Radioligand Receptor Assay for 25-Hydroxyvitamin D2/D3

and la,25-Dihydroxyvitamin D2/D3

APPLICATION TO HYPERVITAMINOSIS D

MMu R. HuGcms, DAVID J. BAYLINK, PATRICIA G. JoNEs, andMARKR. HAUSSLER

From the Departments of Biochemistry, College of Medicine, and Cell andDevelopmental Biology, University of Arizona, Tucson, Arizona, 85724, andthe Veterans Administration Hospital and Department of Medicine,University of Washington, Seattle, Washington 98108

A B S T R A C T A competitive protein binding assay formeasurement of the plasma concentration of la,25-dihy-droxyvitamin D3 [la,25-(OH)aD3] has been extendedto include the immediate precursor of this hormone,25-hydroxyvitamin Ds (25-OHD3). In addition, the assaysystem is capable of measuring the two metabolic prod-ucts of ergocalciferol, namely, 25-hydroxyvitamin D2(25-OHD2) and la,25-dihydroxyvitamin D2 [la,25-(OH) 2D2]. The target tissue assay system consists of ahigh affinity cytosol receptor protein that binds the vita-min D metabolites and a limited number of acceptorsites on the nuclear chromatin. By utilizing a series ofchromatographic purification steps, a single plasmasample can be assayed for any of the four vitamin D me-tabolites either individually or combined. Therefore, theassay procedure allows for both the quantitative andqualitative assessment of the total active vitamin D levelin a given plasma sample.

To show that the binding assay was capable of mea-suring 1a,25-(OH)2D2 as well as la,25 (OH)2D3, twogroups of rats were raised. One group, supplementedwith vitamin D3, produced assayable material that repre-sented la,25- (OH)2D3. The other group, fed only vita-min Da in the diet, yielded plasma containing only la,25-(OH)2D2 as the hormonal form of the vitamin. Thecirculating concentrations of the two active sterols werenearly identical (15 ng/100 ml) in both groups, indi-cating that the competitive binding assay can be used tomeasure both hormonal forms in plasma.

In a separate experiment, la,25-(OH)2D2 was gen-erated in an in vitro kidney homogenate system using

Dr. Baylink is the recipient of Research Career Develop-ment Award DE 19108.

Received for publication 2 September 1975 and in revisedform 8 March 1976.

25-OHD2 as substrate. Comparison of this sterol withla,25- (OH) 2D3 in the assay system showed very similarbinding curves; the Da form was slightly less efficient(77%). Comparison of the respective 25-hydroxy forms(25-OHD2 vs. 25-OHD3) at concentrations 500-foldthat of la,25-(OH)2D3, again suggested that the bindingof the D2 metabolite was slightly less efficient (71%).

Finally, the assay was employed to measure the totalactive vitamin D metabolite pools in the plasma of nor-mal subjects and patients with varying degrees of hy-pervitaminosis D. The normal plasma levels of 25-OHDand la,25-(OH) 2D measured in Tucson adults were25-40 ng/ml and 2.1-4.5 ng/100 ml, respectively. Bothsterols were predominately (> 90%) in the form of vi-tamin D3 metabolites in this environment. Typical casesof hypervitaminosis D exhibited approximately a 15-foldincrease in the plasma 25-OHD concentration, and adramatic changeover to virtually all metabolites existingin the form of D2 vitamins. In contrast, the circulatingconcentration of la,25- (OH)2D was not substantiallyenhanced in vitamin D-intoxicated patients. We there-fore conclude that hypervitaminosis D is not a result ofabnormal plasma levels of la,25-(OH)2D but may becaused by an excessive circulating concentration of25-OHD.

INTRODUCTION

It is now well established that 25-hydroxyvitamin D3(25-OHD3)1 and la,25-dihydroxyvitamin D3 [la,25-

'Abbreviations used in this paper: 25-OHD2,25-hydroxy-vitamin D2 or 25-hydroxyergocalciferol; 25-OHD3, 25-hy-droxyvitamin D3 or 25-hydroxycholecalciferol; la,25- (OH) 2D2, la,25-dihydroxyvitamin D2; la,25- (OH)2D3, la,25-dihy-droxyvitamin D3. When no subscript is present after the D,both D, and D3 are implied.

The Journal of Clinical Investigation Volume 58 July 1976- 61-70 61

(OH)2Ds] are two metabolites of the native vitamin in-volved in regulating calcium and phosphorus metabolism(1). Formation of 25-OHD3 occurs in the liver, intes-tine, and kidney (2), and subsequent activation of thesterol to la,25-(OH)2D3 occurs exclusively in the kid-ney (3), producing what is considered to be the hor-monal form of the vitamin. la,25- (OH)2D3 has beenshown to be the most active metabolite of vitamin D instimulating intestinal calcium transport (4) and bonecalcium mobilization (5). Although the exact sequenceof events that mediate the effect of la,25-(OH)2Ds onits target tissue is not yet completely understood, anearly step has been shown to be the interaction betweenthe hormone and a specific macromolcular component inthe target tissue cytoplasm (6). The hormone-receptorcomplex subsequently translocates to the nuclear compart-ment where it attaches to specific acceptor sites (6-8).Our laboratory has utilized this series of events to de-velop a sensitive competitive binding assay for the hor-monal form of vitamin D (9). The method utilizes asaturable, reconstituted receptor system consisting ofintestinal mucosa cytosol and nuclear chromatin. An iso-tope-dilution curve is obtained by incubating reconsti-tuted cytosol-chromatin and [8H] la,25- (OH)2Ds withincreasing amounts of nonradioactive la,25-(OH) 2D3and then separating the free and bound sterol by trap-ping the sterol-receptor-chromatin supercomplex on glassfiber filters. This technique has been shown to be a re-liable and valid method for quantitation of hormone con-centrations as low as 1.0 ng/100 ml plasma (10).

The present study was undertaken for the followingreasons: (a) to determine if the presently available chro-matography methods could separate vitamin D3 fromvitamin Da metabolites, (b) to ascertain if the discrimi-nation against ergocalciferol metabolites by the chickis at the level of the sterol's target tissue receptor, (c)to extend the la,25-(OH)2D3 assay (9, 10) to incorpo-rate 25-hydroxyvitamin D2 (25-OHDo), 25-OHD3, andla,25-dihydroxyvitamin Da [la,25-(OH)2D2], thus de-veloping a single method for the qualitative and quanti-tative assessment of these biologically active forms ofvitamin D and, (d) to measure these metabolites in hy-pervitaminosis D patients, possibly allowing a more com-plete delineation of the disorder.

METHODS

Animals. Chicks used were White Leghorn cockerels(donated by Demler Farms, Anaheim, Calif.) that wereraised on a vitamin D-deficient diet (11) for 3-4 wk. Maleweanling Holtzman rats were used in the experiment gen-erating la,25-(OH)2Ds and la,25-(OH)2D2 in vivo.

Sterols. Crystalline 25-OHD2 and 25-OHD, were a gen-erous gift from Dr. Jack Hinman and Dr. John Babcock ofthe Upjohn Company, Kalamazoo, Mich. [26,27-methyl-3H]-25-OHD3 (6.5 Ci/mmol) was obtained from Amersham/Searle Corp., Arlington Heights, Ill. [26,27-methyl-'H] la,

25-(OH)2D3 (6.5 Ci/mmol) was produced in vitro, by amodification (12) of the method of Lawson et al. (13) andpurified by column chromatography as previously described(7, 14). Nonradioactive la,25-(OH)2D3 was either preparedand purified via similar procedures (12) or obtained as thecrystalline, synthetic compound from Dr. M. Uskokovic ofHoffman-LaRoche Inc., Nutley, N. J. Biologically generatedand synthetic nonradioactive la,25-(OH)2D3 yielded identi-cal quantitative competition in the binding assay, verifyingthe equivalency of sterol from these two sources. Nonradio-active la,25-(OH)2D2 was generated in vitro in a similarfashion (12). Sterols were quantitated by ultraviolet ab-sorption spectrophotometry and stored in distilled ethanol at-200C.

Chromatography. After addition of 2,000 cpm each oftritiated 25-OHD3 and la,25-(OH)2D3 (6.5 Ci/mmol) tothe plasma and extraction of sterols as described elsewhere(10), chromatographic procedures were carried out as fol-lows. (See Fig. 1 for schematic diagram.) Sephadex LH-20(Pharmacia Fine Chemicals Inc., Piscataway, N. J.) columnchromatography was performed as described previously(10). One 1 X 15-cm (5 g) column resulted in the separa-tion of 25-OHD from la,25-(OH)2D; 25-OHD emergedbetween 0 and 40 ml elution volume, whereas la,25-(OH) 2Deluted between 40 and 95 ml. Throughout the remainder ofthe purification scheme the sterols were treated separately.25-OHD was purified on silicic acid (Bio-Rad Laboratories,Richmond, Calif., minus 325 mesh; activated by heating at120°C for 24 h in a vacuum oven) by eluting first with 20ml hexane-diethyl ether (1: 1) and then with 20 ml diethylether; 25-OHD emerged in the second fraction with theether. la,25- (OH) 2D was purified similarly on silicic acidby elution with diethyl ether (10 ml) and then acetone (8ml); the hormone emerged in the second fraction withthe acetone. The entire procedure was miniaturized to 0.8X 6.4-cm columns to provide a very rapid and simple meansof removing contaminating lipid. Final purification of thesterols was accomplished by Celite (Johns-Manville, NewYork) liquid-liquid partition chromatography (14). Thesolvent system (System I) used for 25-OHD was 100%hexane (mobile) and 15% water in methanol (stationary);the solvent system (System II) for the purification of la,25-(OH)2D was 10% ethyl acetate in hexane (mobile) and45% water in ethanol (stationary). Celite chromatographywas miniaturized to a "micro" scale (0.8 X 8.0 cm), whenrapid purification without separating D3 and D2 forms wasdesired (Fig. iB, D). 10-ml glass pipettes, fitted with glasswool plugs, serve as columns. Elution was carried out withmobile phase, and each sterol emerged between 7 and 22 mlon its respective column. When 12 samples were processedsimultaneously, -the total time from the initial drawing ofthe blood to a purified assayable sample is approximately2 h/sample per technician using the micro-silicic acid andmicro-Celite chromatography method (Fig. 1B, D). If it isnecessary to separate 25-OHD2 from 25-OHD3 and la,25-(OH)2D2 from 1a,25-(OH)2D3, a longer (1 X 35-cm) col-umn (with 5-ml fractions) was required (Fig. 1A, C). Aslow flow rate (15 min/fraction) produced defined peakswith minimal overlap. When sample purification on a suc-cessive series of Sephadex LH-20, silicic acid, and Celitechromatography was utilized, any possibili.ty of cross-con-tamination between the mono- and dihydroxy-metaboliteswas avoided (14). All samples for sterol assay were evap-orated under nitrogen and solubilized in 400 #1 ethanol. Afterdetermination of yield by counting an aliquot of each samplefor tritium (yields range from 40 to 80%), radioreceptorassays were performed on 100-lI portions. Triplicate assays

62 M. R. Hughes, D. J. Baylink, P. G. Jones, and M. R. Haussler

20 ml PLASMA

ADD 2000 cpm EACH [3H25-OH D3 and[!H] lcx,25-(OH)2D3

EXTRACT WITH MeOH-'CHCI3(2:1)

IX15cm SEPHADEXLH-20 COLUMN(CHCI 3-HEXANE,65:35)

25- OH-D

MICRO-SILICIC ACIDHEXANE-ETHER(1:1);X ETHER

LONG CELITESYSTEMI

25-OHD2 AND25-OHD3 ASSAYED

SEPARATELY

MICRO-CELITESYSTEMI

41TOTAL 25-OHD

ASSAYED

kx,25-(OH)2-D

MICRO-SILICIC ACIDETHER; ACETONE

/~~~~~"-LONG CELITE

SYSTEMr

lcr.25-(OH)2D2 ANDlo. 25-(OH)2D3 ASSAYED

SEPARATELY

©F

MICRO- CELITESYSTEMIr

TOTAL Io,25-(OH)2DASSAYED

FIGURE 1 Purification scheme for 25-OHD2 and(or) 25-OHD3 (A, B), and la,25-(OH)ID2and(or) la,25-(OH)2D3 (C, D). For a complete vitamin D metabolite profile of the plasma,pathways A and C are used. When distinctions between D2 and D3 metabolites are not required,the micro-Celite is employed as a final purification step (pathways B and D) and results areexpressed as the "total" D-metabolite concentration. All assays are run in triplicate. Extractionof plasma is carried out as described elsewhere (10). Celite solvent systems I and II andother details are described under Methods.

were carried out on all samples in the present study, andinterassay variation was 10-15%. All solvent used in chro-matographic procedures and sterol storage were reagentgrade and glass distilled before use.

Isolation of subcellular fractions. Intestinal mucosa (6 g)from two rachitic chicks was homogenized in 25 ml of 0.25M sucrose in 0.05 MTris-HCl (pH 7.4), 0.025 M KCI, and0.005 M MgC12. Cytosol fraction was obtained by centri-fugation at 100,000 g for 1 h. Chromatin was prepared fromcrude nuclei (isolated from original homogenate by centri-fugation at 1,000 g for 10 min) by homogenizing successivelyin one 25-ml portion of 0.8 mMEDTA, 25 mMNaCl, pH8; one 25-ml portion of lo Triton X-100 (Rohm & HaasCo., Philadelphia, Pa.), 0.01 M Tris-HCl, pH 7.5; and one25-ml portion of 0.01 M Tris-HCl, pH 7.5. The chromatinwas harvested by sedimentation at 30,000 g for 10 minafter each wash. The entire chromatin pellet from 6 g ofmucosa was reconstituted with the cytosol fraction byhomogenization to create a cytosol-chromatin receptor sys-tem for the competitive binding assay. The reconstitutedmixture was prepared fresh daily and immediately beforeuse. All operations were performed at 0-4oC.

Binding assay. To each assay tube containing [3H]la,25-(OH)2D3 and unlabeled sterol (dried together wi,th a streamof nitrogen) were added 20 ,l of distilled ethanol and 200

Al of reconstituted cytosol-chromatin system. The ethanolthat was added just before the receptor system aids in solu-bilizing the sterols to achieve a higher binding efficiencyand reduces nonspecific binding; this represents a modifi-cation of our earlier procedures (9, 10). The final concen-tration of [3H] 1a,25- (OH)2D3 was 4.3 nM. After incuba-tion, for 20 min at 250C, the quantity of labeled sterolbound to chromatin was determined by filtration (MilliporeCorp., Bedford, Mass., 30-place manifold). 1 ml of cold 1%Triton X-100 in 0.01 M Tris (pH 7.5) was added to eachassay tube, and the entire mixture was applied to a Type Aor A/E glass fiber filter (Gelman Instrument Co., AnnArbor, Mich.) at very low vacuum. After 2-4 min thevacuum was increased to achieve uniform flow rates ofapproximately 1 ml/min, and each of the filters was washedwith 2 ml of 1% Triton X-100, 0.01 M Tris, pH 7.5. Afterfiltration, the filters were placed in liquid scintillation vialswith 5 ml of methanol-chloroform (2:1, vol/vol). After 20min the methanol-chloroform was evaporated on a hot plateunder a stream of air. Without removal of the glass filter,sterols were solubilized in 5 ml of toluene counting solutionconsisting of 4 g of PPO and 50 mg of POPOPper literof toluene. Samples were counted in a refrigerated Beck-man LS-233 liquid scintillation spectrometer (BeckmanInstruments, Inc., Scientific Instruments Div., Irvine,

Assay for Active Vitamin D Metabolites in Plasma 63

0

xCPcr

z0

cr(-

0

0

InNC

6

0 IO 20

x2 "

zO0

05

tein does not discriminate between them but rather bindsla,25-(OH)2D3 and la,25-(OH)2D2 with approximatelythe same affinity.

In vitro generation of la,25-(OH) Ds. In the nextexperiment, a more direct approach was undertaken togenerate la,25- (OH) 2D2. Since the kidney is the solelocation of the la-hydroxylase enzyme that produces thehormonal form of the vitamin, renal tissue from 4-wk-old, severely rachitic chicks (12 animals, plasma Ca++< 5.5 mg%) was removed. Kidney homogenates wereincubated in a phosphate buffer and NADPHgeneratingsystem as previously described (2, 12). The reaction wasinitiated with 4,500 IU (292.5 nmol) crystalline 25-OHD2and was allowed to proceed under air at 37°Cfor 2 h; the generated dihydroxy-sterol was purified onsilicic acid, Sephadex LH-20, and Celite and finallyquantitated by ultraviolet absorption spectrophotometry.These identical conditions are routinely employed forthe production of la,25-(OH)2D3 from 25-OHD3; it isreasonable that the substrate 25-OHD2 will producela,25- (OH)2D2 in this system. Moreover, the sterol dis-played a uv spectrum with a minimum at 228 nm and amaximum at 265 nm which is characteristic of la,25-(OH)2D. In addition, the dihydroxy-metabolite migratedon Celite to the same position (fraction 14) as the invivo-produced sterol of the previous experiment.

Competitive binding of la,25-(OH)xDs and 25-OHD3.To verify that the target tissue receptor system efficientlybinds la,25- (OH)2D2, the in vitro-generated sterol wasadded to the reconstituted cytosol-chromatin system andallowed to compete with [3H] la,25-(OH)2D3 for the re-ceptor. Fig. 3A illustrates the competitive bindingcurves obtained when either la,25-(OH)2D3 or la,25-(OH)2D2 is the competing sterol. Results indicate thatla,25- (OH)2D3 is approximately 1.3 times more efficientthan the dihydroxyergocalciferol metabolite. Neverthe-less, the assay receptor system can bind and thereforedetect both the D2 and D3 forms of the hormone.

In the next experiment 25-OHD binding to the recep-tor was explored. Nonradioactive 25-OHD3 is approxi-mately 500 times less effective than nonradioactive la,25-(OH)2D3 in the ability to displace [3H] la,25-(OH)2D3from its intestinal receptor (18). Therefore, if cytosol-chromatin is incubated with the standard amount (360pg) of [3H] la,25-(OH)2D3, a 500-fold excess (180 ng)of 25-OHD3 will result in 50% competition at the recep-tor (Fig. 3B). Since the dihydroxy-D2 metabolite wascapable of binding (Fig. 3A), the monohydroxy-D2 me-tabolite (25-OHD2) was also tested for its competitiveability. Fig. 3B illustrates that nanogram quantities ofboth 25-OHD3 and 25-OHD2 are capable of competingwith picogram amounts of [3H] la,25-(OH)2D3 but that25-OHD3 is approximately 1.4 times more efficient. Thisis consistent with the results of Fig. 3A, indicating that

N'0

x

cx

w

U-

E0.

301 I

A

22 -

loi,25-(OH)2D2-14 -

6 - lcx,25- OH)2D3

I I I I I I 4j0 100 200 300 400 500 600 70

30

22I

14

6

lot,25-(OH)2D (pg)0

0 50 100 150 200 250 300 35025-OH D (ng)

FIGURE 3 Competitive binding standard curves for la,25-(OH)2D3 or la,25-(OH)2D2 (A), and 25-OHD3 or 25-OHD2 (B). A sample of [3H]la,25-(OH)2D3 (360 pg; 6.5Ci/mmol) was incubated with increasing amounts of non-radioactive sterol in the reconstituted cytosol-receptor sys-tem. The amount of bound tritiated compound (as deter-mined by filtration) is plotted as a function of the amount ofnonradioactive sterol in the incubation mixture. Each pointrepresents the average of quadruplicate assays±+SEM.

cholecalciferol metabolites bind slightly better thanergocalciferol metabolites. Therefore, standard competi-tive binding curves can be obtained for all four of thesterols of interest, and this allows for the measurementof these metabolites in plasma.

Isolation and measurement of plasma vitamin D me-tabolites. Before the circulating concentration of thesemetabolites can be measured by the competitive bindingmethod, the plasma sample must first be purified to sepa-rate the sterols of interest and remove contaminatinglipid. Fig. 1 is a schematic diagram depicting the puri-fication process and the chromatographic options thatare employed in the isolation of these biologically activevitamin D sterols from plasma. A single 20-ml plasmasample (with added radioactive tracers to quantitatepurification yields) is processed in such a way that the

Assay for Active Vitamin D Metabolites in Plasrma 65

I I I I,iB

-A £25-OHD2

2

- ~~25-OHD3

I I I I I ,

-I

p

I

25-hydroxymetabolites are separated from the la,25-dihydroxy metabolites on Sephadex LH-20 (Fig. 1).Each of the two groups are then purified on micro silicicacid columns using elution systems appropriate for theparticular metabolite (see Methods). Final resolution of25-OHD3 from 25-OHD2 and la,25-(OH)2Da fromla,25-(OH)2D2 is achieved by Celite liquid-liquid parti-tion chromatography on 1 X 35-cm columns (Fig. 1,pathways A and C). Subsequent radioreceptor assay ofthe separated sterols results in a complete metaboliteprofile of the four most biologically active forms of vita-min D in the circulation.

In many instances (see Discussion) it is not necessaryto distinguish between the Da and D3 metabolite concen-

NORM/

* 25-OHD3 Q I,,

200 32n

z0

o HYPERVITAMIN

25-OHD2 3(n 592 ng/ml [3H]25-OHD 5

4000-00 7Z 3

0z4 2000z

trations. Consequently, a simplified version of the puri-fication scheme has been developed that utilizes a "mi-cro" (0.8 X 8.0-cm) Celite column that does not resolveDa from Ds metabolites (Fig. 1, pathways B and D). Todemonstrate the inability of the micro-Celite system toseparate the two hormonal sterols, rats, which apparentlymetabolize Da and D3 equivalently, were raised on eithervitamin D2 or D3 in a similar manner to the experimentillustrated in Fig. 2. After purification (as outlined inFig. ID) the average la,25-(OH)2D in the D2- and Ds-raised rats was 18.8±2.0 ng/100 ml (SD) (n = 3; 30total rats) and 17.1±3.0 ng/100 ml (SD) (n=6; 60total rats), respectively. Thus, the micro-Celite chroma-tography system is necessary to remove contaminating

L

5-(OH)2D3ng/oom 04 1) jo

z 0

0.2Do

lo l0 ~ ~ 0)SIS D2 0 UL.

!5-(OH)2D2 [3H]I.25-(OH)2D3 Clnig/1OOM1

&~~~~~0.2 Cw 10cr E

0.10o

20 30 l0 20CELITE FRACTION NUMBER

FIGuRE 4 Celite resolution and radioligand receptor assay of D metabolites in normal sub-jects and in patient A with hypervitaminosis D. All Celite columns were 1 X 35 cm and 5-mlfractions were collected. The plasma from normals and a hypervitaminosis D patient waspurified as outlined in Fig 1A and C and as described under Methods. The plasma from thenormal subjects (four donors; 40 ml) yielded assayable 25-OHD3 (0), which migrates withauthentic [3H]25-OHD3 (0) in system I solvents (top left), and la,25-(OH)2D3, (A) whichmigrates with ['H] la,25-(OH)2D3 (0) in system II solvents (top right). The plasma (23ml) from the hypervitaminosis D patient (patient A; see Methods) yielded assayable materialthat does not migrate with authentic [3H]25-OHD3 in system I solvents (lower left), or ['H]-la,25- (OH)2D3 in system II solvents (lower right). These assayable peaks (fractions 12 and14, respectively) migrate to the exact positions of crystalline 25-OHD2 and in vitro generatedla,25- (OH)2D2 (see Results). Purification yields were as follows: pooled normal subjects =49%o 25-OHD3 and 56%o la,25-(OH)2D3; hypervitaminosis D2 subject= 76% 25-OHD3 and60% 1cr,25-(OH)2D3. In calculating the 25-OHD. and la,25-(OH)2D2 plasma concentrationsthe factors 1.4 and 1.3, respectively, are included since the assay standard curves are obtainedusing the respective D,3 metabolites (see text). Yields are assumed to be identical for D2 andDa forms.

66 M. R. Hughes, D. J. Baylink, P. G. Jones, and M. R. Haussler

lipid from la,25-(OH)2D and permit valid radioreceptorassay (10), but the column yields a mixture of la,25-(OH)2D2 and la,25-(OH)2D3. This column greatlysimplifies the final purification step before assay and al-lows for the routine processing of numerous samples toyield "total" 25-OHD and la,25-(OH)2D concentrationsin plasma.

With the use of this technique, the biologically activemetabolites of vitamin D have been measured in humans.Normal plasma la,25-(OH)2D in humans (78 subjects)is 3.3±0.6 ng/100 ml (SD) or approximately 0.1 nM.Utilizing 2 SD in either direction from the average, weobtain a normal range of 2.1-4.5 ng/100 ml for 95% ofall normals assayed. The concentration of 25-OHD innormal human plasma as measured by this assay is 25-40ng/ml and this value correlates well with data obtainedby other assay techniques (19-21).

Determination of D. vs. Di forms in the humiian andapplication of assay to hypervitaminosis D. The vita-min D metabolite assay, coupled with long Celite col-umns (Fig. 1, pathways A and C), was next used tomonitor the plasma concentrations of 25-OHD2, 25-OHD3, la,25- (OH) 2D2, and la,25-(OH) 2D3 in healthyhuman subjects and patients with varying degrees ofvitamin D intoxication (Figs. 4 and 5). All of the as-sayable 25-OHD in normal human plasma migrates withauthentic [3H] 25-OHD3 at a concentration of 32 ng/ml(Fig. 4, top left); 25-OHD2 was undetectable.2 Whenthe la,25- (OH) 2D concentration was examined (Fig. 4,top right) all of the assayable material migrated with[3H] la,25-(OH)2D3; 1a,25-(OH)2D2 was undetectable 2in these normal volunteers. A circulating level of 4.0 ng/100 ml was obtained when the column peak fractionswere assayed. Therefore, Celite preparative chromatog-raphy and the radioreceptor assay has allowed for themeasurement of both 25-OHD and la,25-(OH) 2D inhuman plasma and demonstrates that greater than 90% 2of the circulating concentration in the normal subject isin the form of cholecalciferol metabolites.

In contrast to these normal values, the assayed plasmafrom a hypervitaminosis D patient (Methods, subject A)contained 592 ng/ml 25-OHD (Fig. 4, lower left), andall of the assayable material migrated on the Celite col-umn (fractions 10-15) as 25-OHD2. Similarly, the onlymeasurable hormone concentration in the plasma of thishypervitaminosis D patient was also the ergocalciferolmetabolite; the la,25- (OH) 2D2 concentration was deter-

' The sensitivity of the dual assay is 17 pg of la,25-(OH)2D hormone (10) and 8.5 ng of precursor 25-OHD.With these values, if a 20-ml plasma sample is purified toa 50% sterol yield, it can be calculated that > 90% of thecirculating vitamin D metabolites in a normal (Tucson)human will be in the D3 form. This value correlates well withthe results of Haddad and Hahn (22) who measured 25-OHDin normal (St. Louis, Mo.) subjects.

mined to be 5.2 ng/100 ml of plasma and la,25-(OH)2D3was undetectable (Fig. 4, lower right). This value isonly slightly above the normal human range of 2.1-4.5ng/100 ml, whereas the assay value for 25-OHD of592 ng/ml represents more than a 15-fold increase overthe normal human subjects. The plasma from two addi-tional patients with postsurgical hypoparathyroidism,who were receiving large doses of vitamin D2 (patientsB and C) was also studied (Fig. 5). The circulatingconcentration of 25-OHD was elevated 10- to 16-foldabove normal values (Fig. 5, left panels) while the la,25-(OH) 2D concentration only slightly increased (Fig. 5,right panels). It is possible that the mass action effectof the elevated 25-OHD substrate on the renal 25-hy-droxy-lIc-hydroxylase may account for the slight increaseseen in the hormone's concentration. In any case, theresults in Figs. 4 and 5 indicate that vitamin D-intoxi-cated patients with intact kidneys maintain near normalhormone levels in the form of la,25-(OH)2D2, but dis-play highly elevated circulating concentrations of 25-OHD2.

DISCUSSIONElucidation of the chick intestinal cytoplasmic and nu-clear receptor system for la,25-(OH) 2D3 by Brumbaughand Haussler (6-8) has allowed for the development ofa competitive binding radioreceptor assay for this hor-mone (9, 10). The value obtained by this assay for thecirculating concentration of la,25- (OH) 2D in normalhumans (0.1 nM; see reference 10) has recently beenconfirmed by Hill et al. (23) using a bioassay. However,several questions remained regarding the applicabilityof the radioreceptor assay to the measurement of activevitamin D metabolites in animals and patients. Most no-table of these was the uncertainty as to whether la,25-(OH)a2D2 was also detected by the radioreceptor assay.Haddad and Hahn (22) have shown that patientstreated with therapeutic doses of vitamin D2 have a sig-nificant plasma concentration of 25-OHD2. Since manyhuman disorders such as hypoparathyroidism, renal os-teodystrophy, and vitamin D-resistant rickets are rou-tinely treated with ergocalciferol, the measurement ofla,25-(OH)2D2 is critical to the assessment of the vita-min D status in these cases. Because the binding systemutilized for the radioreceptor assay is derived from thechick intestine and the chick is known to discriminateagainst the D2 vitamins (16), it was essential to deter-mine directly the effectiveness of la,25-(OH) 2D2 bind-ing to the intestinal receptor. Furthermore, in the puri-fication scheme of Brumbaugh et al. (10) it was notknown whether la,25- (OH)2D2 was resolved from la,25-(OH)2D3. Thus, la,25-(OH)2D2 could have been eitherexcluded completely from measurement by the chroma-tography or quantitated in combination with the la,25-(OH)2D3.

Assay for Active Vitamin D Metabolites in Plasma 67

2, PATIENT B [ HJI.25-(oH)2D3)o .t , 30o25-OHD2 -OHD3 1.25-(OH)2D2

518rI ng |bng + 5.6ngio00mi1

)o - t *o ., ,I I,\l\ 'sIo 1., I

0 0

I-F

o'' PATII25-OHD2 I

319 ng/mi ' I

' [3H]25-OHD3

! I °I'I, I

I I1 ,~~~~~~~~~~~ _ ENT CI[3H]I.25-(OH)2D31,25-(OH)2D2 ,°o4.8ng-IOOmI ,d 'A III... A I I 40.6 z0U03

discrimination against D2 can be accounted for by theslightly lower affinity of D2 metabolites for the intestinalreceptor system.

Another extension of the radioreceptor assay for la,25-(OH)aD3 which is reported here is the application ofthis binding system to the measurement of circulating25-OHD3 and 25-OHD2. As reported previously (18),this receptor will bind 25-OHDa but only when thissterol is present in concentrations in 500-fold excess ofthat required for hormone binding. Although the com-petitive binding system is 500-fold less sensitive for25-OHD3 compared to la,25-(OH)aDa (Fig. 3), thisassay becomes possible because 25-OHD3 circulates inapproximately 1,000-fold greater concentrations thanla,25- (OH) 3D3 in man. Other workers have successfullyquantitated 25-OHD by competitive binding methodsutilizing several proteins. Belsey et al. (19) have re-ported a sensitive, competitive binding assay for mea-suring both vitamin D and 25-OHD which uses a spe-cific protein in rat serum.- More recently, this group hasdeveloped a rapid system (without chromatography) formeasuring 25-OHD in plasma(21). Haddad and Chyu(20) have published a sensitive competitive bindingassay for 25-OHD using a specific protein from rat kid-ney cytosol. Both of the latter assays are approximately40-fold more sensitive than the radioreceptor assay for25-OHD reported here, and they require considerablyless sample volume and preparation. However, the useof intestinal receptor to measure 25-OHD has certaindistinct advantages. For instance, 25-OHDs as well as25-OHD3 (after chromatographic resolution) can bequantitated separately with this binding protein sinceboth sterols bind with similar affinities (Fig. 3B). Yetthe most important consideration regarding this assaylies not in its insensitivity in measuring 25-OHD, but inits ability to quantitate the four major circulating chole-calciferol and ergocalciferol metabolites with a singlecompetitive binding system.

The ability to monitor these biologically active metab-olites in plasma may prove important in the understand-ing of many vitamin D-related diseases. Of particular in-terest in this regard is hypervitaminosis D or vitaminD intoxication. Brumbaugh and Haussler (25), basedupon the fact that 25-OHD3 (in higher concentrations)will bind to the la,25- (OH) 2D3 intestinal receptor, origi-nally hypothesized that 25-OHD may be the toxic agentin hypervitaminosis D. Recently, it has been reported byCounts et al. (26) that vitamin D intoxication and hy-percalcemia can occur in the anephric human; the se-rum concentration of 25-OHD in their patient was nearly28 times larger than the normal circulating concentra-tion of this metabolite. Since la,25- (OH)2D3 is producedexclusively in renal tissue (3), and since anephric hu-man subjects have undetectable lca,25-(OH)aD in theplasma (27), it is apparent that the hypercalcemia in

this patient was probably due to the markedly elevated25-OHD concentrations in the plasma. In the presentstudy (Figs. 4 and 5), we measured the 25-OHD in twohypervitaminosis D patients with intact kidneys and inone patient (normal calcium) being treated with a largedose of vitamin D2. Our results confirm the finding that25-OHD is considerably increased in D intoxication. Inaddition, we found that the la,25- (OH) 2D concentra-tion is near normal in these patients (Figs. 4 and 5)when the kidneys are present. Although la,25-(OH) 2D3is the physiologic hormone regulating calcium trans-location, a 100- to 500-fold greater concentration of pre-cursor 25-OHD3 will mediate (in vitro) bone calciumresorption (5) and intestinal receptor binding (25).Therefore, the causative agent in hypervitaminosis D isprobably the highly elevated 25-OHD, rather than theslightly elevated la,25-(OH)2D hormone. This conjec-ture is supported by the findings of Shen et al. (28)that patients with idiopathic hypercalciuria have simi-lar (slightly elevated) la,25- (OH)2 D levels as we pres-ently report for D-intoxicated patients, but are able tomaintain a normal serum calcium concentration. It shouldbe mentioned that these data do not rule out the possi-bility that some other vitamin D metabolite such as1,24,25-trihydroxyvitamin D (29) might be involved inthis disorder. Nevertheless, it seems reasonable to con-clude that the hypercalcemia seen in D intoxication isnot a result of abnormal plasma levels of la,25- (OH) 2D,but is caused by an excessive circulating concentrationof 25-OHD.

It is of interest that there are undetectable ergocal-ciferol metabolites in normal Tucson Caucasian adultswho undoubtedly ingest vitamin D2 in their diet. Haddadand Hahn (22), using a more sensitive assay for mea-suring 25-OHD in serum, have reported that St. Louisadults exhibit approximately 90% of their circulating25-OHD as the cholecalciferol metabolite. The presentreport also confirms the finding (22) that 25-OHD2 isthe predominant form when a patient is receiving largedoses of ergocalciferol, and it extends this work to in-clude the measurement of la,25- (OH) aD. It is possiblethat preferential metabolism of the two forms of vitaminD occur under different physiological conditions, how-ever further assessment of the metabolic disposition ofnatural and synthetic sources of vitamin D is required.

Vitamin D2 and D3 metabolites are not equivalent inthis receptor system, the D3 forms being 1.3-1.4 timesmore potent. Therefore, when rapid micro-Celite puri-fication is used, equivalent displacement of radioactivela,25- (OH)2D3 by a typical sample might represent onepart vitamin D3 metabolite, 1.3-1.4 parts vitamin D2 me-tabolite, or a variable combination of the two sterols.Nevertheless, meaningful clinical information has beenobtained utilizing this technique for total la,25- (OH)2Dmeasurement. Abnormal circulating levels of la,25-

Assay for Active Vitamin D Metabolites in Plasma 69

(OH)sD have been found in numerous human disordersincluding untreated hypoparathyroidism (15), pseudo-hypoparathyroidism (15), primary hyperparathyroidism(30), idiopathic hypercalciuria (28), and patients withchronic renal failure (9). Further clinical insights awaitrefinement and simplification of the radioligand receptorassay, especially the advent of higher specific activity[3 H] la,25- (OH)2Ds to allow for reduced size of plasmasamples and less arduous chromatographic purificationrequirements.

Note added in proof. We have recently confirmed thestructure of the in vitro produced la,25-(OH)2D2 used inthis study (see Results) by direct-probe mass spectrometry.

ACKNOWLEDGMENTS

The authors wish to thank Ms. Kristine Bursac for hertechnical assistance in this study. Special thanks go toDorothea E. Hellman, M. D. and Stanley Schneider, M. D.for supplying the plasma from their patients.

This investigation was supported by U. S. Public HealthService grants AM15781 and DE02600 and U. S. PublicHealth Service training grant GM01982.

REFERENCES

1. DeLuca, H. F. 1974. Vitamin D: The vitamin and thehormone. Fed. Proc. 33: 2211-2219.

2. Tucker, G., III, R. E. Gagnon, and M. R. Haussler. 1973.Vitamin Ds-25-hydroxylase: Tissue occurrence and ap-parent lack of regulation. Arch. Biochem. Biophys. 155:47-57.

3. Fraser, D. R., and E. Kodicek. 1970. Unique biosynthesisby kidney of a biologically active vitamin D metabolite.Nature (Lond.). 228: 764-766.

4. Haussler, M. R., D. W. Boyce, E. T. Littledike, andH. Rasmussen. 1971. A rapidly acting metabolite ofvitamin D3. Proc. Nati. Acad. Sci. 68: 177-181.

5. Raisz, L. G., C. C. Trummel, M. F. Holick, and H. F.DeLuca. 1972. 1,25-dihydroxycholecalciferol: a potentstimulator of bone resorption in tissue culture. Science(Wash. D. C.). 175: 768-769.

6. Brumbaugh, P. F., and M. R. Haussler. 1974. la,25-dihydroxycholecalciferol receptors in intestine. Tempera-ture-dependent transfer of the hormone to chromatin viaa specific cytosol receptor. J. Biol. Chem. 249: 1258-1262.

7. Brumbaugh, P. F., and M. R. Haussler. 1974. la,25-dihydroxycholecalciferol receptors in intestine. Associa-tion of la,25-dihydroxycholecalciferol with intestinalmucosa chromatin. J. Biol. Chem. 249: 1251-1257.

8. Brumbaugh, P. F., and M. R. Haussler. 1975. Specificbinding of la, 25-dihydroxycholecalciferol to nuclearcomponents of the chick intestine. J. Biol. Chem. 250:1588-1594.

9. Brumbaugh, P. F., D. H. Haussler, R. Bressler, andM. R. Haussler. 1974. Radioreceptor assay for la,25-dihydroxyvitamin D3. Science ( Wash. D. C.). 183:1089-1091.

10. Brumbaugh, P. F., D. H. Haussler, K. M. Bursac, andM. R. Haussler. 1974. Filter assay for la,25-dihydroxy-vitamin D3. Utilization of the hormone's target tissuechromatin receptor. Biochemistry. 13: 4091-097.

11. McNutt, K. W., and M. R. Haussler. 1973. Nutritionaleffectiveness of la,25-dihydroxycholecalciferol in pre-venting rickets in chicks. J. Nutr. 103: 681-689.

12. Haussler, M. R. 1972. Characterization of the metabo-lites of vitamin D3 in the chick. Steroids. 20: 639-650.

13. Lawson, D. E. M., D. R. Fraser, E. Kodicek, H. R.Morris, and D. H. Williams. 1971. Identification of la,25-dihydroxycholecalciferol. A new kidney hormone con-trolling calcium metabolism. Nature (Lond.). 230: 228-230.

14. Haussler, M. R., and H. Rasmussen. 1972. The metabo-lism of vitamin D3 in the chick. J. Biol. Chem. 247:2328-2335.

15. Haussler, M. R., D. J. Baylink, M. R. Hughes, P. F.Brumbaugh, J. E. Wergedal, F. H. Shen, R. L. Nielsen,S. J. Counts, K. M. Bursac, and T. A. McCain. 1976.The assay of la,25-dihydroxyvitamin D3: Physiologicand pathologic modulation of circulating hormone levels.Clin. Endocrinol. 5: 151s-165s.

16. Norman, A. W. 1968. The mode of action of vitamin D.Biol. Rev. (Camb.). 43: 97-137.

17. Hughes, M. R., P. F. Brumbaugh, M. R. Haussler, J.E. Wegedal, and D. J. Baylink. 1975. Regulation ofplasma la,25-dihydroxyvitamin Ds by calcium and phos-phate in the rat. Science (Wash. D. C.). 190: 578-580.

18. Hughes, M. R., P. F. Brumbaugh, and M. R. Haussler.1975. A dual assay for measuring the plasma concen-tration of la,25-dihydroxyvitamin D3 and 25-hydroxy-vitamin D3. Fed. Proc. 34: 893. (Abstr.)

19. Belsey, R. E., H. F. DeLuca, and J. T. Potts, Jr. 1971.Competitive binding assay for vitamin D and 25-OHvitamin D. I. Clin. Endocrinol. Metab. 33: 554-557.

20. Haddad, J. G., and K. J. Chyu. 1971. Competitive pro-tein-binding radioassay for 25-hydroxycholecalciferol. J.Clin. Endocrinol. Metab. 33: 992-995.

21. Belsey, R. E., H. F. DeLuca, and J. T. Potts, Jr. 1974. Arapid assay for 25-OH-vitamin D3 without preparativechromatography. J. Clin. Endocrinol. Metab. 38: 1046-1051.

22. Haddad, J. G., Jr., and T. J. Hahn. 1973. Natural andsynthetic sources of circulating 25-hydroxyvitamin D inman. Nature (Lond.). 244: 515-517.

23. Hill, L. F., E. B. Mawer, and C. M. Taylor. 1975.Determination of plasma levels of 1,25-dihydroxy-chole-calciferol in man. In Vitamin D and Problems Relatedto Uremic Bone Disease. A. W. Norman, K. Schaefer,H. G. Grigoleit, D. V. Herrath, and E. Ritz, editors.Walter De Gruyter & Co., Berlin. 755-762.

24. Belsey, R. E., H. F. DeLuca, and J. T. Potts, Jr. 1974.Selective binding properties of vitamin D transport pro-tein in chick plasma in zitro. Nature (Lond.). 247:208-209.

25. Brumbaugh, P. F., and M. R. Haussler. 1973. la,25-di-hydroxyvitamin D3 receptor: Competitive binding ofvitamin D analogs. Life Sci. 13: 1737-1746.

26. Counts, S. J., D. J. Baylink, F-H. Shen, D. J. Sher-rard, and R. 0. Hickman. 1975. Vitamin D intoxicationin an anephric child. Ann. Intern. Med. 82: 196-200.

27. Haussler, M. R. 1974. Vitamin D: Mode of action andbiomedical applications. Nutr. Rev. 32: 257-266.

28. Shen, F., D. Baylink, R. Nielsen, D. Sherrard, and M.Haussler. 1976. A study of the pathogenesis of idio-pathic hypercalciuria. Clin. Res. 24: 157a. (Abstr.)

29. Holick, M. F., A. Kleiner-Bossaller, H. K. Schnoes, P.M. Kasten, I. T. Boyle, and H. F. DeLuca. 1973. 1,24-25-Trihydroxyvitamin D3. A metabolite of vitamin Dseffective on intestine. J. Biol. Chem. 248: 6691-6696.

30. Haussler, M. R., K. M. Bursac, H. Bone, and C. Y. C.Pak. 1975. Increased circulating la,25-dihydroxyvitaminDs in patients with primary hyperparathyroidism. Clin.Res. 23: 322a. (Abstr.)

70 M. R. Hughes, D. J. Baylink, P. G. Jones, and M. R. Haussler