Synthesis of Deuterium-Labeled Vitamin D Metabolites as Internal Standards for LC-MS Analysis

Blood levels of the vitamin D3 (D3) metabolites 25-hydroxyvitamin D3 (25(OH)D3), 24R,25-dihydroxyvitamin D3, and 1α,25-dihydroxyvitamin D3 (1,25(OH)2D3) are recognized indicators for the diagnosis of bone metabolism-related diseases, D3 deficiency-related diseases, and hypercalcemia, and are generally measured by liquid-chromatography tandem mass spectrometry (LC-MS/MS) using an isotope dilution method. However, other D3 metabolites, such as 20-hydroxyvitamin D3 and lactone D3, also show interesting biological activities and stable isotope-labeled derivatives are required for LC-MS/MS analysis of their concentrations in serum. Here, we describe a versatile synthesis of deuterium-labeled D3 metabolites using A-ring synthons containing three deuterium atoms. Deuterium-labeled 25(OH)D3 (2), 25(OH)D3-23,26-lactone (6), and 1,25(OH)2D3-23,26-lactone (7) were synthesized, and successfully applied as internal standards for the measurement of these compounds in pooled human serum. This is the first quantification of 1,25(OH)2D3-23,26-lactone (7) in human serum.

The concentrations of metabolites 2 and 3 in blood have been measured for clinical purposes by radioimmunoassay (RIA) or chemiluminescent enzyme immunoassay (CLEIA) [16]. They, however, have disadvantages such as the need to handle radioactive materials, and insufficient discrimination of vitamin D metabolites by antibodies [17]. More recently, a liquid-chromatography tandem mass spectrometry (LC-MS/MS) method has been developed to determine the concentration of multiple vitamin D metabolites simultaneously in blood [18]. However, LC-MS/MS-based measurement also has some problems, such as the low ionization efficiency of vitamin D derivatives and interference by contaminants including multiple D3 metabolites in the blood. To address these issues, several approaches have been investigated. Cookson-type reagents have been developed to improve the ionization efficiency of D3 metabolites, affording high sensitivity even at low abundance [19,20]. The isotope dilution method has also been applied to avoid interference from contaminants in the blood. This method requires a stable isotope-labeled compound as an internal standard, and so far, deuterium-labeled 25(OH)2D3 (2), 1,25(OH)2D3 (3), and 24R,25(OH)2D3 (4), in which deuterium is introduced at C26, C27, C6, and C19, have been synthesized ( Figure 2) [21][22][23][24][25]. The concentrations of metabolites 2 and 3 in blood have been measured for clinical purposes by radioimmunoassay (RIA) or chemiluminescent enzyme immunoassay (CLEIA) [16]. They, however, have disadvantages such as the need to handle radioactive materials, and insufficient discrimination of vitamin D metabolites by antibodies [17]. More recently, a liquid-chromatography tandem mass spectrometry (LC-MS/MS) method has been developed to determine the concentration of multiple vitamin D metabolites simultaneously in blood [18]. However, LC-MS/MS-based measurement also has some problems, such as the low ionization efficiency of vitamin D derivatives and interference by contaminants including multiple D 3 metabolites in the blood. To address these issues, several approaches have been investigated. Cookson-type reagents have been developed to improve the ionization efficiency of D 3 metabolites, affording high sensitivity even at low abundance [19,20]. The isotope dilution method has also been applied to avoid interference from contaminants in the blood. This method requires a stable isotope-labeled compound as an internal standard, and so far, deuterium-labeled 25(OH) 2 D 3 (2), 1,25(OH) 2 D 3 (3), and 24R,25(OH) 2 D 3 (4), in which deuterium is introduced at C26, C27, C6, and C19, have been synthesized ( Figure 2) [21][22][23][24][25].
In the synthesis of the deuterium-labeled metabolites 2-4-d 6 , deuterium was introduced into the side chain at C26 and C27 by reacting esters 8 with deuterated Grignard reagent, CD 3 MgBr ( Figure 2B) [21,22]. On the other hand, 2-3-d 3 were synthesized by reacting SO 2 adducts of cyclic compounds 9 derived from D 3 with deuterium oxide (D 2 O) [23][24][25]. In both strategies, the range of metabolites that can be synthesized is limited due to the restrictions imposed by the use of steroid precursors. Therefore, a more versatile approach is required. Convergent strategies, with coupling between CD-ring and A-ring moieties, have been widely applied for the synthesis of D 3 derivatives [26,27]. Since the CDring structures of the metabolites are diverse, whereas the A-ring structures are relatively constant, we considered that deuterium-labeled A-ring synthons would be suitable for the preparation of a variety of deuterium-labeled D 3 metabolites ( Figure 2D). In addition, labeling in the A-ring has an advantage in metabolism studies because the side chains of the D3 are well known to be enzymatically metabolized easily. In this study, we have developed a synthesis of deuterium-labeled A-ring precursors 13-d 3 and 16-d 3 incorporating three deuterium atoms. These precursors were coupled with CD-ring moieties 17 and 18 to afford deuterium-labeled 25(OH)D 3 -d 3 (2-d 3 ) and vitamin D lactones 25(OH)D 3 -23,26-lactone-d 3 (6-d 3 ) and 1,25(OH) 2 D 3 -23,26-lactone-d 3 (7-d 3 ). We also confirmed that the concentrations of 2, 6, and 7 in human serum could be measured by LC-MS/MS using the corresponding deuterium-labeled compounds as the internal standards (IS) (see Supplementary Materials). In the synthesis of the deuterium-labeled metabolites 2−4-d6, deuterium was introduced into the side chain at C26 and C27 by reacting esters 8 with deuterated Grignard reagent, CD3MgBr ( Figure 2B) [21,22]. On the other hand, 2−3-d3 were synthesized by reacting SO2 adducts of cyclic compounds 9 derived from D3 with deuterium oxide (D2O) [23][24][25]. In both strategies, the range of metabolites that can be synthesized is limited due to the restrictions imposed by the use of steroid precursors. Therefore, a more versatile approach is required. Convergent strategies, with coupling between CD-ring and A-ring moieties, have been widely applied for the synthesis of D3 derivatives [26,27]. Since the CD-ring structures of the metabolites are diverse, whereas the A-ring structures are relatively constant, we considered that deuterium-labeled A-ring synthons would be suitable for the preparation of a variety of deuterium-labeled D3 metabolites ( Figure 2D). In addition, labeling in the A-ring has an advantage in metabolism studies because the side chains of the D3 are well known to be enzymatically metabolized easily. In this study, we have developed a synthesis of deuterium-labeled A-ring precursors 13-d3 and 16-d3 incorporating three deuterium atoms. These precursors were coupled with CD-ring moieties 17 and 18 to afford deuterium-labeled 25(OH)D3-d3 (2-d3) and vitamin D lactones 25(OH)D3-23,26lactone-d3 (6-d3) and 1,25(OH)2D3-23,26-lactone-d3 (7-d3). We also confirmed that the concentrations of 2, 6, and 7 in human serum could be measured by LC-MS/MS using the corresponding deuterium-labeled compounds as the internal standards (IS).

Results
We employed a convergent strategy using the palladium-catalyzed coupling reaction of enyne-type deuterium-labeled A-ring precursors 13-d 3 and 16-d 3 with bromoolefins 17 and 18 as the CD-ring moieties. The deuterium atoms in 13-d 3 and 16-d 3 were introduced by the H/D exchange at the a-position of the alcohol, as reported by Sajiki et al. [28]. Our synthesis of deuterium-labeled enyne 13 commenced with the H/D exchange reaction of alcohol 10, which was obtained from L-(-)-malic acid in 4 steps (Scheme 1) [29]. The alcohol 10 was subjected to the H/D exchange reaction with a catalytic amount of Ru/C in D 2 O at 80 • C under an H 2 atmosphere to afford 10-d 3 deuterium-labeled at C3 and C4 in a 96% yield with over 93% deuteride content [28]. In this reaction, the stereochemistry at C3 was isomerized (4:1 ratio of α-10a and β-10b). The deuterium-labeled alcohol 10 (enantiomeric mixture) was converted into alkyne 11 by tosylation of the primary alcohol followed by epoxidation with NaH and reaction with TMS-acetylene (22% yield from 10-d 3 ). The hydroxyl group in alkyne 11 was protected with TBS ether, followed by deprotection of the TMS and pivaloyl groups with NaOMe in MeOH to give the alcohol 12 in an 86% yield from 11. Enyne 13 was obtained in a 51% yield from 12 via 4 steps, (i) tosylation of the primary alcohol; (ii) cyanation with NaCN; (iii) reduction of the nitrile group with DIBAL-H to aldehyde; and (iv) a Wittig reaction with Ph 3 PCH 3 I and NaHMDS. It was confirmed by 1 H-NMR that the deuteration rate did not decrease in these reaction steps [30]. droxyl group in alkyne 11 was protected with TBS ether, followed by deprotection of the TMS and pivaloyl groups with NaOMe in MeOH to give the alcohol 12 in an 86% yield from 11. Enyne 13 was obtained in a 51% yield from 12 via 4 steps, (i) tosylation of the primary alcohol; (ii) cyanation with NaCN; (iii) reduction of the nitrile group with DIBAL H to aldehyde; and (iv) a Wittig reaction with Ph3PCH3I and NaHMDS. It was confirmed by 1 H-NMR that the deuteration rate did not decrease in these reaction steps [30]. Scheme 1. Synthesis of deuterium-labeled enyne 13-d3.

Derivatization of 2, 6, 7 for LC-MS/MS, and Preparation of Calibration Curves
With the deuterium-labeled D3 metabolites of 2-d3, 6-d3, and 7-d3 in hand, we next examined the quantitative analysis of the three D3 metabolites in pooled human serum by LC-MS/MS. First, we confirmed that our deuterium-labeled D3 metabolites were suitable as the internal standards for the isotope dilution method in an LC-MS/MS analysis. As described above, D3 and its metabolites have low ionization efficiency in an LC-MS/MS, and derivatization is necessary to improve the ionization efficiency. Thus, the D3 metabolites 2, 6, and 7, as well as 2-d3, 6-d3, and 7-d3, were derivatized with a recently developed reagent DAP-PA (4-(4′-dimethylaminophenyl)-1,2,4-triazoline-3,5-dione-phenyl anthracene) [20], and the ion peaks of the DAP adducts were detected by selective reaction monitoring (SRM) under the LC-MS/MS conditions shown in Table 1 (Figure 3).

Derivatization of 2, 6, 7 for LC-MS/MS, and Preparation of Calibration Curves
With the deuterium-labeled D 3 metabolites of 2-d 3 , 6-d 3 , and 7-d 3 in hand, we next examined the quantitative analysis of the three D 3 metabolites in pooled human serum by LC-MS/MS. First, we confirmed that our deuterium-labeled D 3 metabolites were suitable as the internal standards for the isotope dilution method in an LC-MS/MS analysis. As described above, D 3 and its metabolites have low ionization efficiency in an LC-MS/MS, and derivatization is necessary to improve the ionization efficiency. Thus, the D 3 metabolites 2, 6, and 7, as well as 2-d 3 , 6-d 3 , and 7-d 3 , were derivatized with a recently developed reagent DAP-PA (4-(4 -dimethylaminophenyl)-1,2,4-triazoline-3,5-dione-phenyl anthracene) [20], and the ion peaks of the DAP adducts were detected by selective reaction monitoring (SRM) under the LC-MS/MS conditions shown in Table 1 (Figure 3).  In the case of the DAP-adducts of 2 and 2-d3 ( Figure 3A,B), we observed identical ion peaks at the retention time of 5.60 min (abbreviated as tR: 5.60 min). Similarly, 6 and 6-d3 showed the same tR of 3.50 min, and 7 and 7-d3 showed the same tR of 2.15 min, indicating that the deuterium-labeled compounds are suitable as internal standards for the isotope dilution method. We also observed small peaks at the retention times of 5.20 min ( Figure  3A,B), 2.65 min ( Figure 3C,D), and 2.37 min (Figure 3E,F) for 2/2-d3, 6/6-d3, and 7/7-d3, respectively. These peaks are due to the epimers at C6 of the DAP adducts, because DAP-PA reacts from both the α-and β-faces.
Next, the calibration curves were prepared as follows (Figure 4). A total of 100 μL of In the case of the DAP-adducts of 2 and 2-d 3 ( Figure 3A,B), we observed identical ion peaks at the retention time of 5.60 min (abbreviated as t R : 5.60 min). Similarly, 6 and 6-d 3 showed the same t R of 3.50 min, and 7 and 7-d 3 showed the same t R of 2.15 min, indicating that the deuterium-labeled compounds are suitable as internal standards for the Molecules 2022, 27, 2427 6 of 8 isotope dilution method. We also observed small peaks at the retention times of 5.20 min ( Figure 3A,B), 2.65 min ( Figure 3C,D), and 2.37 min (Figure 3E,F) for 2/2-d 3 , 6/6-d 3 , and 7/7-d 3 , respectively. These peaks are due to the epimers at C6 of the DAP adducts, because DAP-PA reacts from both the αand β-faces.
Next, the calibration curves were prepared as follows (Figure 4). A total of 100 µL of each one of the calibrator solutions was mixed with 200 µL of the internal standard solution and evaporated to dryness. After derivatization with DAP-PA, an LC-MS/MS analysis of the unlabeled and labeled DAP-adducts was performed, and the calibration curves were prepared by plotting the concentration of unlabeled DAP-adduct against the ion peak area ratio of unlabeled versus labeled DAP-adduct. All of the calibration curves showed good linearity. In the case of the DAP-adducts of 2 and 2-d3 ( Figure 3A,B), we observed identical ion peaks at the retention time of 5.60 min (abbreviated as tR: 5.60 min). Similarly, 6 and 6-d3 showed the same tR of 3.50 min, and 7 and 7-d3 showed the same tR of 2.15 min, indicating that the deuterium-labeled compounds are suitable as internal standards for the isotope dilution method. We also observed small peaks at the retention times of 5.20 min ( Figure  3A,B), 2.65 min ( Figure 3C,D), and 2.37 min ( Figure 3E,F) for 2/2-d3, 6/6-d3, and 7/7-d3, respectively. These peaks are due to the epimers at C6 of the DAP adducts, because DAP-PA reacts from both the α-and β-faces.
Next, the calibration curves were prepared as follows (Figure 4). A total of 100 μL of each one of the calibrator solutions was mixed with 200 μL of the internal standard solution and evaporated to dryness. After derivatization with DAP-PA, an LC-MS/MS analysis of the unlabeled and labeled DAP-adducts was performed, and the calibration curves were prepared by plotting the concentration of unlabeled DAP-adduct against the ion peak area ratio of unlabeled versus labeled DAP-adduct. All of the calibration curves showed good linearity.

Quantification of the D3 Derivatives in Human Serum
The levels of 2, 6, and 7 in pooled human serum were quantified by the LC-MS/MS using the isotope dilution method with the constructed calibration curves. The serum was pretreated as follows. An aliquot of serum (100 μL) was mixed with the internal standards solution (200 μL). Each sample was loaded onto a supported liquid extraction column (ISOLUTE SLE+ 300 μL sample Volume, Biotage, Uppsala, Sweden) and eluted three times with 600 mL hexane/ethyl acetate (1/1, v/v) using a PRESSURE+48 positive pressure manifold (Biotage, Uppsala, Sweden). The combined eluates were evaporated to dryness

Quantification of the D 3 Derivatives in Human Serum
The levels of 2, 6, and 7 in pooled human serum were quantified by the LC-MS/MS using the isotope dilution method with the constructed calibration curves. The serum was pretreated as follows. An aliquot of serum (100 µL) was mixed with the internal standards solution (200 µL). Each sample was loaded onto a supported liquid extraction column (ISOLUTE SLE+ 300 µL sample Volume, Biotage, Uppsala, Sweden) and eluted three times with 600 mL hexane/ethyl acetate (1/1, v/v) using a PRESSURE+48 positive pressure manifold (Biotage, Uppsala, Sweden). The combined eluates were evaporated to dryness in a centrifugal evaporator. The ion peaks of the metabolites matched well with those of the corresponding internal standards in the pretreated samples ( Figure 5). in a centrifugal evaporator. The ion peaks of the metabolites matched well with those of the corresponding internal standards in the pretreated samples ( Figure 5). The concentrations of 2, 6, and 7 in human serum were calculated to be 5.1 ng/mL, 38.3 pg/mL, and 8.9 pg/mL, respectively, based on the area ratios of the detected peaks. The concentrations of 2 and 6 were in agreement with previously reported values [37], while this is the first quantification of 1α-lactone 7 in human serum.

Conclusions
We synthesized deuterium-labeled A-ring-d3 synthons 13 and 16 and utilized them for the convergent synthesis of deuterium-labeled D3 derivatives 25(OH)D3 (2), 25(OH)D3-23, 26-lactone (6), and 1,25(OH)2D3-23, 26-lactone (7). These deuterium-labeled D3 metabolites were successfully applied as internal standards for the quantification of the metabolites in pooled human serum by LC-MS/MS using the isotope dilution method. This is The concentrations of 2, 6, and 7 in human serum were calculated to be 5.1 ng/mL, 38.3 pg/mL, and 8.9 pg/mL, respectively, based on the area ratios of the detected peaks. The concentrations of 2 and 6 were in agreement with previously reported values [37], while this is the first quantification of 1α-lactone 7 in human serum.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/molecules27082427/s1, Experimental procedures for synthesis and characterization of compounds, Experimental procedure for LC-MS/MS analysis using the isotope dilution method, 1 H and 13 C NMR spectra.