each represented by a multiple sequence alignment and Hidden Markov Model. PANTHER subPSEC scores are continuous from 0 to −10. A value of 0 is interpreted as a functionally neutral variant; the more negative the subPSEC value, the more deleterious the substitution. The cutoff value suggested is −3. 2-4

PMUT

(http://mmb2.pcb.ub.es:8080/PMut/) uses neural networks that have been trained with a large database of disease-associated and neutral variants to predict the impact of a given amino acid substitution. The output gives a neural network (NN) value between 0 and 1 (the higher this value, the more deleterious the variant) and a confidence value between 0 and 9 (the higher this value, the more reliable the NN) 5

SIFT

(http://sift.jcvi.org/) Scale-invariant feature transform (SIFT) predicts whether an amino acid substitution affects protein function. SIFT prediction is based on the degree of conservation of amino acid residues in sequence alignments derived from closely related sequences, collected through PSI-BLAST. SIFT can be applied to naturally occurring nonsynonymous polymorphisms or laboratory- induced missense mutations.6

BLOSSUM

Blosum62 (ftp://ftp.ncbi.nih.gov/blast/matrices/BLOSUM62; BLOcks of Amino Acid SUbstitution Matrix) is a substitution matrix for pairwise protein sequence alignments. You will encounter Blosum62 in a number of bioinformatics applications that align protein sequences or analyze the homology between sequences. It contains similarity scores for all permutations of two amino acids, assigning higher (better) scores to similar amino acids. Scores within a BLOSUM are log-odds scores that measure, in an alignment, the logarithm for the ratio of the likelihood of two amino acids appearing with a biological sense and the likelihood of the same amino acids appearing by chance. A positive score is given to the more likely substitutions while a negative score is given to the less likely substitutions

SNAP

SNAP (screening for non-acceptable polymorphisms; http://cubic.bioc.columbia.edu/services/snap/) predicts the functional effects of single amino acid substitutions. Single Nucleotide Polymorphisms (SNPs) represent a very large portion of all genetic variations. SNPs found in the coding regions of genes are often non-synonymous, changing a single amino acid in the encoded protein sequence.7 SNPs are either "neutral" in the sense that the resulting point-mutated protein is not functionally discernible from the wild-type, or they are "non-neutral" in that the mutant and wild-type differ in function. The ability to identify non-neutral substitutions in an ocean of SNPs could significantly aid targeting disease causing detrimental mutations, as well as SNPs that increase the fitness of particular phenotypes.

PREDICTION SOFTWARES FOR SPLICE-SITE MUTATIONS

The effect of splice site variations was also evaluated, using different analysis programs, including the splice site prediction tool from the Berkeley Drosophila Genome Project (BDGP) web site (http://www.fruitfly.org/seq_tools/splice.html). This is based on a generalized Hidden Markov Model to predict the strength of the possible splice site, using a neural network that has been trained by a set of 793 unrelated human genes Berkeley Drosophila Genome Project (BDGP) web site).8 Another tool used was NetGene2 (http://www.cbs.dtu.dk/services/NetGene2/), a service producing neural network predictions of splice sites in human, C. elegans and A. thaliana DNA.9 References 1. Ramensky V, Bork P, Sunyaev S. Human non-synonymous SNPs: server and survey. Nucleic Acids Res. 2002;30(17):3894–3900.

2. Thomas PD, Campbell MJ, Kejariwal A, Mi H, Karlak B, et al. PANTHER: a library of protein families and subfamilies indexed by function. Genome Res. 2003;13(9):2129–2141.

3. Thomas PD, Kejariwal A. Coding single-nucleotide polymorphisms associated with complex vs. Mendelian disease: evolutionary evidence for differences in molecular effects. Proc Natl Acad Sci U S A. 2004;101(43):15398–15403.

4. Thomas PD, Kejariwal A, Guo N, Mi H, Campbell MJ, et al. Applications for protein sequence-function evolution data: mRNA/protein expression analysis and coding SNP scoring tools. Nucleic Acids Res. 2006;34(Web Server issue):W645–650.

5. Ferrer-Costa C, Gelpi JL, Zamakola L, Parraga I, de la Cruz X, et al. PMUT: a web-based tool for the annotation of pathological mutations on proteins. Bioinformatics. 2005;21(14):3176–3178.

6. Kumar P, Henikoff S, Ng PC. Predicting the effects of coding non-synonymous variants on protein function using the SIFT algorithm. Nat Protoc. 2009;4(7):1073-81.

7. Bromberg Y, Rost B. SNAP: predict effect of non-synonymous polymorphisms on function. Nucleic Acids Res. 2007;35(11):3823–3835

8. Reese MG, Eeckman FH, Kulp D, Haussler D. Improved splice site detection in Genie. J Comput Biol. 1997;4(3):311–323.

9. Hebsgaard SM, Korning PG, Tolstrup N, Engelbrecht J, Rouze P, Brunak S. Splice site prediction in Arabidopsis thaliana DNA by combining local and global sequence information. Nucleic Acids Res. 1996; 24(17):3439-3452.