Elucidation of metabolic pathways of 25-hydroxyvitamin D3 mediated by CYP24A1 and CYP3A using Cyp24a1 knockout rats generated by CRISPR/Cas9 system

CYP24A1-deficient (Cyp24a1 KO) rats were generated using the CRISPER/Cas9 system to investigate CYP24A1-dependent or -independent metabolism of 25(OH)D3, the prohormone of calcitriol. Plasma 25(OH)D3 concentrations in Cyp24a1 KO rats were approximately twofold higher than in wild-type rats. Wild-type rats showed five metabolites of 25(OH)D3 in plasma following oral administration of 25(OH)D3, and these metabolites were not detected in Cyp24a1 KO rats. Among these metabolites, 25(OH)D3-26,23-lactone was identified as the second major metabolite with a significantly higher Tmax value than others. When 23S,25(OH)2D3 was administered to Cyp24a1 KO rats, neither 23,25,26(OH)3D3 nor 25(OH)D3-26,23-lactone was observed. However, when 23S,25R,26(OH)3D3 was administered to Cyp24a1 KO rats, plasma 25(OH)D3-26,23-lactone was detected. These results suggested that CYP24A1 is responsible for the conversion of 25(OH)D3 to 23,25,26(OH)3D3 via 23,25(OH)2D3, but enzyme(s) other than CYP24A1 may be involved in the conversion of 23,25,26(OH)3D3 to 25(OH)D3-26,23-lactone. Enzymatic studies using recombinant human CYP species and the inhibitory effects of ketoconazole suggested that CYP3A plays an essential role in the conversion of 23,25,26(OH)3D3 into 25(OH)D3-26,23-lactone in both rats and humans. Taken together, our data indicate that Cyp24a1 KO rats are valuable for metabolic studies of vitamin D and its analogs. In addition, long-term administration of 25(OH)D3 to Cyp24a1 KO rats at 110 μg/kg body weight/day resulted in significant weight loss and ectopic calcification. Thus, Cyp24a1 KO rats could represent an important model for studying renal diseases originating from CYP24A1 dysfunction.

There are two vitamin D responsive elements in the promoter region of the CYP24A1 gene, and the binding of active vitamin D to VDRs leads to profound transcriptional induction of CYP24A1. The resultant CYP24A1 protein inactivates the active form of vitamin D. This mechanism is crucial for maintaining constant plasma levels of active vitamin D. CYP24A1 dysfunction causes elevated plasma levels of 1,25(OH) 2 D3 and is associated with idiopathic infantile hypercalcemia or kidney stones, which indicates the importance of this enzyme for vitamin D metabolism (3,4).
We expressed rat and human CYP24A1 in Escherichia coli and examined their enzymatic properties (5,6). We identified that rat CYP24A1 catalyzes a six-step reaction commencing with 24R-position hydroxylation, leading to calcitroic acid, which is designated as the C-24 pathway. In contrast, human CYP24A1 catalyzes a four-step reaction starting at position 23S, leading to the formation of a 26,23-lactone.
The ratios of the C-23 and C-24 pathways vary among species. For human CYP24A1, the ratio in the initial reactions of C-23 and C-24 is approximately 1:4, whereas in rat CYP24A1, it is approximately 1:25 (6,7), based on the comparison of first step metabolic reaction. In species such as guinea pig and opossum, the C-23 pathway is, in contrast, predominant (7,8).
Recently, it has become possible to quantify various vitamin D metabolites more sensitively using LC/MS/MS, and Kaufmann et al. have succeeded in quantifying various metabolites by performing analysis after derivatization by DMEQ-TAD (13)(14)(15)(16). They demonstrated that metabolites in the C-24 and C-23 pathways, such as 24-oxo-25(OH)D3 (C-24), 24-oxo-23,25(OH) 2 D3 (C- 24), and 25(OH)D3-26,23-lactone (C-23), are not detected in Cyp24a1 KO mice (13), suggesting that Cyp24A1 plays a crucial role in both C-23 and C-24 pathways in mice. Our focus is investigating which steps of the C-24 and C-23 pathways are carried out only by CYP24A1 and which steps contribute to the extent of involvement by other enzymes. Although Cyp24a1 KO mice appear to be useful, it is difficult to evaluate detailed time courses of metabolites in terms of blood sampling volumes possible in small rodents.
Recently, we successfully generated genetically modified (GM) rats deficient in Cyp27b1 or Vdr genes (17). Type II rickets model rats with a mutant Vdr (R270L), which recognizes 1,25(OH) 2 D3 with an affinity equivalent to that of 25(OH)D3, were also generated. Although Cyp27b1-knockout (KO), Vdr-KO, and Vdr (R270L) rats exhibit rickets symptoms, they are significantly different to each other. Comparison among WT and three types of GM rats led to novel concepts regarding vitamin D actions, including genomic and nongenomic actions.
In this study, we generated Cyp24a1 KO rats and succeeded in measuring the time course of 25(OH)D3 plasma levels and its metabolites after the administration of 25(OH)D3. Furthermore, we revealed for the first time that not only CYP24A1, but also CYP3A1/2 is involved in the bioconversion of 23S,25,26(OH) 3 D3 to 25(OH)D3-26,23-lactone, which is known to have a biological activity to reduce serum Ca level (18).

Simultaneous quantification of 25(OH)D3 and its metabolites
Ion spectra involving DMEQ-TAD adducts of 25(OH)D3, 24,25(OH) 2 D3, and 25(OH)D3-26,23-lactone were similar to those shown in our previous study (11). The ion product spectra of DMEQ-TAD adducts of 24,25(OH) 2 (Fig. S1). The ion fragment at m/z 468 was selected for multiple reaction monitoring (MRM) analysis, as shown in Table S1, because of greater specificity and lower background than the fragment at m/z 247. Chromatograms for each metabolite are shown in Figure S1. Two peaks for each metabolite were detected, which represented the 6R-and 6Sisomers of the DMEQ-TAD adduct of each metabolite, as in the case of 25(OH)D3 reported in a previous study (20). The mixtures of each authentic standard could be quantitatively analyzed simultaneously using MRM scan mode.

Generation of CYP24a1 KO rats
Twenty-six offspring were obtained, one of which had a 1 bp insertion at a target site. Consequently, 30 amino acid residues from 439 to 469 were substituted, resulting in the introduction of a premature stop codon at residue 469 (Fig. S2). This frame shift causes loss of the cysteine residue at position 462, which is the fifth ligand of heme iron and the active center, and causes a deletion involving the C-terminal portion of CYP24A1, which consists of 514 amino acids. Thus, this truncated CYP24A1 could not show enzymatic activity. Homozygotes for this mutant CYP24A1 allele were used in this study. We confirmed that no off-target site (OTS) events occurred in potential OTSs searched using the CRISPR Direct tool for CYP24A1 sgRNA (Fig. S3).
Appearance, growth, and bone-metabolism-related parameters in the blood of Cyp24a1 KO rats No significant difference was observed in the appearance and growth rate between Cyp24a1 KO and WT rats (Fig. 2). Femoral length in Cyp24a1 KO rats at 15 weeks of age was Study on metabolic pathway of 25(OH)D3 using Cyp24a1 KO rats similar to that of WT rats. CT scanning, von Kossa staining, and toluidine blue staining of femora revealed no noticeable differences between Cyp24a1 KO and WT rats (Fig. S4).
St-Arnaud (21) demonstrated that 50% of Cyp24a1 KO mice (homozygous mutant) died before 3 weeks of age, but no such situation was observed with Cyp24a1 KO rats. μCT scans showed no particular symptoms such as soft tissue calcification (data not shown). Neither plasma Ca and P levels nor parathyroid hormone (PTH) levels were significantly different between Cyp24a1 KO and WT rats ( Table 1).
In the next step, 25(OH)D3 was orally administered to WT or Cyp24a1 KO rats at 200 μg/kg body weight, and plasma concentrations of 25(OH)D3 and its metabolites were examined periodically over 0-48 h. After 25(OH)D3 administration, plasma 25(OH)D3 concentrations in Cyp24a1 KO rats were markedly elevated than in WT animals (Figs. 3 and 4). The maximum plasma concentration (C max ) of 25(OH)D3 in Cyp24a1 KO rats was four times higher than that in WT Each value represents the mean ± SD (n = 5-8; biological replicates).  Study on metabolic pathway of 25(OH)D3 using Cyp24a1 KO rats rats (Table 3 and Fig. 4).  Table 3). Previous studies demonstrated that 25(OH)D3-26,23lactone is one of the major metabolites in rats and mice upon 25(OH)D3 administration (9, 10), which agrees with our present results. The T max values of 25(OH)D3-26,23-lactone compared with those of other metabolites, with no detection of 25(OH)D3-26,23-lactone in our in vitro system containing recombinant rat CYP24A1 (Fig. 1), strongly suggested the likely involvement of enzyme(s) other than CYP24A1 in the formation of 25(OH)D3-26,23-lactone.

23S,25D3 administration
When 25(OH)D3 and ketoconazole to WT rats were simultaneously administered, plasma 25(OH)D3-26,23-lactone level was reduced to 50% compared with that without ketoconazole (Fig. S5, A and D). When 23S,25R,26(OH)3D3 and ketoconazole were simultaneously administered to WT and CYP24A1 rats, 25(OH)D3-26,23-lactone levels were reduced to 44% and 27% at 6 h after administration, respectively, compared with those without ketoconazole (Fig. S5, B, C, E, and F). Because the lactone formation in CYP24A1 KO rats appears to depend on CYP3A1/2, it is possible that the activity of CYP3A1/2 was decreased to 27% by the administered ketoconazole. Based on these results, and when the contribution of CYP3A1/2 and CYP24A1 is X and Y, respectively, the following equations could be represented.
X + Y = 1, and 0.27 × X + Y = 0.44 Thus, X and Y are calculated to be 0.77 and 0.23, respectively. These results indicate that the contribution ratio of CYP3A1/2 is 77%, and that of CYP24A1 is 23%.
Effects of daily administration of 25(OH)D3 to Cyp24a1 KO rats 25(OH)D3 was also administered daily to Cyp24a1 KO rats to clarify the effects of long-term administration. As described in "Materials and Methods," Cyp24a1 KO rats were fed normal chow before 9 weeks of age and were then converted to 25(OH) D3-containing chow at a dose of approximately 110 μg/kg body weight/day. Significant weight loss was observed compared with rodents on a normal diet (Fig. 10A). Food consumption was also markedly reduced, with initial food consumption of 73.0 ± 12.7 g/kg bw/day (n = 5), which was approximately halved to 36.5 ± 14.6 g/kg bw/day after 6 weeks of 25(OH)D3 treatment. Furthermore, ectopic calcification was observed in Cyp24a1 KO rats receiving 25(OH)D3-containing chow, based on μCT images. Figure 10B illustrates the calcification of the aorta in the chest. On the contrary, when WT rats were fed the same 25(OH)D3-containing diet for 6 weeks, body weight increased as anticipated, with no reduction in food consumption or noticeable phenotypic changes.
Plasma Ca concentrations in the Cyp24a1 KO rats receiving 25(OH)D3-containing chow for 0, 1, and 2 weeks were 8.8 ± 1.4, 10.1 ± 1.3, and 11.0 ± 0.9 mg/dl, respectively. Thus, plasma Ca levels increased after daily administration of 25(OH)D3, but remained at approximately 11 mg/dl thereafter. Some Cyp24a1 KO rats died after daily 25(OH)D3 administration for 14--17 weeks. Although this study was conducted at a dose of only approximately 110 μg/kg body weight/day, the effects of longterm administration of 25(OH)D3 on CYP24A1 dysfunction should be investigated in greater detail in the future.

Discussion
In in vitro systems containing recombinant rat CYP24A1 expression system, 25(OH)D3 metabolites in the C-24 pathway were predominant compared with those in the C-23 pathway, and metabolites in the C-23 pathway, except for 23S,25(OH) 2 D3, were not detected. However, the end product of the C-23 pathway, 25(OH)D3-26,23-lactone, has been detected as a major metabolite in the plasma of WT rats  Study on metabolic pathway of 25(OH)D3 using Cyp24a1 KO rats administered with 25(OH)D3. Although this metabolite has also been reported to be present in humans at elevated 25(OH) D3 levels, the plasma 25(OH)D3-26,23-lactone level was lower than that in mice (16). This appeared to be inconsistent with the fact that compared with rat and mouse CYP24A1, human CYP24A1 displayed higher metabolism in the C-23 pathway (5). Thus, these discrepancies in in vitro and in vivo 25(OH) D3-26,23-lactone formation between rats, mice, and humans still remain unexplained.
In this study, we revealed that CYP3As are involved in the conversion of 23,25,26(OH) 3 D3 into 25(OH)D3-26,23-lactone, whereas CYP24A1 is responsible for the conversion of 25(OH) D3 into 23,25,26(OH) 3 D3 via 23,25(OH) 2 D3 (Fig. 7). Comparison of the amount of 25(OH)D3-26,23-lactone produced by administration of 23S,25R,26(OH) 3 D3 between WT and Cyp24a1 KO rats suggests that the contribution of CYP3A1/2 to this reaction could be larger than that of CYP24A1 in WT rats. Enzymatic studies using recombinant human CYP species and the inhibitory effects of ketoconazole strongly suggest that CYP3A plays an essential role in the conversion of 23,25,26(OH) 3 D3 into 25(OH)D3-26,23-lactone in both rats and humans. It is noted that human CYP3A4 is the most important CYP species among human drug-metabolizing P450 species belonging to cytochrome P450 families 1, 2, and 3. Although CYP3A4 catalyzes 4β, 23R, and 24S-hydroxylation of 25(OH)D3 and 1,25(OH) 2 D3, it has not been given much importance with regard to vitamin D metabolism. However, we revealed another role of CYP3A4 in vitamin D metabolism in this study. Thus, from the viewpoint of drug/drug interaction, CYP3A4-dependent drug/vitamin D interaction may be important. Recently, novel rickets type III has been discovered, We conducted a metabolic study using Cyp24a1 KO rats because the body size of these rats was about ten times larger than that of mice, and enough blood could be collected periodically over 48 h. This allowed a detailed analysis of the pharmacokinetics of the various metabolites, as shown in Figure 8 and Table 3. Such analyses appear to be difficult in mice, even if the analyses are sensitive. It is noted that comparisons between WT and Cyp24a1 KO rats have resolved previous inexplicable issues on 23,26-lactone-25(OH)D3 formation in WT rats.
Recently, SULT2A1 and UGT1A4 have been reported as metabolic enzymes of 25(OH)D3 and 1,25(OH)D3, respectively, in addition to CYP3A (27,28). In particular, the plasma concentration of sulfate conjugates formed by SULT2A1 was high. The plasma concentrations of 25(OH)D3, 25(OH)D3-3-O-sulfate, 24R,25(OH) 2 D3, and 4β,25(OH) 2 D3 in healthy humans were 131, 96, 12, and 0.26 nM, respectively (27). Furthermore, some studies demonstrated that the increased expression of VDR induced the transcription of CYP3A (29) and SULT2A1 (30). Based on these results, it is likely that CYP3A and SULT2A1 play critical roles in vitamin D metabolism, especially in cases of CYP24A1 dysfunction. The plasma concentration of sulfate conjugates was not examined in this study, but we will investigate it in more detail in the future.
In humans, CYP24A1 dysfunction sometimes causes elevated plasma levels of 1,25(OH) 2 D3 and is associated with idiopathic infantile hypercalcemia or kidney stones. However, Pronicka et al. (31) reported that the plasma 1,25(OH) 2 D3 levels are almost normal in many cases of human CYP24A1 defects. In this study, Cyp24a1 KO rats showed normal concentrations of 1,25(OH) 2 D3, and its concentration was slightly lower than those in WT rats. It has been reported that 1,25(OH) 2 D3 concentration in the plasma of Cyp24a1 KO mice is slightly lower than that of Cyp24a1-hetero mice (32), which is consistent with our present study. In the case of mice, it has been reported that 50% of Cyp24a1 KO homogeneous pups with Cyp24a1 KO parents died before the age of 3 weeks and displayed impaired bone mineralization (32). Since Cyp24a1 and Vdr doubleknockout mice normalized bone formation without impairing bone mineralization, St-Arnaud et al. (32) speculated that a rapid increase in plasma 1,25(OH) 2 D3 during late gestation in female Cyp24a1 KO mice, resulting in hypercalcemia, affects calcification during development. In the present study, however, Cyp24a1 KO rats showed no abnormal phenotype and pups that had Cyp24a1 KO parents grew normally. At present, the reason Study on metabolic pathway of 25(OH)D3 using Cyp24a1 KO rats for this difference between mice and rats is unclear. Although single 25(OH)D3 administration to Cyp24a1 KO rats did not cause toxic effects, long-term administration of 25(OH)D3 to the Cyp24a1 KO rats at 110 μg/kg b.w./day showed significant weight loss and ectopic calcification. In contrast, no side effects were observed in WT rats under the same conditions (Fig. 10). Thus, Cyp24a1 KO rats treated with vitamin D, 25(OH)D3, or 1,25(OH) 2 D3 could be used as models of renal diseases originating from CYP24A1 dysfunction, although further studies are needed.
In this study, we performed detailed metabolic analysis of 25(OH)D3 in vivo by using Cyp24a1 KO rats generated by genome editing. We conducted a metabolic analysis of natural 25(OH)D3 and concluded that the Cyp24a1 KO rats could be useful in developing various vitamin D derivatives as pharmaceuticals in the future.

Animals and diets
Jcl:Wistar rats were obtained from CLEA Japan Inc. Embryonic microinjection for genome editing was performed by KAC Co, Ltd. The generated Cyp24a1 KO rats were kept at room temperature (22-26 C) and in 50-55% humidity with a 12 h light/dark cycle. They were allowed food and water ad libitum and fed F-2 formula diet (Oriental Yeast Co) containing 0.74 w/w% calcium and 2000 IU vitamin D3/kg diet.
All experimental protocols using animals were performed in accordance with the Guidelines for Animal Experiments at Toyama Prefectural University and were approved by the Animal Research and Ethics Committee of Toyama Prefectural University.

Generation of Cyp24a1 KO rats and validation of off-target
Cyp24a1 KO rats were generated by CRISPR/Cas9 genome editing system as well as our previous study (17). The target site was selected to delete the cysteine at position 462 in Exon 10, which is the fifth ligand of heme iron and an active center of CYP24A1 (Fig. S2).
Potential off-target sites (OTS) of CYP24A1 in rat genomes were searched by the CRISPR Direct tool (http://crispr.dbcls. jp/) to valid the off-target events. The OTS region was amplified, purified, and then directly sequenced (Fig. S3).

Founder generation
Pups delivered from the transferred embryos were obtained by caesarean section on the day of birth and nursed by foster mothers. After weaning, tail tips were biopsied for mutation analysis. Crude genomic DNA was extracted from the tail tips by Lyppo (Wako Pure Chemicals). PCR was performed with KOD Fx Neo (TOYOBO) using the set of primer 5 0 -ATGAGGA-GAATCAGTGGTCCTCCTGGCTGCC-3 0 , 5 0 -AGGACTTT-TACCCAGCAGAGAGCCAGGTGG-3 0 . PCR products were purified using FastGene Gel/PCR Extraction Kit (Nippon Gene) according to the manufacture's protocol and directly sequenced and analyzed on an Applied Biosystems 3500 DNA sequencer (Thermo Fisher Scientific) using the BigDye Terminator v3.1 Cycle Sequencing Kit (Thermo Fisher Scientific).

Measurement of plasma Ca, phosphorus (P), and parathyroid hormone (PTH) concentrations
The plasma Ca and P concentrations were measured using the Calcium E-Test Wako and Phospha C-Test Wako (FUJI-FILM Wako Pure Chemical Corporation), respectively. The Study on metabolic pathway of 25(OH)D3 using Cyp24a1 KO rats plasma PTH concentration was determined using the Rat Intact PTH ELISA Kit (Immutopics Inc).

Phenotype of femura
Micro CT-scans analysis of femura was performed using an X-ray CT system (Latheta LCT-200; Hitachi Aloka Medical). Parameters used for the CT scans were same as described in our previous study (17). Von Kossa staining and Toluidine Blue staining were performed as described in our previous study (17).
In order to investigate the effect of CYP3A inhibition in vivo, 30 mg/kg-body weight of ketoconazole was coadministrated with each of 25(OH)D3 or 23S,25R,26(OH) 3 D3, and the blood samples were collected at 6 h after administration.
Daily administration of 25(OH)D3 to Cyp24a1 KO rats 25(OH)D3 was also daily dietary administrated to Cyp24a1 KO rats to clarify effects of long-term administration. 25(OH) D3 containing food was prepared as described in our previous study (17). Briefly, pellet diet containing 1.5 mg 25(OH)D3 per 1 kg F-2 normal diet was prepared by Oriental Yeast Co. Normal F-2 diet was fed prior to 9-week age, and after that 25(OH)D3-containing diet was fed. The average daily food intake was 20.0 ± 5.0 g in the beginning of dietary 25(OH)D3 administration, and the food intake per kg body weight was 73.0 ± 12.7 g/kg bw/day (N = 5). Thus, the dose of the 25(OH) D3 was calculated to be 110 ± 19 μg/kg bw/day in the beginning of dietary 25(OH)D3 administration.

D3 in each fraction
The liver mitochondrial and microsomal fractions were prepared from WT and Cyp24a1 KO rats, and the productivity of 25(OH)D3-23,26-lactone in each mitochondrial and microsomal fraction was analyzed by modified methods described in the previous study (11). The mitochondrial fraction was sonicated on ice for 1 min in 10 s bursts with 1 min of cooling. It was then incubated in 100 mM Tris-HCl buffer (pH 7.4) containing 5000 nM ADX, 500 nM ADR, 10 μM 23S,25R,26(OH) 3 D3, and 1 mM NADPH or NADH at 37 C for 30 min. The microsomal fraction was incubated in 100 mM phosphate buffer (pH 7.4), 10 μM 23S,25(OH) 2 D3 or 23S,25R,26(OH) 3 D3, and 1 mM NADPH or NADH at 37 C for 30 min. The substrate and its metabolites were extracted and analyzed as described in the section, Measurement of plasma 25(OH)D3 and its metabolites concentration by LC/MS/MS analysis.

Statistical analysis
Student's t-test was used to identify significant differences between WT and Cyp24a1 KO rats (p < 0.05).

Data availability
All data generated or analyzed in the present study are included in this article or the data repositories listed in References.