24-Hydroxylation of 1,25-dihydroxyergocalciferol: An unambiguous deactivation process

THEJOURNAL OF BIOLOGICAL CHEMISTRY

Vol. 261,No. 20, Issue of July 15,pp. 9250-9256, 1986 Printed in U.S.A.

24-Hydroxylationof 1,25-Dihydroxyergocalciferol AN UNAMBIGUOUS DEACTIVATION PROCESS* (Received for publication, December 20, 1985)

Ronald L. Horst$, Timothy A. ReinhardtS, Charles F. Rambergg, Nick J. KoszewskiS, and Joseph L. Napolin From the $National Animal Disease Center, Agricultural Research Service, United States Department of Agriculture, Ames, Iowa 50010, the $University of Pennsylvania, Kennett Square, Pennsylvania 19348, and lTDepartment of Biochemistry, State University of New York-Buffalo,Buffalo, New York 14214

1,24,25-Trihydroxyergocalciferolwas isolated from nutritional evolution since many species evolved consuming bovine kidney homogenates incubated with 1,25-diplant tissue rich in ergocalciferol and/or receiving cholecalhydroxyergocalciferol and from chick kidney homog- ciferol following exposure of their skin to ultraviolet irradiaenates incubated with 24,25-dihydroxyergocalciferol. tion. Both cholecalciferol and ergocalciferol require C-25 hyThe identity was established by ultraviolet absorbance, droxylation as the initialstep in expression of biological sensitivity to periodate, nuclear magnetic resonance, activity. This initial step is followed by C-1 hydroxylation to and mass spectrometry. The new metabolite had an form the 1,25-dihydroxy metabolites which are nowwidely 1,24,25-trihydroxycholecalciferolfor considered as the most biologically active forms of either affinity equal to the bovine-thymus andchick-intestinal1,25-dihycholecalciferol or ergocalciferol. droxyvitamin D receptor and hadan affinity twice that The metabolite 1,25-dihydroxycholecalciferol(1,25-(OH)zof 1,24,25-trihydroxycholecalciferolfor the rat-intestinal receptor. It was 3- and 6-fold less competitive D3)l is known to undergo further side-chain oxidation at than either 1,25-dihydroxycholecalciferolor 1,24,25- C-23, C-24, and/or C-26 to form a number of metabolites in(6-9), trihydroxycholecalciferol, respectively,forthe rat cluding 1,24,25-(OH)3D3(3-51, 24-ket0-1,25-(OH)~D~ 24-ket0-1,23,25-(0H)~D~ (7, 9, lo), 1,25,26-(OH)3D3(11, 12), plasma vitamin D transport protein. 1,24,25-Trihydroxyergocalciferol was at least 10-fold less active 1,23,25-(OH)3D3(13, 14), and 1,25-(OH)zD3-26,23lactone (6, than 1,25-dihydroxycholecalciferol, 1,25-dihydrox- 7, 15). Similar oxidative steps have been shown to occur yergocalciferol, and 1,24,25-trihydroxycholecalci- during 25-OHD3 metabolism (16-18). The metabolic imporferol at stimulating intestinal-calcium transport and tance of these side-chain modifications is still controversial, was also relatively ineffective at stimulating bone- but is most likely related to deactivation and rapid clearance calcium resorptionin rats. Moreover, in rats, [‘HI of the 1,25-(OH)zD3steroid hormone, particularly in the case 1,24,25-trihydroxyergocalciferol wasclearedfrom of C-23 oxidation (13, 19). plasma approximately 40% faster than [3H]1,24,25Although C-24 oxidation of the 25-OHDz side chain has trihydroxycholecalciferol. These data suggest that C- been described (20), the metabolism of 1,25-(OH)zDz, an 24 hydroxylation of 1,25-dihydroxyergocalciferol equally active compound as 1,25-(OH)zD3in mammals (21), represents a significant in vivo deactivationstep, has not been investigated. Both quantitative and qualitative whereas equivalent deactivation of 1,25-dihydroxy- differences in metabolism between 1,25-(OH)2D3and 1,25cholecalciferol seems to involve metabolic steps sub- (OH)zDz are anticipated because of the structure of the ergosequent to C-24 hydroxylation (C-24 ketonization). C- calciferol side chain with its C-22-alkene and 24(S)-methyl 24 ketonization of 1,25-trihydroxyergocalciferol group. Although production of the metabolite 1,24,25-(OH)3D2 would not be anticipated due to the presence of the has yet to be demonstrated in uiuo, its formation would be 24(S)-methyl group. These results reveal further dis- anticipated. Once formed, 1,24,25-(OH)3Dz,unlike 1,24,25similarities between ergocalciferol andcholecalciferol metabolism in mammals and suggest a mechanism for (OH)&, wouldbe unlikely to undergo C-24 ketonization the lesser tendency of ergocalciferol to cause hyper- because the 24(S)-methylgroup would probably block further oxidation. Similarly, the compounds 1,23,25-(OH)& or 1,25calcemia relative tocholecalciferol. (OH)zDz-26,23lactone are not anticipated because the C-22 position of 1,25-(OH)zDzis unsaturated. These probable metabolic differences may have biological relevance because erCholecalciferol (vitamin D3) is the major naturally occur- gocalciferol is less likely to cause hypercalcemia than cholering form of vitamin D in animals, whereas ergocalciferol calciferol (22), and there is discrimination either for ergocal(vitamin Dz) is a major naturally occurring form of vitamin The abbreviations used are: 1,25-(OH)&, 1,25-dihydroxycholeD in plants. Endogenous synthesis of these sterols occurs following ultraviolet irradiation of their respective precursors, calciferol; 1,25-(OH)zDz, 1,25-dihydroxyergocalciferol; 1,24,25(OH)&, 1,24,25-trihydroxycholecalciferol;1,24,25-(OH)~D~, 1,24,25i.e. 7-dehydrocholesterol (precursor to cholecalciferol) which trihydroxyergocalciferol; 1,25,28-(OH)3D~, 1,25,28-trihydroxyergois present in the skin of animals (1)and ergosterol (precursor calciferol; 24-ket0-1,25-(OH)~D3, 24-keto-1,25-dihydroxycholecalto ergocalciferol) which is present in plant tissue (2). Both ciferol;1,23,25-(OH)sD3, 1,23,25-trihydroxycholecalciferol;1,25,26forms of vitamin D cantherefore be considered important in (OH)3D3, 1,25,26-trihydroxycholecalciferol;1,25-(OH)zD3-26,23-1ac* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

tone, 1,25-dihydroxycholecalciferol-26,23-lactone; 25-OHDz, 25-hydroxyvitamin Dz; 25-OHD3,25-hydroxyvitamin D3; HPLC, high performance liquid chromatography; NMR, nuclear magnetic resonance; ICT, intestinal calcium transport, BCR,bone calcium resorption; EI, electron impact.

9250

9251

24-Hydroxylationof 1,25-Dihydroxyergocalciferol ciferol or cholecalciferol, depending upon the species-dependent preference (23, 24). Both phenomena could be related to differences in the nature and/or rate of side-chain metabolism. During the course of studying 1,25-(OH)& metabolism by bovine kidney homogenates, we isolated a quantitatively major metabolite. This paper will report the structure of the metabolite as 1,24,25-(OH)3D2and will demonstrate that C 24 hydroxylation targets 1,25-(OH)& to deactivation. Metabolic steps subsequent to (2-24 hydroxylation seem to be required for equivalent deactivation of 1,25-(OH)2D3.

1 -HYDROXYLATION

i

24.25-(0E),

+

84-HYDROXYLATION

I

I 'E1Dp

MATERIALS ANDMETHODS

High-performance liquid chromatography was performed with Waters Model ALC/GPC 204 liquid chromatography equipment (Waters Associates, Milford, MA)? Calciferols were detected at 254 nm. HPLC columns were purchased from DuPont Instruments(Wilmington, DE), unlessotherwise noted. Zorbax-Si1 refers to microparticulate silica gel columns. Zorbax-NHZ refers to the 3-aminopropyltriethoxy derivative of the microparticulate silica gel columns. Whatman ODS refers to the octadecasilane derivative of the microparticulate silica gel columns. HPLC solvents were purchased from Burdick and Jackson Laboratories (Muskegon, MI) or Fisher. Ultraviolet spectra were determined in ethanol with a Beckman Model 25 recording spectrophotometer. Molar absorptivities of 18,200 and 19,400 M" cm" were used for cholecalciferol and ergocalciferol metabolites, respectively. Mass spectra were obtained at 70 eV using the solid probe of a Finnigan Model 4021 EI/CI gas chromatography/mass spectrometry system coupled with an INCOS 2000 data system. To obtain spectra, the probe was heated from ambient to 320 "C over a 10-min period at anionizer temperature of 250 "C.Chloride addition, negative ion, chemical ionization (Cl-NCI) mass spectra were obtained with dichlorodifluoromethane as reagent gas. Proton NMR were taken in CDC1, with a 300 MHz Nicolet Fourier transform NMRspectrometer. 1,25-Dihydroxyvitamin Dz, (24R)1,24,25-(OH),D, (253)1,25,26(OH)&?, and1,25,28-(OH)3Dzwere gifts from Dr. M. R. Uskokovic of Hoffmann-La Roche. The 24,25-(OH)zD2and 24,25-(OH)3[3H]D2 (2.3 Ci/mmol) were prepared in vivo as previously described (23). The (90 Ci/mmol) 1,25-(OH)z[3H]DZ(90 Ci/mmol) and 1,25-(0H)~[~H]D3 were synthesized by allowing [3H]methyl magnesium bromide to react with l-hydroxy-27-nor-25-keto-vitamin D2 and 1-hydroxy-27-nor-25keto-vitamin D3, respectively. Their purity and biological activity were verified by HPLC and by their binding propertiesin 1,25(OH),D receptor binding assays (25, 26). 24-Ket0-1,25-(OH)~D,was synthesized and purified as previously described {6). Production and Purification of 1,24,25-(oH)& and 1,24,25(OH)3~H/D2-Inorder to substantiate the structural assignment, 1,24,25-(OH)3Dzwas prepared by two separate procedures using different starting materials. A diagram of the production methods is shown in Fig. 1. Preparation of kidney homogenates and reaction mixtures in all cases was as described by Engstrom et al.(27). Fractions 1and 2 were prepared by incubating 24,25-(OH)~Dz(two flasks, 10 pg each) and 24,25-(OH)2[3H]D2(one flask, 1 pCi) for 60 min with homogenates prepared from kidneys of vitamin D-deficient chicks. Vitamin D deficiency stimulates kidney 1-hydroxylase in chicks (27, 28). Fractions 3 and 4 were prepared by incubating 1,25-(OH)& (80 flasks, 15 pg each) and 1,25-(OH)2[3H]Dz(two flasks, 500 pCi each) with homogenates prepared from kidneys of Jersey bull calves (6-8 weeks pg at 24 and 4 h prior to sacrifice). old) treated with 1,25-(OH)2D2 (100 This treatment regime stimulates the kidney 24-hydroxylase (27). Lipids were extracted in all samples using the methylene chloride/ methanol procedure of Horst et d.(29). Sephadex LH-20 chromatography for fractions 1 and 2 was achieved by applying the extracted lipids to a column (0.6 X 15.5 cm) developed in 0.015:1:1:9 water/ chloroform/methanol/hexane (mobile phase 1). The first 8 ml were discarded and the solvent changed to 0.015:2:2:8 water/chloroform/ methanol/hexane (mobile phase 2). Fifteen ml of mobile phase 2 were added to elute the calciferols. Fractions 3 and 4 were purified in a ~

~~

Mention of a trade name, proprietary product, or vendors does not constitute a guarantee or warrantyby the United States Department of Agriculture and does not imply its approval to theexclusion of other products orvendors that may be suitable.

1

1.24.25-(OH)s Dg

1,24,25-(Olib Dz

4

.1,24,25-(OH)gl%lDz l o r Racer kinetics

FIG. 1. Flow diagram for the procedures used in preparing radioinert and radioactive1,24,25-(OH)~D2.The 1,24,25(OH)~DZ was synthesized either by the 1-hydroxylation of 24,25(OH)zD2 (fraction 1) or 24-hydroxylation of 1,25-(OH)2Dz(fraction 3). Fraction 2,which contained radiolabeled 24,25-(OH)zDz and 1,24,25-(OH)Dz, was added to each fraction 1and fraction 3 prior to HPLC.

2.0

0.I

8

0.I

0

0

20 ml

FIG. 2. HPLC profile of the metabolitesproduced by chick kidneyhomogenatesincubatedwith 24,25-(OH)aDa(Panel A ) and from calf kidney homogenates incubated with 1,25(OH)2D1(Panel B ) . HPLC was achieved using a Zorbax-Si1 (0.9 X 25 cm) column developed in isopropanol/methylene chloride/hexane (52570). similar manner except that 18 ml of mobile phase 1 were added to each column to elute the starting material followed by 15 mlof solvent system 2 to elute the reaction products. Prior to HPLC, fraction 2 was divided equally between fractions 1 and 3 to allow for radioisotope monitoring of products. Fractions 1 and 3 were applied separately to a semipreparative Zorbax-Si1column (0.9 X 25 cm) developed in 2-propanol/methylene chloride/hexane (5:25:70). Each 1,24,25-(OH)3Dzpeak (Fig. 2 A , and B ) was collected separately, the solvent was evaporated, and the residue was resuspended in 2-propanol/hexane (1090) for further purification on an analytical (0.25 X 25 cm) Zorbax-Si1 column

9252

24-Hydroxylutionof 1,25-Dihydroxyergocakiferol

developed in the same mobile phase. The major peak from each (3438 ml) was collected and eluted through a Zorbax-NHZcolumn (0.25 X 25 cm) developed in 2-propanol/hexane (14:W). The major peak (35-40 ml) was recycled once (for a total of two passes across the column) and collected, the solvent was evaporated, and themetabolite was dissolved in ethanolfor spectral and chemical analyses. The [3H] trihydroxy-l,25-(0H)~D~ metabolites of fraction 4 were purified in the samemanner, except fraction 2 was not added. The 1,24,25(OH)3[3H]Dzpurified from fraction 4 was used in plasma tracer kinetic studies. Periodate Cbauage-One pg of 1,24,25-(OH)3Dz,purified from bovine kidney homogenates after incubation of 1,25-(OH)&, was dissolved in 20 pl of methanol. Fifteen pl of 5% sodium metaperiodate in water were added, and the mixture was incubated at 45 “C. After 1.5 h, the solvent was evaporated under Nz,and the residue was dissolved in 2-propanol/hexane (1090). The 1,24,25-(OH)3Dzwas separated from the cleavage product on a Zorbax-Si1 column developed in the same solvent. Silylation-l,24,25-Dihydroxyvitamin Dz (500 ng), purified from bovine kidney homogenates incubated with 1,25-(OH)zDz, was allowed to react with N-methyl-N-(trimethylsily1)trifluoroacetamide (50 pl) at 90 “C for 2 h to affect persilylation. The silylated product was eluted with 0.1% methylene chloride in methanol from a Whatman ODs-3 HPLCcolumn (0.42 X 25 cm). Biological Evaluation-Male weanling rats (Holtzman, Madison, WI) were housed individually in overhanging wire cagesand were fed a vitamin D-deficient diet containing low calcium (0.005%) and normal phosphorus (0.3%). Test compounds were dissolved in 10-20 p1 of ethanol. Plasma (1.8 ml), collected from vitamin D-deficient rats, was added to the ethanol solution. The test compounds were given intrajugularly in 0.3 ml of the ethanol/plasma carrier solution. Controls received carrier alone. Eighteen hours after injection, the rats were decapitated and theirduodena were used to measure intestinal transport of &Ca by the everted gut sac technique (30). Blood was collected in heparanized tubes and centrifuged. The resulting plasma was measured for calcium concentration by atomic absorption spectroscopy to determine the degree of bone resorption. Since rats were feda diet essentially devoid of calcium, plasma calcium increases reflected mobilization of calcium from bone, not intestinal calcium absorption. Competitive Protein Binding Assays-The ability of1,24,25(OH),DZ and 1,24,25-(OH)3D3to compete for 1,25-(OH)Z[3H]D3receptor binding sites was evaluated with receptors prepared from bovine thymus, chick intestine, and ratintestine. The receptors were prepared and theassays were performed as previously described (25). In addition, eachmetabolite was tested for its ability to compete with 25-OH[3H]D3for binding sites on the ratplasma vitamin D-binding protein. The assay was performed as described previously (23). P h n a Tracer Kinetics-Each of six male rats, fed a stock diet (320-340 g), were injected intravenously with 1 pCi (11 pmol) each of 1,24,25-(OH)3[3H]Dz and 1,24,25-(OH)3[3H]D3.The injections were prepared by dissolving the radioactive metabolites in 50 pl of ethanol and suspending the dissolved materials in 3 ml of plasma collected from rats maintained on a stock diet. Each rat received 0.3 ml of the mixture. One ml of blood was taken from each of three rats at each time point. Blood was collected from the jugular vein in a heparinized syringe while the rats were maintained under isoforane anesthesia. The ratswere divided into two groups of three. Bleeding times were alternated between the two groups, and no rat was bled more than four times duringthe 8-hcourse of the experiment. Plasma lipids were extracted and prepared for analysis of 1,24,25by HPLC (31). Separation of (OH)3-[3H]Dzand 1,24,25-(OH)3[SH]Ds these two metabolites was achieved using a Zorbax-Si1column developed in 1550135isopropanol/methylene chloride/hexane. Radioinert 1,24,25-(OH)3D3 was used to monitor recovery by comparing UV peak heights of unknowns with those of standards. The tracer kinetic data were analyzed using CONSAM (32), a conversational version of the SAAM computer program (33). The program solves the differential equations for specific kinetic model, iteratively adjusts the parameter values to achieve a least squares fit to the tracer data, and calculates the standard deviations of the parameter values with respect to the inputdata. A variety of program features, such as automatic calculation of steady solutionsand capacity for simultaneous analysis of multiple sets of data, facilitate hypothesis testing inthe model building.

RESULTS

Kidney homogenates, prepared from 1,25-(OH)2D2-dosed calves, produced a major polar metabolite upon incubation with radioinert 1,25-(OH)2Dz(Fig. 2B). The metabolite was purified with a total of three chemically distinct HPLC systems, capable of separating all known 1,25-(OH)2Dmetabolites (34). The purified metabolite had a UV absorbance spectrum with a ,A, at 265 nm and a ,A, at 228 nm. The ratio was 1.8. These results are consistent with a vitamin D-like 5(E),7,10(19)-triene chromophore. On thebasis of the spectrum, approximately 45 pg of the metabolite was isolated. The E1 mass spectrum of the metabolite showed a molecular ion at m/z 444 which is consistent with hydroxylation of the parent compound (Fig. 3). Peaks at m/z 426,408, and 390 represent losses of 1,2, and3 molecules of water, respectively. The peak at mJz 368 results from cleavagebetween c-24 and C-25 with subsequent loss of water involving the C-24 hydroxyl group. Loss of water from this fragment provided the peak at m/z 350. Loss of the side chain and a molecule of water provided the peak at m/z 269. Loss of water from m/z 269 provided mlz 251. The peak at 134 is typical of lahydroxylated vitamin D compounds; it resulted from cleavage between C-6 and C-7, followed byloss of water. The base peak at m/z 59 results from cleavage between C-24 and C-25 to produce ((CH3)P= OH)’. Persilylation of the metabolite produced a derivative that had a molecular ion at m/z 732, indicating a total of four hydroxyl groups were present (Fig. 4). Peaks at m/z 717,642, and 627 represent losses of a methyl group, trimethylsilylOH, and both, respectively. Peaks at m/z 601 (M+ 131) and 131 (C3HGOSi(CH3),)resulted from cleavage between c-24 and C-25 and indicate that the additional hydroxyl group is not at (2-26. The peak at mJz 552 results from loss of2trimethylsilyl-OH groups. The peak at m/z 511 was produced by loss of a trimethyl-OH group and cleavage between C-24 and C-25, i.e. loss of m/z 131 at the same time. The peak at m/z 480 likely represents loss of CH2Si(CH3)2from m/z 552. Loss of water from m/z 480, generated from the oxygen atom remaining after loss of CH2Si(CH3)2,provided m/z 462. The intense peak at 206 is the silylated equivalent of mlz 134 in the mass spectrum of the underivatized sample. The metabolite was sensitive to sodium periodate treatment (data notshown). The product had a chloride ion (NCI) mass

-

205

’ 7 ’ ”

II I -1 II I

0

MI 2 FIG. 3. E1 mass spectrometry of peak I (1,24,26-(OH)~D-d produced from 1,26-(OH)aD~.

24-Hydroxylationof 1,25-Dihydroxyergocalciferol r IO1

MIZ

FIG.4. E1 mass spectrometry of the tetrasilylated derivative of 1,24,26-(OH)sDa.

RI

I

+

i

i

i

i

i

9253

upfield portion of the spectrum (see insert, Fig. 5). Due to the chirality present in the side chain, C-26 and C-27 methyl groups are not chemical shift equivalent and give rise to distinct singlets at 61.21 and 61.19. The C-28 methyl group was assigned to the singlet at 61.27. The signal at 61.25 has been verified as an impurity present in the CDCl,. Assignments made with regard to the side chain closely match the known values obtained for 24,25-(OH)2D2 (34). The data are therefore consistent with hydroxylation of 1,25-(OH)2D2 at (2-24. Conversion of 24,25-(OH)ZD2to 1,24,25-(OH)3Dzwould furthersubstantiatethestructural assignment. Therefore, radioinert and radioactive 24,25(OH)zDzwere individually incubated with kidney homogenate from vitamin D-deficient, calcium-deficient chicks. Initial HPLC analysis revealed that the chick kidney incubations containing the radioinert and radioactive 24,25-(Of&D2 resulted in the production of a metabolite (Fig. 2 4 ) that coeluted with the putative 1,24,25-(OH)3Dz(Fig. 2 B ) produced in bovine kidney homogenates. Further analysis of the 1,24,25-(OH)sD2 produced from radioinert 24,25-(OH)2Dz demonstrated co-migration of this metabolite with the metabolite produced from 1,25-(OH)2Dzon two chemically distinct HPLC systems (Figs. 6, 7). No known la-hydroxylated metabolite of cholecalciferolor ergocalciferol co-migrates on both of these systems (34). Furthermore, upon evaluation by E1 mass spectroscopy, the 1,24,25-(0H),D2produced from 24,25(OH)ZD2 showed a molecular ion a t m/z 444 and the diagnostics peak at mlz 426, 408, 390,368,350, 269, 251,134, and 59, with the same relative intensities observed with 1,24,25(OH)3D2 produced from 1,25-(OH)2Dz, demonstrating its identity as 1,24,25-(OH)3D2. Biological Evaluation-The intestinal calcium transport (ICT) and bone calcium resorption (BCR) responses of rats

I VI1

1

I

0

PI%

FIG. 5. Three hundred-MHzprotonNMR of 40 fig of 1,24,25-(OH)sDa isolated from bovine kidney homogenates incubated with 1,25-(OH),Dz. The concentration was 40 pg in 300 pl of CDCL. The peak at S1.25 represents an impurity in the CDCL.

+

spectrum with peaks a t m/z 419 (384 35Cl) and mlz 421 (384 + 37Cl)indicating a molecular weight of 384, i e . cleavage between C-24 and C-25, demonstrating that a vicinal C-241 C-25 diol was in place in the parentcompound. A 300 MHz proton NMR spectrum was also taken with 40 pg of metabolite (Fig. 5). The presence of a In-hydroxyl group is suggested by the appearance of the multiplet located at 64.42 corresponding to theC-16 proton. Additional supporting evidence is provided by the characteristic downfield shifts observed in the olefinic region. The C-6 proton is located at 66.37 and the C-19 protons (2and E ) are assigned to the signals seen at 64.99 and 65.31. These values are in excellent agreement with chemical shifts reportedfor other la-hydroxylated vitamin D compounds (34). The protons on the C-22 olefin bond produced a multiplet at 65.56. This signal is shifted downfield relative to itsknown position in vitamin Dzand is consistent with the chemical shift observed for these protons in 24,25-(OH)2D2(34). The positions of side-chain hydroxylation were determined by the examination of the expanded

0 ml FIG.6. HPLC migration of 1,24,2b-(OH)sDa. A Zorbax-Si1 column developed in isopropanol/methylene chloride (793)was used. The HPLC profile of Panel A was achieved using approximately 50 ng of each metabolite. The 1,24,25-(OH)3D2was isolated from calf kidney homogenates incubated with 1,25-(OH),D,. Panel B represents the same conditions and quantities of metabolites listed in Panel A, but includes 50 ng of 1,24,25-(0H)3& produced from chick kidney homogenates incubated with 24,25-(OH)2D2.

9254

24-Hydroxylation of 1,25-Dihydroxyergocalciferol

-n

0

O I

20

50

40

30

ml FIG. 7. Further HPLC analysis of1,24,26-(OH)sDa. The source and quantities of metabolites for Panels A and B were the same as Panels A and B , respectively, in Fig. 6. The mobile phase for the Zorbax-Si1 column, however, was 2-propanol/hexane (1090).

1,25-(OH)2D3.However, 1,24,25-(OH)3D3was 4-fold less active than 1,%-(OH),D3 at stimulating BCR (i.e. 50 ng of 1,24,25-(OH)3D3gave a similar BCR response as 12.5 ng of 1,25-(OH),D,). CompetitiveProtein Binding Assays-The ability of 1,24,25(OH)3Dz to compete with 1,25-(OH),['H]D3 for the 1,25(0H)ZD receptor was examined in different binding protein preparations. As shown in Fig. 8, 1,24,25-(OH)3D~ competed equally with 1,25-(OH)2D3and 1,24,25-(OH)3D3for receptor prepared from bovine thymus. Likewise, 1,24,25-(OH)3Dz was equivalent to 1,24,25-(OH)3D3in affinity for receptor prepared from chick intestine; each was approximately &fold less potent than 1,25-(OH)&. In receptor prepared from rat intestine, 1,24,25-(OH)3D2had 2-fold higher affinity than 1,24,25(OH),D3, but 6.5-fold less affinity than 1,25-(OH)zD3. We also examined the ability of the metabolites to compete for binding sites on the plasma vitamin D transport protein. The 1,24,25-(OH)3D2had approximately %fold less affinity than 1,24,25-(OH)3D3and approximately 6-fold less affinity than 1,25-(OH)zD3at displacing 25-OH['H]D3 from the rat plasma vitamin D transport protein. Plasma Tracer Kinetics-The data in Fig. 9 show the relative disappearance rates of 1,24,25-(OH)3[3H]Dzand 1,24,25(OH),['H]D, from plasma following intravenous doses to normal rats. At least two exponential terms were required to fit the tracer data for the 480-min sampling period. The best fitting exponential functions for 1,24,25-(OH)3['H]Dz and 1,24,2~5-(OH)~['H]D~ were: c ( t ) = 4.8e - 0.086t 1.6e - 0.002t, and c( t) = 3.7e - 0.051t 2.8e - 0.0015t, respectively, where c( t ) is the plasma radioactivity as "%" of dose per ml and "t"

+

+

TABLE I Biological evaluation of 1,24-(0H)&, 1,25-(0H)&, 24-keto-1,25(OH)&', 1,24,25-(OH)&,and 1,24,25-(OH)3o3wing the rat bioassay Values with different superscriptsare different p < 0.05. (Z k S.D., n = 6). ComDounds ~ v e n

Experiment I Control 1,25-(OH)2D3 1,24,25-(OH)3Dz Experiment I1 Control 1,25-(OH)zDz 1,24,25-(OH)3D2 Experiment I11 Control 1,25-(OH)zDz 1,24,25-(OH)3Dz Experiment IV Control 1,25-(OH)zD, 24-keto-l,25-(OH)2D'

Dosage &rat

12.5 2.9 12.5 25.0 3.6 50.0 3.0

ICY

1.5 f 0.3l5.9 f 0.42 6.5 2.8 f 0.2' 6.0 f 0.5' f 0.8' 6.4

BCRb

f 0.3l f 0.52 f 0.5l 6.2 k 0.5l f 0.6'

12.5 50 100 200 2.9

2.1 f 0.3l4.8 f 0.4l 3.8 f 0.6' 6.3 f 0.4' 5.0 f 0.5l 2.0 f 0.6' 5.3 f 0.4' 3.0 k 0.7' f 0.5' 5.4 f 0.52

12.5 200 2.2 400 800

1.7 f 0.3l 5.4 2.6 f 0.5' f 0.3l 4.4 1.9 f 0.3' 4.5 2.0 f 0.4l 4.4

4.4 f 0.3l f 0.2' f 0.2l f 0.4l f 0.3l

12.5 12.5 50 100

1.4 & 0.2'4.4 3.7 f 0.S3 5.0 4.6 2.8 k 0.3' 4.6 2.3 f 0.4' 2.6 & 0.5' 4.4

f 0.2' f 0.2' f 0.2l f 0.2l f 0.3'

%a (serosal)/"Ca (mucosal).

* Plasma calcium (mg/dl).

receiving a single dose of test compound 18 h prior to experiment are recorded in Table I. The datasuggest that atdoses up to800 ng, the 1,24,25-(OH)3Dzhas little orno ICT orBCR activity. The metabolite 1,24,25-(OH)3D3,on the other hand, had significant ICT at the12.5-ng dose and was equivalent to

q /TU%

FIG. 8. Relative binding of1,25-(oH)2Ds,1,24,26-(OH)sDa, and 1,24,25-(OH)sD3to the 1,25-(OH)2Dreceptor using cytosolic preparations from calf thymus, rat intestine, chick intestine, and the vitamin D-binding protein in rat plasma. Radioactive 1,25-(OH),D3 was used in the cytosol binding assays, while radioactive 25-OHDa was used in the rat plasma binding assay.

1 0.1 : 0

4 LlOtI I

I

I

I

1

100

200

300

400

500

TIME (min) FIG. 9. Plasma turnover of 1,24,26-(OH)s[8H]D2 and 1,24,25-(OH)s[sH]Dsin normal 325-350-g rats.

24-Hydroxylationof 2,25-Dihydroxyergocalciferol is the time in min postinjection. The overall disappearance (tll2)rates for 1,24,25-(OH)3[3H]D2and 1,24,25-(OHM3H1D3 were 352and 470 min, respectively. From the compartmental model (Fig. 9),the sizes of the compartments, rates of intercompartmental exchange, and turnover were calculated (Table 11). A compartmental modelwasemployed tointerpretthe dilution, exchange, and disappearance of 1,24,25-(OH)3[3H) D3 and 1,24,25-(OH)3[3H]D2. The biexponential nature of the disappearance curves indicated the need for two exchanging compartments for each system. After fitting a two-compartment model to each of the data sets, the differences in model parameters for the D, and D3 metabolites were evaluated by simultaneously fitting both sets of data subject to the constraint that one or more parameters were identical for 1,24,25-(OH)3[3H]D3and 1,24,25(OH)3[3H]Dzkinetics (33). Initially, a model having irreversible loss of tracer (L(0,l)) emanating from compartment 1 was studied. Using this model, it was found that differences in two parameters (L(0,l) and L(2,l)) were necessary and sufficient to describe the differences in the two sets of data. The distribution volume for compartment 1 and L( 1,2) and thefractional rateof tracer return from compartment 2 to compartment 1 could be constrained to identical values for both systems. Subsequently, an alternative model having irreversible tracer loss (L(O,2)) emanating from compartment 2 was investigated. Simultaneous fitting of both sets of data to this model showedthat anacceptable fit to both sets of data could be obtained with only a single parameter (L(2,l)) differing between the two systems. On this basis, we concluded that the model having a loss pathway from compartment 2 was preferable since it provided a simpler explanation of the differences in metabolism between 1,24,25-(OH)3[3H]D2and 1,24,25-(OH)3[3H]D3. The parameters and steady-statesolution of this model are presented in Table 11. The fractional rate of transport from compartment 1 to compartment 2 was%fold higher for 1,24,25-(OH)3[3H]D~ than for 1,24,25-(OH)3[3H]D3. This implied a more rapid clearance of the Dz metabolite from the circulation. However, metabolite in compartment 2 had the same probability of being returned to the circulation (or removed irreversibly) for 1,24,25-(OH),[3H1Dz and 1,24,25(OH)3[3H]D3. Although the initial volume of distribution ( V(1)) was identical for 1,24,25-(OH)3[3H]D2and 1,24,25-(OH)3[3H]D3, the plasma-equivalent virtual volume for compartment 2 (V(2)) was also 2-fold higher for 1,24,25-(OH)3[3H]Dzas a TABLEXI Calclllatwns derived from the tracer kinetic model for 1,24,25(OH),fH]D, and 1,24,25-(OH)3fHjD3 L(I,J) is the fractional rate of transport (min") from compartment J t o compartment I. V(r)is the distribution volume for compartment I . Cl(I,J) is the steady-state clearance (ml/min) of the metabolite from compartmentJ to compartment I. Values & S.D. were calculated by the CONSAM program during a simultaneousfit to both 1,24,25(OH)a[3H]D2and 1,24,25-(OH)3[3H]D3data. Parameter 1,24,25-(OH)a[3H]Dz 1,24,25-(OH)aIaH]Ds L(2,1) L(172) "4092) V(1) V(2) ClW) C1(1,2) Cl(0.2)

0.0650 f 0.015 0.0200 & 0.005 0.0026 f 0.0005 16.0 +- 1.3 45.0 f 5.1 1.00 & 0.17 0.90 f 0.17 0.12 f 0.012

0.0300 4 0.0074 0.0200 2 0.005 0.0026 f 0.0005 16.0 k 1.3 20.0 f 2.6 0.460 +. 0.086 0.410 _+ 0.084 0.053 +- 0.0061

9255

consequence of the more rapid clearance into this peripheral compartment from the circulation. DISCUSSION

This report identifies 1,24,25-(OH)3Dzand demonstrates that it is a significant 1,25-(OH)zDzmetabolite of kidney homogenates prepared from 1,25-(OH)2Dz-dosedcalves. The structural assignment as 1,24,25-(OH)3D~ was determined by chromatographic and spectral means and by comparing it to the product produced from the 1-hydroxylation of 24,25(OH)2D2.Following the initial confirmation of its structure, this new metabolite was evaluated extensively for biological activity using several methods. As shown in Fig. 8, the ratintestinal receptor bound 1,24,25(OH)JD~with approximately %fold greater affinity than 1,24,25-(OH)&.. In previous experiments conducted by our laboratories, we have shown the rat to be unique regarding other aspectsof ergocalciferolmetabolism. In theexperiments where animals were dosed orally with equal amounts of ergocalciferol and cholecalciferol, the rat produced more25OHD2 than 25-OHD3, whereas the pig and chick produced more 25-OHD3than !%-OH& (35). This data, as well as the receptor data presented, are consistent with ergocalciferol being the metabolicalIy perferred vitamin D source for rats. Like other investigators (36), we have shown the 1,24,25(OH)&, has significant biological activity and is roughly equal to 1,25-(OH)2D3inits ability to stimulate ICT. Its ability to stimulate BCR, however, is markedly attenuated (Table I). The 1,24,25-(OH)3Dz,on the other hand, has very little activity and no BCR activity at doses 10-50 times that Of 1,25-(OH)zD3. The metabolite 1,24,25-(OH)3D3is converted into 24-keto1,25-(OH)2D3 in vivo (6, 8, 9). 24-Keto-1,25-(OH),D3 has affinity a t least equivalent to 1,24,.%-(OH)3D3for the chick intestinal andbovine thymus cytosolic 1,25-(OH)zDreceptors (6), and,therefore, 24-ketonization of 1,24,25-(OH)3D3apparently is not a deactivation event per se. We, therefore, tested 24-keto-1,25-(OH),D3for ICT andBCR activity to determine whether the activity of the 24-keto metabolite would provide an obvious explanation for the differences between 1,24,25(OH),DZ and 1,24,25-(OH)3D3.24-Ket0-1,25-(OH)~D~, however, at least within the pharmacological constraints of the experiment, was not as active as 1,24,25-(OH)3D3in uiuo. Thus, this approach did not readily provide insight into the mechanisms of the differences. Insight into the differences in activity was provided by the plasma turnover experiments. Notably, 1,24,25-(OH)3Dzwas cleared from plasma of rats approximately 40% faster than 1,24,25-(OH)3D3 (Fig.9). This difference in turnover may be largely responsible for the observed dissimilarities in biological activity of the two metabolites. The difference in plasma clearance may be the result of the relatively low affinity that 1,24,25-(OH)& has for the plasma vitamin D transport protein, rendering 1,24,25-(OH)3Dzavailable for more extensive degradation in tissues. We have studied an analogous situation with regard to thecompounds 1,25-(OH)zDzand epi1,25-(OH)2Dz. Both compounds have equal affinity for the 1,25-(OH)2D receptor, but the latter binds 3- to 6-fold less avidly to the plasma vitamin D-binding protein. The epi compound is also cleared from plasma approximately 4-fold faster than 1,25-(OH)zDzand is approximately &fold less potent at stimulating ICT andBCR.3 These results, with the different forms of seco-steroid hormones, are in contrast with results obtained with other steroid T. A. Reinhardt and R. L. Horst, unpublished results.

9256

24-Hydroxylationof 1,25-Dihydroxyergocalciferol

hormones. Previous work with naturally occurring estrogens has shown similar differences in binding, but with a different outcome. Estradiol-17P binds to plasma transport components 10- to 100-fold better than estriol, and therefore free estriol exists in blood at a higher concentration and is more available for tissue interaction (37-39). As a result, even though estradiol-17P has a 10-fold higher affinity for the estrogen receptor than estriol, both are equally effective at stimulating early uterotropic events. The discrepancy between predicting and realized potency, based on receptor binding, results from differential binding of the two compounds to plasma components. Based on this example, we might have predicted that 1,24,25-(OH)3Dz wouldhave much greater biological activity than 1,24,25-(OH)3D3.Our data show, therefore, that factors other than rapid entry into targetcells (due to higher concentration of free hormone), along with greater (or equal)receptor affinity, are required for biopotency. In summary, the presence of 1,24,25-(OH)3D2in calf kidney homogenates incubated with 1,25-(OH)zDzhas been demonstrated. This newly described metabolite was a major product of renal 1,25-(OH)zDzmetabolism and was found to be rapidly removed from plasma, and relatively weak in stimulating eitherICT or BCR. The data strongly suggest that C-24 hydroxylation of 1,24,2E1-(OH)~D2 targets 1,25-(OH),D2 to deactivation, whereas C-24 ketonization, not C-24 hydroxylation, is required for similar deactivation of 1,25-(OH)~D3. Thus, these data also add to our understanding of steroid hormones in general by demonstrating the importance of target tissue metabolism to the ultimate expression of hormone activity. Acknowledgments-The kinetic analysis was performed at the University of Pennsylvania School of Veterinary Medicine using the computing facilities supported by the Pennsylvania Department of Agriculture. We wish to thank Derrel Hoy for his expert technical assistance and toAnnette Bates for preparation of the manuscript. REFERENCES 1. Holick, M. F., Frommer, J. E., McNeill, S. C., Richtand, N. M., Henley, J. W., and Potta, J. T., Jr. (1977)Biochem. Biophys. Res. Commun. 76, 107-114 2. Bills, C . E. (1967)The Vitamins (Sebrell, W.H., and Harris, R. S., eds) Vol. 2,pp. 149-173,Academic Press, New York 3. Holick, M. F., Kleiner-Bossallier, A., Schnoes, H. K., Kasten, P. M., Boyle, I. T., and DeLuca, H. F. (1973)J. Biol. Chem. 248, 6691-6696 4. Kumar, R., Schnoes, H. K., and DeLuca, H. F. (1978)J. Biol. Chem. 253,3804-3809 5. Reinhardt, T. A., Napoli, J. L., Beitz, D. C., Littledike, E. T., and Horst, R. L. (1982)Arch. Biochem. Biophys. 213,163-168 6. Napoli, J. L., Pramanik, B. C., Royal, P. M., Reinhardt, T. A., and Horst, R. L. (1983)J. Biol. Chem. 268,9100-9107 7. Napoli, J. L.,and Martin, C . A. (1984)Biochem. J. 219,713-717 8. Ishizuka, S., Ishimoto, S., and Norman, A. W. (1984)Biochemistry 23,1473-1478 9. Mayer, E., Bishop, J. E., Chandraratna, R. A. S., Okamura, W. H., Kruse, J. R., Popjak, G., Ohnuma, N., and Norman, A. W.

(1983)J. Bid. Chem. 268,13458-13465 10. Napoli, J. L., and Horst, R.L. (1983)Biochemistry 22, 58485853 11. Reinhardt, T. A., Napoli, J. L., Praminik, B., Littledike, E. T., Beitz, D. C., Partridge, J. J., Uskokovii., M. R., and Horst, R. L. (1981)Biochemistry 20,6230-6235 12. Tanaka, Y.,Schnoes, H. K., Smith, C.M., and DeLuca, H. F. (1981)Arch. Biochem. Biophys. 210, 104-109 13. Horst, R. L., Wovkulich, P. M., Baggiolini, E. G., Uskokovic, M. R., Engstrom, G. W., and Napoli, J. L.(1984)Biochemistry 23, 3973-3979 14. Napoli, J. L., and Horst, R. L. (1983)Biochem. J. 214,261-264 15. Ohnuma, N., and Norman, A. W. (1982)Arch. Biochem. Bwphys. 204,387-391 16. Jones, G., Kano, K., Yamada, S., Jurusawa, T., Takayama, H., and Suda,T. (1984)Biochemistry 23,3749-3753 17. Jones, G., Jung, M., and Kono, K. (1983)J. Biol. Chem. 258, 12920-12928 18. Mayer, E. G., Reddy, S., Chandrarantna, R. A. S., Okamura, W. H., Kruse, J. R., Popjak, G., Bishop, J. E., and Norman, A. W. (1983)Biochemistry 22,1798-1805 19. Napoli, J. L., Pramanik, B. C., Partridge, J. J., Uskokovik, M. R., and Horst, R. L. (1982)J. Biol. Chem. 257,9634-9639 20. Jones, G., Rosenthal, A., Segev, D.,Mazur, Y.,Frolow, F., Halfon, Y., Rabinovich, D., and Shakked., Z. (1979)Biochemistry 18, 2094-2101 21. Jones, G., Schnoes, H. K., and DeLuca, H. F. (1975)Biochemistry 14,1250 22. Roborgh, J. R., and Demann, T. J. (1960)Biochern. Pharmacol. 3,272-282 23. Horst, R. L., Littledike, E. T., Riley, J. L., and Napoli, J. L. (1981)Anal. Biochem. 116,189-203 24. Sommerfeldt, J. L., Napoli, J. L., Littledike, E. T., Beitz, D. C., and Horst, R. L. (1983)J. Nutr. 113,2595-2600 25. Reinhardt, T. A., Horst, R.L., Orf, J. W., and Hollis,B. W. (1984)J. Clin. Endocrinol. & Metab. 58,91-98 26. Reinhardt, T. A., Horst, R. L., Littledike, E. T., and Beitz, D. C . (1982)Biochem. Biophys. Res. Cornmun. 106,1012-1018 27. Engstrom, G . W., Horst, R. L., Reinhardt, T. A., and Littledike, E.T.(1984)J. Nutr. 114, 119-126 28. Tanaka., Y.. . and DeLuca. H. F. (1981) . . Anal. Bwchem. 110.102107 29. Horst, R. L., Shepard, R. M., Jorgensen, N. A., and DeLuca, H. F. (1979)Arch. Biochem. Biophys. 192,512-523 30. Martin, D. L., and DeLuca, H. F. (1969)Am. J. Physiol. 216, 1351-1359 31. Horst, R. L. (1984)in Vitamin D: Basic andClinical Aspects (Kumar, R., ed) pp. 423-478,Martinus Nijhoff, Boston 32. Boston, R. C., Grief, P. C., and Berman, M. (1981)Comput. Prog. Biowd. 13, 11-119 33. Berman, M. (1963)J, Theor. Biol. 4,229-236 34. Napoli, J. L., Koszewski, N. J., and Horst, R. L. (1986)Methods Enzymul. 123,127-140 35. Horst, R. L.,Napoli, J. L., and Littledike, E. T. (1982)Biochem. J. 204,185-189 36. Castillo, L., Tanaka, Y., DeLuca, H. F., and Ikekawa, N. (1978) Miner. Electrolyte Metab. 1, 198-207 37. Anderson, J. N., Peck, E. J., dr., and Clark, J. H. (1974)J. Steroid Biochem. 6,103-107 38. Anderson, J. N., Peck, E. J., Jr., and Clark, J. H. (1975)Endocrinology 96,160-167 39. Hardin, J. W., Clark, J. H., Glasser, S. R., and Peck, E. J., Jr. (1976)Bioehemisty 15,1370-1374