Establishment of enzyme immunoassay for measuring β-methyldigoxin levels in human serum by specific antiserum

Article abstract of DOI:10.1248/bpb.25.1251

We investigated the specificity of obtained antisera to β-methyldigoxin by the enzyme immunoassay. Three types of hapten-bovine serum albumin (BSA) conjugates were synthesized to obtain high specific antisera to β-methyldigoxin. The haptens were linked to

Full text of DOI:10.1248/bpb.25.1251

October 2002
Biol. Pharm. Bull. 25(10) 1251—1257 (2002)
1251
b
Establishment of Enzyme Immunoassay for Measuring -Methyldigoxin
Levels in Human Serum by Specific Antiserum
Yasuhiko HIGASHI,* Naomi WATANABE, Toshiyuki SASAKI, and Youichi FUJII
Department of Analytical Chemistry, Faculty of Pharmaceutical Sciences, Hokuriku University; Ho-3, Kanagawa-machi,
Kanazawa 920–1181, Japan. Received March 15, 2002; accepted July 17, 2002
We investigated the specificity of obtained antisera to b-methyldigoxin by the enzyme immunoassay. Three
types of hapten-bovine serum albumin (BSA) conjugates were synthesized to obtain high specific antisera to b-
methyldigoxin. The haptens were linked to the carrier protein through hemisuccinate at C-3؅ and C-3؆ positions
in the digitoxose chain and at C-12 position in the aglycone. Anti-b-methyldigoxin 3؅-hemisuccinate–BSA anti-
serum showed a low detection limit (0.2 ng/ml) and possessed high specificity for b-methyldigoxin, exhibiting low
cross-reactions with digoxigenin bisdigitoxoside (8.3%), dihydrodigoxin (4.8%), digitoxin (1.5%), and digoxigenin
monodigitoxoside (0.95%), except for cross-reaction with digoxin (43%). Compared with commercial anti-
digoxin antiserum, clinically used to monitor b-methyldigoxin concentration in human serum, cross-reaction
data of anti-b-methyldigoxin 3؅-hemisuccinate–BSA antiserum showed higher specificity for b-methyldigoxin.
The intra-assay and inter-assay variations using this antiserum were less than 6.9% and 8.1%, respectively. The
recovery tests were good, within the range of 96.2—104.3%. Phenyl boric acid (PBA) column treatment was ef-
fective to rapidly and selectively separate b-methyldigoxin from the mixture of b-methyldigoxin and its metabo-
lites in human serum. The recovery tests of b-methyldigoxin with PBA column in human serum were about
110% and interference of metabolites of b-methyldigoxin was negligible. These results suggest that anti-b-
methyldigoxin 3؅-hemisuccinate–BSA antiserum and PBA column treatment are useful to more precisely moni-
tor the unchanged type of b-methyldigoxin concentration in human serum.
Key words b-methyldigoxin; anti-b-methyldigoxin 3Ј-hemisuccinate–BSA antiserum; phenyl boric acid column; anti-digoxin
antiserum; enzyme immunoassay; cross-reactivity
Cardiac glycosides are clinically used as an important drug cific properties to MDx3 of three obtained antisera and com-
for treatment of congestive heart failure and atrial fibrilla- mercial anti-Dx3 antiserum in enzyme immunoassay (EIA).
tion, and therapeutic monitoring of them has been performed, Then, using a phenyl boric acid (PBA) column, which has
because their effective concentration in serum is very narrow. the ability to adsorb compounds possessing cis-diol group in
b-Methyldigoxin (MDx3) is a compound conjugating methyl the chemical structure,12,13) we carried out the pretreatment to
group at C-4ٞ position of digoxin (Dx3), and it is a more ab- separate MDx3 from mixtures of MDx3 and its metabolites
sorptive drug compared with Dx3 after p.o. administration.
in human serum and investigated the usefulness of a PBA
Immunoassay procedures have been used to monitor car- column to monitor the unchanged type of MDx3.
diac glycosides concentrations in serum. Recently, anti-Dx3
antibody has been clinically used to monitor MDx3 levels in MATERIALS AND METHODS
human serum.1—3) However, because administered MDx3 is
metabolized to mainly Dx3, digoxigenin bisdigitoxoside
Materials MDx3 and DihDx3 were obtained from
(Dx2), digoxigenin monodigitoxoside (Dx1), digoxigenin Boehringer Mannheim (Mannheim, Germany), spironolac-
(Dx0), and dihydrodigoxin (DihDx3) in the body,4—6) low tone, BSA (fraction V), anti-rabbit IgG antiserum developed
specificity of used antibodies is one of the most severe prob- in goat, and anti-Dx3 antiserum from Sigma Chemical Co.
lems for measuring MDx3. To obtain a highly specific anti- (St. Louis, MO, U.S.A.), Dx3 from Aldrich (Milwaukee, WI,
body, the binding position of hapten to carrier protein is a U.S.A.). Sephadex LH-20 were purchased from Pharmacia
key point. Shimada et al.7) and Thong et al.8) used conjugates Fine Chemicals (Uppsala, Sweden), Reversed-phase KC18
where the hapten was linked to the carrier protein at the C-12 plates (5ϫ10 cm) for TLC from Whatman (Clifton, NJ,
(or C-17) and C-22 positions, respectively. These antisera ex- U.S.A.), Kiesel-gel 60 for column chromatography and high
hibited remarkable cross-reactivity with DihDx3, one of the performance thin-layer chromatography (HPTLC) plates
metabolites. Previously, we reported preparation and anti- (5ϫ10 cm) from E. Merck (Darmstadt, Germany). Digitoxin,
genic properties of Dx3–bovine serum albumin (BSA) and b-D-galactosidase (EC 3.2.1.23) from Escherichia coli, and
digitoxin–BSA conjugates linked at the digitoxose C-3Ј and 4-methylumbelliferyl-b-D-galactopyranoside, and other gen-
C-3Љ positions using radioimmunoassay (RIA).9—11) Our pre- eral reagents were obtained from Wako Pure Chemical In-
vious results suggested that antisera to these conjugates dustries (Osaka, Japan). Dx0, Dx1, and Dx2 were prepared
linked at the digitoxose C-3Ј position showed higher speci- by hydrolysis of Dx3 according to the methods of Kaiser and
ficity to metabolites. However, a more convenient and simple co-workers.14) Dx0 12-hemisuccinate was synthesized by the
method is needed, because RIA uses radioactive substances method of Ikeda and Fujii.9) PBA column and VacElut Vac-
and a special facility.
In this paper, we prepared MDx3–BSA conjugates pos- U.S.A.) and GL Science (Tokyo, Japan), respectively.
sessing bridges at the hydroxyl groups of C-3Ј, C-3Љ, or C-12 Apparatus All melting points were determined with a
as binding positions in synthesizing antigens, and show spe- Yanagimoto micro hot-stage apparatus and are uncorrected.
uum Manifold were purchased from Varian (Harbor City,
* To whom correspondence should be addressed. e-mail: y-higashi@hokuriku-u.ac.jp
© 2002 Pharmaceutical Society of Japan

1252
Vol. 25, No. 10
Optical rotations were measured with a JASCO DIP-370 dig- (3H, s, 18-CH3), 0.93 (3H, s, 19-CH3), 1.24—1.37 (9H, m,
ital polarimeter. FAB-MS measurements were made on a sugar-CH3), 2.68 (4H, br s, –CO(CH2)2CO–), 3.40 (3H, s,
JEOL HX-100 instrument equipped with a FAB ion source –OCH3), 5.43 (1H, m, 3Љ-H), 5.94 (1H, s, 22-H).
using glycerol and NaCl as the matrix agents. UV spectra
MDx3 12-Hemisuccinate (3) To a solution of MDx3
were obtained with a Shimadzu UV-3000 recording spec- (1.0 g, 1.3 mmol) in pyridine (70 ml), succinic anhydride
1
trophotometer. H-NMR spectra were recorded using tetra- (4.0 g, 40 mmol) was added, and the mixture was allowed to
methylsilane as an internal standard on a JEOL EX-90A spec- stand at 60 °C for 73 h. The reaction mixture was treated in
trometer at 90 MHz. Abbreviations used: sϭsinglet, dϭdou- the same manner as described in 1. The AcOEt extract and
blet, and mϭmultiplet.
the MeOH eluate were combined and submitted to silica-gel
MDx3 3؅-Hemisuccinate (1) To a solution of MDx3 column (133ϫ2.5 cm i.d.) chromatography using CHCl3–
(1.2 g, 1.5 mmol) in pyridine (70 ml), succinic anhydride MeOH–H2O (92 : 8 : 0.5, v/v) as a mobile phase. The eluate
(1.2 g, 12 mmol) was added, and the mixture was allowed to corresponding to 3 was further purified on a Sephadex LH-
stand at 70 °C for 73.5 h. The reaction mixture was extracted 20 column (47ϫ1.5 cm i.d.) using MeOH as an eluent. The
with AcOEt, and the organic layer was washed with 1% HCl, eluate was recrystallized from acetone–hexane (1 : 2, v/v) to
1% NaHCO3, and H2O, then dried over anhydrous Na2SO4. give 3 (153 mg, 13.6%) as a colorless amorphous solid. mp
The aqueous layer was percolated through an Amberlite 157—160 °C. [a]D22 Ϫ103.4° (cϭ0.20, MeOH). Anal. Calcd
XAD-2 column (80ϫ1.6 cm i.d.). The column was washed for C46H70O17: C, 61.73; H, 7.88. Found: C, 61.37; H, 8.12.
with H2O, then the desired material, which could not be ex- FAB-MS m/z: 939 [MϪHϩ2Na]ϩ, 917 [MϩNa]ϩ, 899
tracted with AcOEt, was eluted with MeOH. The AcOEt ex- [MϪH2OϩNa]ϩ, 625 [digoxigenin monodigitoxoside 12-
tract and the MeOH eluate were combined and submitted to hemisuccinateϪH2OϩNa]ϩ, EI-MS m/z: 472 [digoxigenin
silica-gel column (133ϫ2.5 cm i.d.) chromatography using 12-hemisuccinateϪH2O]ϩ. UV lmax (MeOH) nm (e): 217
1
CHCl3–MeOH–AcOH (95 : 5 : 0.2, v/v) as a mobile phase. (13700). H-NMR (CDCl3) d: 0.89 and 0.92 (6H, s, 18-CH3
The eluate corresponding to 1 was further purified on a or 19-CH3), 1.22—1.30 (9H, m, sugar-CH3), 2.65 (4H, br s,
Sephadex LH-20 column (47ϫ1.5 cm i.d.) using MeOH as –CO(CH2)2CO-), 3.41 (3H, s, –OCH3), 5.97 (1H, s, 22-H).
an eluent. The eluate was recrystallized from acetone–hexane
Partial Hydrolysis of MDx3, 1, 2, or 3 To a solution of
(1 : 2, v/v) to give 1 (97 mg, 7.2%) as a colorless amorphous MDx3, 1, 2, or 3 (ca. 2.0 mg) in MeOH (0.52 ml), 0.05 M
solid. mp 151—154 °C. [a]D21 ϩ34.4° (cϭ0.25, MeOH). HCl (0.08 ml) was added, and the mixture was allowed to
Anal. Calcd for C46H70O17·1/2H2O: C, 61.11; H, 7.92. stand at 60 °C for 4 h. The reaction mixture was then ex-
Found: C, 61.15; H, 7.83. FAB-MS m/z: 939 [MϪHϩ tracted with AcOEt (3 ml) and the extract was washed with
2Na]ϩ, 917 [MϩNa]ϩ, 773. [digoxigenin bisdigitoxoside 3Ј- H2O. After evaporation of the solvent, the residue was spot-
hemisuccinateϩNa]ϩ, 755 [digoxigenin bisdigitoxoside 3Ј- ted on TLC plates.
hemisuccinateϪH2OϩNa]ϩ,. 643 [digoxigenin monodigitox-
MDx3 3؅-Hemisuccinate p-Nitrophenyl Ester (4) Di-
oside 3Ј-hemisuccinateϩNa]ϩ, 621 [digoxigenin monodigi- cyclohexylcarbodiimide (DCC, 43 mg, 0.21 mmol) was added
toxoside. 3Ј-hemisuccinateϩH]ϩ, 603 [digoxigenin monodig- to a solution of 1 (80 mg, 0.089 mmol) and p-nitrophenol
itoxoside 3Ј-hemisuccinateϪH2OϩH]ϩ. UV lmax (MeOH) (410 mg, 2.9 mmol) in dioxane (13 ml), and the mixture was
1
nm (e): 218 (12900). H-NMR (CDCl3) d: 0.80 (3H, s, 18- stirred at room temperature for 8 h. After evaporation of the
CH3), 0.92 (3H, s, 19-CH3), 1.21—1.27 (9H, m, sugar-CH3), solvent, the crude product obtained was submitted to silica-
2.67 (4H, br s, –CO(CH2)2CO–), 3.41 (3H, s, –OCH3), 5.38 gel column (47ϫ1.5 cm i.d.) chromatography using hexane–
(1H, m, 3Ј-H), 5.95 (1H, s, 22-H).
AcOEt (1 : 2, v/v) as a mobile phase, and 4 (61 mg, 67.2%)
MDx3 3؆-Hemisuccinate (2) To a solution of MDx3 was obtained as a yellow oil. 1H-NMR (CDCI3) d: 0.79 (3H,
(1.3 g, 1.6 mmol) in pyridine (90 ml), succinic anhydride s, 18-CH3), 0.93 (3H, s, 19-CH3), 1.21—1.27 (9H, m, sugar-
(1.2 g, 12 mmol) was added, and the mixture was allowed to CH3), 2.72—2.95 (4H, m, –CO(CH2)2CO–), 3.41 (3H, s,
stand at 60 °C for 68 h. The reaction mixture was treated in –OCH3), 5.39 (1H, m, 3Ј-H), 5.93 (1H, s, 22-H), 7.29, 8.27
the same manner as described in 1. The AcOEt extract and (each 2H, d, Jϭ9 Hz, aromatic H).
the MeOH eluate were combined and submitted to silica-gel
MDx3 3؆-Hemisuccinate p-Nitrophenyl Ester (5) DCC
column (133ϫ2.5 cm i.d.) chromatography using CHCl3– (26 mg, 0.12 mmol) was added to a solution of 2 (41 mg,
MeOH–H2O (92 : 8 : 0.5, v/v) as a mobile phase. The eluate 0.045 mmol) and p-nitrophenol (203 mg, 1.4 mmol) in diox-
corresponding to 2 was further purified on a Sephadex LH- ane (17 ml), and the mixture was stirred at room temperature
20 column (47ϫ1.5 cm i.d.) using MeOH as an eluent. The for 6 h. After evaporation of the solvent, the crude product
eluate was recrystallized from acetone–hexane (1 : 2, v/v) to obtained was submitted to silica-gel column (15ϫ0.7 cm i.d.)
give 2 (56 mg, 4.0%) as a colorless amorphous solid. mp chromatography using hexane–AcOEt (1 : 1, v/v) as a mobile
149—152 °C. [a]D21 ϩ37.3° (cϭ0.25, MeOH). Anal. Calcd phase, and 5 (20 mg, 44%) was obtained as a yellow oil. 1H-
for C46H70O17·1/2H2O: C, 61.73; H, 7.88. Found: C, 61.55; NMR (CDCI3) d: 0.79 (3H, s, 18-CH3), 0.92 (3H, s, 19-
H, 7.79. FAB-MS m/z: 939 [MϪHϩ2Na]ϩ, 917 [MϩNa]ϩ, CH3), 1.17—1.29 (9H, m, sugar-CH3), 2.82—2.93 (4H, m,
773. [digoxigenin bisdigitoxoside 3Љ-hemisuccinateϩNa]ϩ, –CO(CH2)2CO–), 3.40 (3H, s, –OCH3), 5.41 (1H, m, 3Љ-H),
755 [digoxigenin bisdigitoxoside 3Љ-hemisuccinateϪH2Oϩ 5.93 (1H, s, 22-H), 7.30, 8.27 (each 2H, d, Jϭ9 Hz, aromatic
Na]ϩ, 543 [digoxigenin monodigitoxosideϩNa]ϩ, 525 H).
[digoxigenin monodigitoxosideϪH2OϩNa]ϩ, 521 [digoxi-
MDx3 12-Hemisuccinate p-Nitrophenyl Ester (6) DCC
genin monodigitoxosideϩH]ϩ, 502 [digoxigenin monodigi- (76 mg, 0.37 mmol) in dioxane (40 ml) was gradually added
toxosideϪH2O]ϩ, 412 [digoxigeninϪHϩNa]ϩ. UV lmax to a solution of 3 (152 mg, 0.17 mmol) and p-nitrophenol
1
(MeOH) nm (e): 217 (12900). H-NMR (CDCl3) d: 0.80 (754 mg, 5.4 mmol) in dioxane (20 ml), and the mixture was

October 2002
1253
stirred at room temperature for 7 h. After evaporation of the
solvent, the crude product obtained was submitted to silica-
gel column (47ϫ1.5 cm i.d.) chromatography using hexane–
AcOEt (1 : 2, v/v) as a mobile phase, and 6 (83 mg, 48.1%)
1
was obtained as a yellow oil. H-NMR (CDCI3) d: 0.90 and
0.91 (6H, s, 18-CH3 or 19-CH3), 1.19—1.30 (9H, m, sugar-
CH3), 2.71—2.81 (4H, m, –CO(CH2)2CO–), 3.42 (3H, s,
–OCH3), 5.98 (1H, s, 22-H), 7.15, 8.27 (each 2H, d, Jϭ9 Hz,
aromatic H).
MDx3 3؅-Hemisuccinate-BSA (7), MDx3 3؆-Hemisucci-
nate–BSA (8), and MDx3 12-Hemisuccinate–BSA (9)
A
solution of BSA (50 mg) in 0.05 M phosphate buffer (pH 7.4,
0.5 ml) was added to a solution of 5 (40 mg, 0.039 mmol), 6
(20 mg, 0.020 mmol), or 7 (52 mg 0.051 mmol) in pyridine
(0.5 ml), and then these mixtures were stirred at room tem-
perature for 8 h to synthesize 7 or 8 and for 5 h to synthesize
9. The resulting solutions were dialyzed against a constant
flow of cold water (20 l) at room temperature overnight.
Lyophilization of these solutions afforded 7, 8, or 9 as a
fluffy powder.
Preparation of b-D-Galactosidase-Labeled Antigens
A solution of 4 (5.8 mg), 5 (6.0 mg), or 6 (2.0 mg) in dioxan
(0.5 ml) were incubated with a solution of b-D-galactosidase
(2.1 mg) in 1 ml of 0.05 M phosphate buffer (pH 9.0) for 4 h
Chart 1. Chemical Structures of Synthesized Compounds
at 4 °C. Then, the reaction mixtures were submitted to K2HPO4 (8.7 g), NaCl (9.0 g), NaN3 (1.0 g), MgCl2 (0.01 g),
Sephadex G-25 column (45ϫ1.5 cm i.d.) chromatography and BSA (1.0 g) in H2O (1000 ml). Standard samples (0.2 ml)
using 0.05 M phosphate buffer (pH 7.0) as a mobile phase. of MDx3 (0.2—20 ng/ml) and diluted antiserum (0.1 ml)
2.7 ml of fractions were collected, and fractions representing were mixed and incubated for 2.5 h at 4 °C. Then, 0.1 ml of
the main peak of the enzyme activity were further submitted synthetic enzyme–labeled antigens (diluted 1 : 1200 in phos-
to Sephadex G-100 column (45ϫ1.5 cm i.d.) chromatogra- phate saline buffer) was added and incubated for 0.5 h at
phy using 0.05 M phosphate buffer (pH 7.0) as a mobile 4 °C. 0.2 ml of goat anti-rabbit IgG antiserum (1.67%) and
phase. Fractions representing the main peak of the enzyme 0.1 ml of normal rabbit serum (1%) were added and allowed
activity were chosen as a labeled antigen in EIA. On the to stand for 12 h at 4 °C. Mixtures were centrifuged at
other hand, each fraction was spotted on TLC plates using 1600ϫg for 20 min, the supernatant was aspirated, and the
CH3Cl–MeOH–AcOH (90 : 10 : 0.8, v/v) as a developing sol- immune precipitate was twice washed by 1 ml of phosphate
vent, and coloration for steroid compounds at origin was ob- saline buffer. The activity of enzyme conjugate bound to
served by spraying with concentrated H2SO4. From these re- each tube was then measured by the addition of 0.5 ml of
sults, three enzyme-labeled antigens (MDx3 3Ј-, 3Љ-, and 12- 4-methylumbelliferyl-b-D-galactopyranoside (2.0ϫ10Ϫ5 M),
hemisuccinate–b-D-galactosidase) were obtained.
followed by incubation of the tubes at 30 °C for 3 h after pre-
Immunization Procedure Three domestic strains of incubation of 0.5 ml of phosphate saline buffer at 30 °C for
male albino rabbits were used for immunization with each 3 min. The enzyme reaction was stopped by addition of
hapten–BSA conjugate. The antigen (3 mg) was dissolved in 2.0 ml of glycine–NaOH buffer (0.1 M, pH 10.3) to each tube,
sterile isotonic saline (0.9 ml) and emulsified with complete and the resulting 4-methylumbelliferone was measured by
Freund’s adjuvant (2.1 ml). The emulsion was injected into spectrofluorometry at wavelengths of 362 nm for excitation
multiple subcutaneous sites along both sides of the back of and 446 nm for emission using a fluorescence spectropho-
each rabbit. The procedure was repeated at intervals of two tometer (F-2000, Hitachi, Tokyo, Japan).
weeks for five months and then the rabbits were boosted
Cross-Reaction Study The antisera raised against 7, 8,
twice a month. After confirmation of the increase in the anti- and 9, and commercial anti-Dx3 antiserum were abbreviated
body titer, blood was collected from the marginal ear veins. to Antiserum-A, Antiserum-B, Antiserum-C, and Antiserum-
Sera were separated by centrifugation at 1400ϫg for 10 min D, respectively. The specificities of these antisera were tested
and stored at Ϫ18 °C in small aliquots. The antisera were by calculating the percentage cross-reactivity with various
thawed, diluted with phosphate saline buffer (pH 7.4) and compounds. Cross-reactivity was determined by the above-
used in the assay.
mentioned assay procedure, by comparing the concentrations
EIA Method EIA is based on the principle of competi- of MDx3 and test compounds necessary for a 50% displace-
tion between enzyme-labeled and unlabeled drugs for an an- ment of the antibody-bound enzyme-labeled MDx3.
tibody, followed by measurement of the marker enzyme ac-
Extraction of MDx3 and Its Metabolites in Human
tivity of the immune precipitate. EIA was performed in phos- Serum 0.2 ml of human serum containing MDx3, Dx3,
phate saline buffer (pH 7.3). Phosphate saline buffer was Dx2, Dx1, Dx0, and DihDx3, 0.1 ml of boric acid (10 mM),
adjusted to pH 7.3 by addition of a solution containing and 0.1 ml of NH4OH (0.3%) were mixed and the mixtures
NaH2PO4 (3.9 g), NaCl (4.5 g), NaN3 (0.5 g), MgCl2 (0.005 g), were extracted into 2 ml of AcOEt at two times. Then, each
and BSA (0.5 g) in H2O (500 ml) to a solution containing 1.8 ml of AcOEt phase was combined, evaporated with a

1254
Vol. 25, No. 10
concentrator (TAITEC, Saitama, Japan), and 1 ml of CH3CN
was added to the residue.
Separation of MDx3 Using PBA Column Separation
procedures were as follows. 1 ml of HCl (0.1 M) was added to
the column, which was allowed to completely run through
with VacElut (GL Science, Tokyo, Japan). Then, 4 ml of
NH4OH (1.0%), 2 ml of (NH4)2SO4 (0.01 M, pH 8.5), and
2 ml of NH4OH (0.3%) were passed through the column, suc-
cessively. The above CH3CN solution was added to the col-
umn, allowed to completely run at gravity, and the column
was extracted four times with 1 ml of CH3CN. About 4 ml of
collected CH3CN solution was evaporated with a concentra-
tor, and was used in the assay.
Fig. 1. Reversed-phase TLC Chromatogram of the Products Obtained by
Partial Hydrolysis of MDx 3Ј-, 3Љ-, and 12-Hemisuccinates
RESULTS AND DISCUSSION
Plate, Whatman KC18; developing solvent, MeOH–0.5 M NaCl (18 : 13, v/v); visual-
ization, spraying with concentrated sulfuric acid followed by heating in an oven at
120 °C for 10 min. 1, hydrolysate of MDx3; 2, hydrolysate of MDx3 3Ј-hemisuccinate;
3, hydrolysate of MDx3 3Љ-hemisuccinate; 4, hydrolysate of MDx3 12-hemisuccinate;
5, Dx0 12-hemisuccinate. a, hemisuccinates; b, Dx0; c, Dx1; d, Dx2; e, MDx3.
Initially, the synthesis of MDx3 3Ј-, 3Љ-, and 12-hemisuc-
cinate–BSA conjugates (7—9) was carried out. In the previ-
ous papers of this series, we reported the preparation of digi-
toxin 3Ј- and 3Љ-hemisuccinates from digitoxin and Dx3 3Ј-
and 3Љ-hemisuccinates from Dx3 by utilizing the reactivities
of the hydroxyl groups in the digitoxose moiety.9—11) In a
similar way, MDx3 3Ј-, 3Љ-, and 12-hemisuccinates (1—3)
could be prepared from MDx3 with succinic anhydride in
pyridine. The resulting crude product was submitted to silica-
gel column chromatography using a CHCl3–MeOH–AcOH
(95 : 5 : 0.2, v/v) or CHCl3–MeOH–H2O (92 : 8 : 0.5, v/v) as
an eluent, and the chromatography provided a satisfactory
separation of three isomeric hemisuccinates. In the 1H-NMR
spectra of 1, 2, and 3, the four-methylene protons appeared
as a broad singlet, showing that one succinyl moiety had
been introduced.
The position of the succinyl group was elucidated by par-
tial hydrolysis of these compounds. Reversed-phase TLC of
their hydrolyzate is shown in Fig. 1. When MDx3 was treated
with 0.05 M HCl under mild conditions, the glycosidic bond
was partially hydrolyzed to give a mixture of Dx0, Dx1, Dx2,
and the intact starting material. When 1 and 2 were treated in
Fig. 2. Standard Curves of MDx3 by Homologous Assay Using Anti-
MDx3 3Ј-Hemisuccinate–BSA Antiserum (Antiserum-A), Anti-MDx3 3Љ-
Hemisuccinate–BSA Antiserum (Antiserum-B), and Anti-MDx3 12-
Hemisuccinate–BSA Antiserum (Antiserum-C)
Antiserum-A (᭹), Antiserum-B (᭝), Antiserum-C (ᮀ).
the same way, Dx0 12-hemisuccinate was not produced from serum-A, Antiserum-B, and Antiserum-C bound approxi-
1 and 3, Dx0 was only produced from 1 and both Dx0 and mately 50% of MDx3 at final dilutions of 1 : 1500, 1 : 15000,
Dx1 from 2. Therefore, 1 and 2 were designated as the 3Ј- 1 : 2000, respectively. The plots of percent bound fluores-
hemisuccinate and 3Љ-hemisuccinate of MDx3, respectively. cence intensity vs. logarithm of the concentration of non-la-
When 3 was treated in a similar fashion, Dx0 12-hemisucci- beled MDx3 showed a linear relationship at range 0.2 to
nate was only produced from 3. Therefore, 3 was designated 20 ng/ml using Antiserum-A. However, Antiserum-B and
as the 12-hemisuccinate of MDx3. The results of FAB-MS of Antiserum-C showed a linear relationship over range 0.5 to
three compounds also supported these structures.
10 ng/ml. It was shown that the detection limit of Antiserum-
To couple the hapten with BSA, 1, 2, and 3 were trans- A was the best of three obtained antisera.
formed into their p-nitrophenyl esters (4—6) by treatment
The specificities of Antiserum-A, Antiserum-B, and Anti-
with p-nitrophenol and DCC in dioxane. These p-nitrophenyl serum-C were assessed by cross-reaction tests with various
esters were linked with BSA, followed by dialysis of reaction related compounds. The percentage cross-reactivity was cal-
mixture against cold water to give the desired BSA conju- culated at 50% displacement of the antibody-bound labeled
gates (7—9).
MDx3, and the results are listed in Table 1. Antiserum-A
Each immunogen thus obtained was administered to three possessed high specificity, exhibiting low cross-reactions
rabbits to produce the antibody. Among the antisera elicited with Dx2 (8.3%), DihDx3 (4.8%), and digitoxin (1.5%), ex-
in rabbits, the most specific antiserum to MDx3 was selected cept for cross-reaction with Dx3 (43%). Also, there was no
for characterization in each group. The properties of antisera significant cross-reaction with Dx1 (0.95%). All other com-
were investigated by EIA with MDx3 hemisuccinate–b-D- pounds tested showed negligible values of Ͻ0.05%. In con-
galactosidase. The separation of bound and free fractions was trast, Antiserum-B exhibited considerable cross-reactivity
performed using formation of an immunoprecipitate.
with Dx3 (82%), Dx2 (65%), Dx1 (55%), Dx0 (58%), and
The standard curves of MDx3 using homologous assay by digitoxin (36%) and low cross-reaction with DihDx3 (4.7%).
the three obtained antisera are presented in Fig. 2. Anti- Antiserum-C showed considerable cross-reactivity with Dx3

October 2002
1255
Table 1. Cross-reaction Data for EIA of Antiserum-A, Antiserum-B, Anti-
serum-C, and Antiserum-D
% Cross-reactivity (50%)
Antiserum-A Antiserum-B Antiserum-C Antiserum-D
MDx3
Dx3
Dx2
Dx1
100
43
8.3
0.95
Ͻ0.05
4.8
100
82
65
55
58
100
39
Ͻ0.05
Ͻ0.05
Ͻ0.05
35
100
83
103
95
63
3.7
0.55
Ͻ0.05
Ͻ0.05
Ͻ0.05
Ͻ0.05
Ͻ0.05
Dx0
DihDx3
Digitoxin
Spironolactone
Digitoxose
Progesterone
Teststerone
Cholesterol
4.7
36
1.5
41
Ͻ0.05
Ͻ0.05
Ͻ0.05
Ͻ0.05
Ͻ0.05
Ͻ0.05
Ͻ0.05
Ͻ0.05
Ͻ0.05
Ͻ0.05
Ͻ0.05
Ͻ0.05
Ͻ0.05
Ͻ0.05
Ͻ0.05
Fig. 3. Standard Curves of MDx3 Using Antiserum-A and Commercial
Anti-Dx3 Antiserum (Antiserum-D)
Antiserum-A (᭹), Antiserum-D (᭺).
Antiserum-D which bound approximately 50% of MDx3 at a
final dilution of 1 : 500 showed a linear relationship in the
range 0.5 to 5 ng/ml, while Antiserum-A showed a linear re-
lationship in the range 0.2 to 20 ng/ml. These results show
that Antiserum-D has poorer sensitivity and measurable
range to MDx3 as compared with those of Antiserum-A.
The percentage of cross-reactivity was calculated at 50%
displacement of the antibody-bound labeled MDx3, and the
results are listed in Table 1. Antiserum-D possessed much
lower specificity to MDx3 compared with antiserum-A, ex-
hibiting high cross-reactions with Dx3 (81%), Dx2 (103%),
Dx1 (95%), and Dx0 (63%), and low cross-reaction with
DihDx3 (3.7%). All other compounds tested showed negligi-
ble values of Ͻ0.05%. Antiserum-D showed high cross-reac-
tivity with metabolites formed by the successive cleavage of
the digitoxose residues, indicating that Antiserum-D is much
inferior to Antiserum-A in monitoring of unchanged type of
MDx3, because Antiserum-D is antiserum produced by per-
oxidate-oxidized Dx3–BSA conjugate as an antigen at the
position of terminal digitoxose.3)
As shown in Table 2, we investigated the interference of
MDx3 metabolites in human serum by using Antiserum-A
and Antiserum-D. Each metabolite (Dx3, Dx2, Dx1, Dx0,
and DihDx3) at ratios of 10%, 30% and 100% to MDx3 con-
centration were added to the MDx3 in human serum, and the
recovery test was performed. The recovery ratios using Anti-
serum-A at 10%, 30%, and 100% were in the range of 103—
110%, 128—146%, and 175—291%, respectively. In con-
trast, those using Antiserum-D at 10%, 30%, and 100% were
Ͼ146%, Ͼ226%, and Ͼ493%, respectively. These results in-
dicate that the interference of MDx3 metabolites by using
Antiserum-A in human serum was much lower than that of
Antiserum-D, because of its low specificity to MDx3.
Values are calculated on a molar basis.
(39%), DihDx3 (35%), and digitoxin (41%). Methyl group at
C-4ٞ of MDx3 may be poor at recognizing the obtained anti-
bodies because of their very small molecular size. Another
discussion on the cross-reaction study is that, among Anti-
serum-A, Antiserum-B, and Antiserum-C, Antiserum-B ex-
hibited about two times higher cross-reactions with Dx3
(82%). The binding position of BSA of 8 is the nearest to the
methyl group at C-4ٞ of MDx3 of 7, 8, and 9. For this rea-
son, Antiserum-B may be dependent on the distance between
methyl group at C-4ٞ of MDx3 and the binding position of
BSA. It is evident from the cross-reaction data that the anti-
body produced by an antigen in which the hapten is linked to
a carrier protein through the digitoxose C-3Ј position, remote
from the terminal digitoxose and steroid nucleus, is able to
recognize both the sugar moiety and the lactone ring of
MDx3. From the present results and the previous study in-
volving anti-digitoxin antiserum and anti-Dx3 antiserum, the
C-3Ј position seems to be a favorable site for the attachment
of BSA on the cardiac glycoside in the production of the
specific antisera.9—11) However, the antiserum produced by
MDx3–BSA conjugate possessing a bridge at the C-3Љ posi-
tion near the terminal digitoxose is inferior in terms of speci-
ficity with respect to changes in the digitoxose part. The anti-
serum produced by the MDx3–BSA conjugate possessing a
bridge at C-12 position of steroid nucleus could not recog-
nize the lactone ring part. These results suggest that the Anti-
serum-A possessed the highest specificity for MDx3 of these
antisera, and Antiserum-A is useful to monitor MDx3 in
human serum.
To investigate the precision of EIA, we performed intra-
and inter-assay (nϭ7) in human at 0.5, 1, 2, and 5 ng/ml by
homologous assay using Antiserum-A which showed the
highest selectivity to MDx3. In intra-assay, the range of coef-
ficient variation to the average was within 4.5 to 6.9%, and
the recovery was within 96.2 to 103.8%. In inter-assay, the
range of standard deviation was 6.0—8.1%, and the recovery
was 97.1—104.3%. These data shows that the precision of
EIA is satisfactory. On the basis of these results, Antiserum-
D (commercial anti-Dx3 antiserum) was adopted as the stan-
dard to compare the specific properties of Antiserum-A.
The standard curves of MDx3 using Antiserum-A and An-
tiserum-D are presented in Fig. 3. The standard curve using
Antiserum-A showed much lower cross-reactivity in com-
parison with Antiserum-D, indicating that Antiserum-A may
be used for more precise determination of intact MDx3 in
human serum. However, it is considered that these low cross-
reactivities may result in error in the monitoring of MDx3.
Removal of cross-reactive compounds in human serum is
a good method. Recently, the PBA column has been used
for the rapid and selective isolation of urinary catec-
holeamine.12,13) Dx3, Dx2, Dx1, and DihDx3, which showed
cross-reactivity (4.8—43%), possess a cis-diol group at the
terminal glycoside in the structures, while MDx3 does not.
We tested rapid and selective separation of MDx3 from mix-

1256
Vol. 25, No. 10
Table 2. Interference of MDx3 Metabolites on the Assay of MDx3 in Human Serum
Antiserum-A
Antiserum-D
Measured (ng/ml)
Added
Measured (ng/ml)
Recovery
(%)
Recovery
(%)
meanϮS.D. (nϭ7)
meanϮS.D. (nϭ7)
MDx3 (0.5 ng/ml)
ϩMetabolites (each 0.05 ng/ml)
ϩMetabolites (each 0.15 ng/ml)
ϩMetabolites (each 0.5 ng/ml)
MDx3 (1 ng/ml)
0.55Ϯ0.03
0.68Ϯ0.06
0.88Ϯ0.07
110
136
175
0.73Ϯ0.06
1.20Ϯ0.10
2.47Ϯ0.21
146
240
493
ϩMetabolites (each 0.1 ng/ml)
ϩMetabolites (each 0.3 ng/ml)
ϩMetabolites (each 1 ng/ml)
MDx3 (2 ng/ml)
1.07Ϯ0.06
1.36Ϯ0.12
2.09Ϯ0.16
108
136
209
1.48Ϯ0.11
2.26Ϯ0.20
Ͼ5.0
148
226
(—)
ϩMetabolites (each 0.2 ng/ml)
ϩMetabolites (each 0.6 ng/ml)
ϩMetabolites (each 2 ng/ml)
MDx3 (5 ng/ml)
2.06Ϯ0.14
2.92Ϯ0.24
4.74Ϯ0.40
103
146
237
3.85Ϯ0.34
Ͼ5.0
Ͼ5.0
192.5
(—)
(—)
ϩMetabolites (each 0.5 ng/ml)
ϩMetabolites (each 1.5 ng/ml)
ϩMetabolites (each 5 ng/ml)
5.17Ϯ0.34
6.39Ϯ0.55
14.6Ϯ1.21
103
128
291
Ͼ5.0
Ͼ5.0
Ͼ5.0
(—)
(—)
(—)
Metabolites: Dx3, Dx2, Dx1, Dx0, and DihDx3.
Table 3. Recovery of MDx3 in Phosphate Buffer and Human Serum by
Homologous Assay Using Antiserum-A after PBA Column Treatment
In phosphate buffer (nϭ9)
Measured (ng/ml)
meanϮS.D.
Added (ng/ml)
C.V. (%)
Recovery (%)
0.5
1
2
0.51Ϯ0.05
0.97Ϯ0.08
2.06Ϯ0.19
5.20Ϯ0.50
9.4
8.2
9.3
9.7
102
97.3
103
104
5
In human serum (nϭ9)
Fig. 4. Rapid and Selective Separation of MDx3 from Mixtures of MDx3
and Its Metabolites in Human Serum Using PBA Column Treatment
Measured (ng/ml)
meanϮS.D.
Added (ng/ml)
C.V. (%)
Recovery (%)
Merck HPTLC (silica-gel); developing solvent, CHCl3–MeOH–H2O (85 : 15 : 1, v/v);
visualization, spraying with concentrated sulfuric acid followed by heating in an oven at
120 °C for 10 min. 1, mixtures of MDx3, Dx3, Dx2, Dx1, Dx0, DihDx3; 2, PBA col-
umn treatment by CH3CN (1 ml); 3, PBA column treatment washed by CH3CN (1 ml);
4, PBA column treatment washed by CH3CN (1 ml); 5, PBA column treatment washed
by CH3CN (1 ml); 6, PBA column treatment washed by CH3OH (1 ml). a, MDx3; b,
Dx0; c, Dx1; d, Dx2; e, Dx3; f, DihDx3.
0.5
1
2
0.54Ϯ0.04
1.12Ϯ0.11
2.20Ϯ0.20
5.61Ϯ0.53
9.7
9.3
9.0
9.5
108
112
110
112
5
tures of MDx3 and its metabolites in human serum using a
About 4 ml of CH3CN solution developed in the PBA col-
PBA column. Adding whole CH3CN solution to a PBA col- umn were collected and evaporated in vacuo. Then, MDx3
umn after extraction of the human serum with AcOEt, evapo- was measured by homologous EIA using Antiserum-A after
ration of the extracted organic layer, and addition of 1 ml of PBA column treatment in phosphate buffer and human serum
CH3CN to a test tube, we could detect MDx3 only on was investigated. The data are listed in Table 3. Recovery
HPTLC as shown in Fig. 4. By the second addition of 1 ml of rates in phosphate buffer and human serum were about 100%
CH3CN to PBA column, we could detect MDx3 and Dx0 on and about 110%, respectively. As shown in Table 4, we inves-
HPTLC. By the third addition, we could detect MDx3 and tigated the effect of PBA column treatment on the interfer-
Dx0 on HPTLC. However, using the same procedure 4 times, ence of metabolites. Each metabolite (Dx3, Dx2, Dx1, Dx0,
nothing could be detected on HPTLC. Dx3, Dx2, Dx1, and and DihDx3) at an equal concentration with the MDx3 con-
DihDx3 could be detected on HPTLC by washing the col- centration was added to the MDx3 in human serum, and the
umn with MeOH after eluent of CH3CN solution at 4 times, recovery test was performed. The recovery ratios at non-PBA
indicating that these compounds interacted with boric acids column treatment and PBA column treatment were in the
in the PBA column because they possess a cis-diol group. range of and 175—291% and 107—114%, respectively
These results suggest that MDx3 and Dx0, which did not These data indicated that no significant interference by
possess a cis-diol group in the molecular structure, can be se- metabolites of MDx3 in our EIA was confirmed.
lectively separated from the mixture of MDx3 and its
metabolites in human serum by PBA column treatment.
When PBA column was used for the pretreatment of
human serum sample containing Dx3, Dx2, Dx1, Dx0, and

October 2002
1257
Table 4. Interference of MDx3 Metabolites on the Assay of MDx3 by Homologous Assay Using Antiserum-A in Human Serum after PBA Column Treat-
ment
Non-PBA Treatment
PBA Treatment
Added
Measured (ng/ml)
meanϮS.D. (nϭ7)
C.V.
(%)
Recovery
(%)
Measured (ng/ml)
meanϮS.D. (nϭ7)
C.V.
(%)
Recovery
(%)
MDx3 (0.5 ng/ml)
ϩMetabolites (each 0.5 ng/ml)
MDx3 (1 ng/ml)
ϩMetabolites (each 1 ng/ml)
MDx3 (2 ng/ml)
ϩMetabolites (each 2 ng/ml)
MDx3 (5 ng/ml)
ϩMetabolites (each 5 ng/ml)
0.88Ϯ0.07
2.09Ϯ0.16
4.74Ϯ0.40
14.6Ϯ1.21
8.0
7.7
8.4
8.3
175
209
237
291
0.54Ϯ0.05
1.10Ϯ0.10
2.28Ϯ0.21
5.62Ϯ0.48
9.3
9.1
9.2
8.5
107
110
114
112
Metabolites: Dx3, Dx2, Dx1, Dx0, and DihDx3.
DihDx3, MDx3 and Dx0 were separated (Fig. 4). Antiserum- REFERENCES
D showed high cross-reactivity with Dx0 (63%), while Anti-
1) Valdes R., Jr., Brown B. A., Graves S. W., Am. J. Clin. Pathol., 82,
210—213 (1984).
2) Wahyono D., Piechaczyk M., Mourton C., Bastide J. M., Pau B., Hy-
bridoma, 9, 619—629 (1990).
3) Miller J. J., Straub R. W., Jr., Valdes R., Jr., Clin. Chem., 40, 1898—
1903 (1994).
4) Gault M. H., Longerich L. L., Loo J. C. K., Ko P. T. H., Fine A., Vas-
dev S. C., Dawe M. A., Clin. Pharmacol. Ther., 35, 74—82 (1984).
5) Hinderling P. H., Garrett E. R., Wester R. C., J. Pharm. Sci., 66, 242—
253 (1977).
6) Hinderling P. H., Garrett E. R., Wester R. C., J. Pharm. Sci., 66, 314—
325 (1977).
7) Shimada K., Mizusawa S., Ohkubo T., Nambara T., Chem. Pharm.
Bull., 32, 2301—2306 (1984).
8) Thong B., Soldin S. J., Lingwood C. A., Clin. Chem., 31, 1625—1631
(1985).
serum-A showed negligible value of Ͻ0.05%. It is consid-
ered that EIA by Antiserum-D using PBA column shows the
significant measurement error. In addition, the measurement
range of EIA by Antiserum-A (0.2—20 ng/ml) is much sen-
sitive compared with that of Antiserum-D (0.5—5 ng/ml). In
these points, Antiserum-A of our novel specific antiserum is
superior to Antiserum-D.
Consequently, our results suggest that unchanged type of
MDx3 concentration in human serum after administration
will be more precisely monitored using both Antiserum-A
and PBA column treatment. We expect that our newly devel-
oped assay system for MDx3 will be clinically popularized in
the world in the near future.
9) Ikeda Y., Fujii Y., Biol. Pharm. Bull., 23, 906—910 (2000).
10) Fujii Y., Ikeda Y., Yamazaki M., Clin. Chem., 40, 172—173 (1994).
11) Fujii Y., Ikeda Y., Fujii M., Yamazaki M., Biol. Pharm. Bull., 17,
467—471 (1994).
Acknowledgements This work was supported in part by
the Special Research Fund of Hokuriku University.
12) Dimson P. A., Analytichem. International (Current), 3, 4—5 (1984).
13) Alan H. B. W., Terrie G. G., Clin. Chem., 31, 298—302 (1985).
14) Haack E., Kaiser F., Spingler H., Naturwissenschaften, 44, 633—634
(1957).