Determination of Individual Bile Acids in Serum by High Performance Liquid Chromatography ~~

~~

Ganfeng Wangt, and Neil1 H. Stacey* National Institute of Occupational Health and Safety, The University of Sydney, NSW 2006, Australia

John Earl Royal Alexandra Hospital for Children, The University of Sydney, NSW 2006, Australia

A high performance liquid chromatographic (HPLC) method for analysis of 4 free and 8 conjugated bile acids in submicromolar quantities in serum is described using precolumn derivatization with 4-bromomethyl-7-methoxycoumarin (BMC) and fluorescence detection. Bile acids were extracted from serum with 0.4 M sodium bicarbonate, adsorbed onto a Sep-Pak C,8 cartridge and eluted with methanol. The extract was derivatized with B M C in acetonitrile using 18-crown-6 crown ether as catalyst and the B M C labelled glycine conjugates and free bile acids were analysed using acetonitrile+ methanol + water gradient elution and detection a t 320/385 nm. Using a novel and simple approach, taurine conjugates were isolated by extracting the dried, derivatized material with water, in contrast to previous methods which required column chromatography cleanup to isolate the taurine conjugates prior to derivatization. The isolated taurine conjugates were then hydrolysed enzymatically, extracted, derivatized and analysed as free-bile acids. Recoveries of individual bile acids varied from 83 - 96% for free and glycine conjugates and 72 - 83% for taurine conjugates. Coefficients of variation were in the range of 5.1 - 12.5%. In addition to the simpler and shorter procedure for taurine conjugates, this method has increased sensitivity over most other procedures and improved H P L C separation for the various bile acids and conjugates with equivalent recovery and reproducibility compared with other published methods.

INTRODUCTION

Recent interest in serum bile acids (SBA) has arisen from their potential use as a test of liver function for workers exposed to chemicals (Edling and Tagesson, 1984, Franco et al., 1986, Liss et al., 1985) and as a sensitive and specific indicator for liver diseases (Linnet e f al., 1982, Skrede et al., 1978, Okuda et al., 1984). For these reasons several attempts to develop a reliable method for identification and determination of individual bile acids and their conjugates in biological fluids have been made. High performance liquid chromatography (HPLC) with UV detection (Nakayama and Nakagaki, 1980, Ruben and van BergeHenegouwen, 1982, Hernanz and Codoce, 1985) or with fluorescence detection (Okuyama et al., 1979, Kamada et al., 1983, Goto et al., 1983, Andreolini et al., 1988) are the most widely used methods for determination of individual bile acids in various body fluids. Methods using fluorescence detection are sufficiently sensitive, but fractionation of free bile acids, glycine conjugates and taurine conjugates creates problems in obtaining consistent reproducibility and good recoveries, and is a very time-consuming procedure. In order to solve these problems we studied optimum conditions for the extraction of bile acids from serum with Sep-Pak C,, cartridges, isolation and enzymatic hydrolysis of taurine conjugates, derivatization with 4-bromomethyl-7-

’Author to whom correspondence should be addressed Dr N. H . Stacey at: NIOHS, GPO Box 58, Sydney, NSW, 2001 Australia. iPermanent address: Institute of Occupational Medicine, Chinese Academy of Preventive Medicine, 29 Nanwei Road, Beijing, People’s Republic of China.

methoxycoumarin (BMC) and gradient elution on a reversed-phase column using acetonitrile + methanol + water as the mobile phase. A sensitive, reproducible and shorter procedure providing much improved separation has been developed. EXPERIMENTAL Reagents. Cholic acid (CA), ursodeoxycholic acid (UDCA), deoxycholic acid (DC), chenodeoxycholic acid (CDC), glycocholic acid (GC), glycodeoxycholic acid (G D C), glycochenodeoxycholic acid (GCDC), glycolithocholic acid (GLC), taurocholic acid (TC), tauroursodeoxycholic acid (TUDC), taurochenodeoxycholic acid (TCDC), and taurodeoxycholic acid (TDC) were purchased from Sigma (St. Louis, MO, USA). 4-Bromomethyl-7-methoxycoumarin (BMC), 1,4,7,10,13,16-hexaoxacyclooctadecane(18-crown-6), and cholylglycine hydrolase (42 units/mg solid) were also purchased from Sigma. Sep-Pak CIX cartridges were from Waters (Milford, MA, USA) while the ARCOIM LC13 disposable filter assembly, 0.45 pm, was obtained from Gelman Sciences (Sydney, Australia). Stock solutions of bile acids were prepared by dissolving each bile acid in methanol to a final concentration of 5 pmol/mL. A standard solution of mixed bile acids was prepared by combining and diluting the stock solutions with acetonitrile to 50 nmoles/mL of individual bile acid. Instruments. A Waters HPLC system with a Baseline 810 chromatography workstation (Waters Chromatography Div, Milford, MA, USA) (dual Model 501 pumps, a U6K lnjector and a Blue Chip personal computer) and a Shimadzu RF-535 fluorescence monitor (Shimadzu, Tokyo, Japan) were used. Sample Preparation. The general scheme for preparation and determination of bile acids in serum is shown in Fig 1. The

0269-3879/90/0136-0140$5.00 136 BIOMEDICAL CHROMATOGRAPHY, VOL. 4, NO 4, 1990

0 John

Wiley & Sons Limited, 1990

ASSAY OF BILE ACIDS BY HPLC

I

Sep-Pak

4

!+I

Hydrolysis

1

Deriv I

HPLC

acetonitrile. A convex gradient (Waters system + 2) at 1 mL/min from 0 to 63% I1 over 47 min was used. The column was then washed with 100% I1 for 8 min at 1.5 mL/min and then allowed to re-equilibrate at 0% I 1 for 6 min before the next injection (see Table l(a)). The excitation and emission wavelengths were 320 and 385 nm, respectively.

' r

p'

Extr action

1

'

Ta"W

Figure 1. General scheme for the separation and determination of bile acids in serum.

detail of the procedure follows. 100 or 200pL serum was diluted with 4 mL 0.4 M N a H C 0 3 solution and passed through a Sep-Pak C L 8cartridge primed previously with 5 mL methanol followed by 5 mL water. After washing the cartridge with 20 mL water, bile acids were eluted with 2 mL methanol. The eluate was dried down at 45 "C under a stream of nitrogen. Derivatization procedure. Sodium carbonate (0.4 mg), 100 p L BMC solution (2 mg/mL in acetonitrile) and 50 p L 18-crown-6 solution (20 mg/mL in acetonitrile) were added to a standard mixture of free and glycine-conjugated bile acids or dried serum eluate. In some assays, varying concentrations of sodium hydroxide in methanol instead of sodium carbonate were added to determine the effect of alkalinity. The mixtures were allowed to react at 40 "C for 1 h. After reaction, the standard and samples were made up to 2.0 mL and 0.5 mL, respectively, with acetonitrile and 10 )*L aliquots were injected into the HPLC. Isolation and hydrolysis of taurine conjugated bile acids. After HPLC analysis of free and glycine conjugated bile acids requiring 1OpL of sample, the remainder was evaporated just to dryness (evaporated samples should not remain in the stream of nitrogen or dissolution will be difficult). Water (170 pL) was added, it was allowed to stand for lOmin, vortexed for 30 s and drawn u p into a 2 mL syringe. Two further washes with 170 p L of water were then carried out and combined with the first 170 pL. The pooled sample was then filtered through an ARCO'M LC13 disposable filter assembly. Then 2 0 0 p L 0.025 M sodium acetate buffer (pH 5.6) and 100 pL cholylglycine hydrolase solution (80 units/mL in 0.005 M Na,HPO,, pH 7.0) were added to the filtrate. The samples were incubated at 37 "C for 20 min. The reaction mixture was then extracted and derivatized as described for serum samples above.

HPLC conditions for taurine conjugated bile acids. The isolated taurine conjugates, converted to free bile acids and derivatized, were run on the same column with the same mobile phases. A convex gradient was again used but at 1.2 mL/min from 0 to 60% I1 over 40 min. The column was washed with 100% 11 for 6 min at 1.5 mL/min and then allowed to re-equilibrate at 0% I1 for 6 rnin before the next injection (see Table l ( b ) ) .

RESULTS AND DISCUSSION Effect of temperature upon derivatization with BMC The effect of temperature of derivatization with BMC upon the quantity of fluorescent product obtained is shown in Fig. 2. Reaction was incomplete at 20°C for 1 h but was almost complete for most bile acids and conjugates after reaction at 40 "C for 1 h. Although reaction at 60 "C for 1 h gave a further slight increase in some bile acids this was accompanied by an increase in background contaminant peaks which caused problems in measuring low levels of bile acids. The optimum temperature for the reaction was thus determined to be 40 "C. Effect of alkali upon derivatization with BMC The derivatization requires alkaline catalysis with a previous report having used potassium hydroxide at 25 yg/mL reaction mix (Kamada et al. 1983). To optimize the concentration of alkali, varying concentrations of sodium hydroxide were added ranging from 10 to 240yg/mL reaction mix. Figure 3 clearly shows that increasing the sodium hydroxide from 10 to 30 bg/rnL increased the quantity of fluorescent product produced from eight different bile acids but higher concentrations 6

HPLC conditions for free and glycine conjugated bile acids. The column used was biova-PakTM ODS ( 5 pm particle size, 15 cm x 3.9 mm ID). Mobile phase 1 was acetonitrile + methanol+water, 15: 13.8:71.2 v/v/v/, and phase I 1 was

A P)

0 X

9

Table 1. The gradient programs" for separation of bile acids

A

B a

B

Composition

Time

Flow

(min)

(mL/rnin)

%I

%II

0.0 47.0 47.1 55.0 0.0 39.9 40.0 46.0

1.o 1 .o 1.5 1.5 1.2 1.2 1.5 1.5

100.0

0.0 63.0 100.0 100.0 0.0

37.0 0.0 0.0 100.0 40.0 0.0 0.0

60.0 100.0 100.0

.-m

type

W

I 3 O

+2 +9 0 +4 +9 0

Program A is for free and glycine conjugated bile acids; program is for taurine conjugated bile acids.

@ John Wiley & Sons Limited, 1990

tJ 37

Curve

2

a, Q

-

0

20

40

Temperature

60

("C)

Figure 2. Effect of reaction temperature on peak height of bile acids at a reaction time of 60 min.

BIOMEDICAL CHROMATOGRAPHY, VOL. 4, NO 4, 1990 137

G. WANG ET AL.

P::

u

U

0

~

20

30

40

120

240

NaOH Concentrations,pg/ml Figure 3. Effect of NaOH concentration and Na,CO, in the medium on formation of BMC-derivatives of bile acids (average of all those assayed).

i I

reduced the yield. The optimum concentration for sodium hydroxide was found t o be 30 pg/mL. Solid sodium carbonate was tested as an alternative to sodium hydroxide and was found to give the highest yield of all with more reliable and consistent derivatization and was therefore selected as the alkaline catalyst in our procedure. The choice of 0.4 mg was initially based on “the few grains” of potassium carbonate used by Andreolini et at., (1988). Isolation and hydrolysis of taurine conjugated bile acids The taurine conjugates d o not react with BMC and need to be isolated from the free and glycine conjugated bile acids, hydrolysed, derivatized and measured separately. In previous reports taurine conjugates were isolated by a variety of relatively time-consuming column cleanup techniques including PHP-LP-20 column (Kamada et al., 1983; Goto et al., 1983) and Sep-Pak SIL cartridge (Street et al., 1985; Andreolini et al., 1988). It was found that the reproducibility and recovery for some of these separations could not easily be replicated in our laboratory. By the simple procedure of performing the isolation of the taurine conjugates after derivatization of free and glycine conjugates, we could extract taurine conjugates with water from the water-insoluble BMC derivatives obviating the need for a column separation. A recovery of 90% or greater was obtained for each taurine conjugate using this method. Lack of contamination by glycine or free bile acids was confirmed by repeat HPLC analysis prior to hydrolysis.

25

.

,

30

.

,

35

.

,

40

‘

I

45

minutes

Figure 4. Chromatogram of a standard mixture of free and glycineconjugated bile acids ranging from 1.4 to 3.3 pmol for amounts injected. For the abbreviations, see text.

chromatograms but the bile acids were well separated from these. Figure 6 shows the analysis of glycine conjugates and free bile acids in serum from a worker exposed to solvents with glycochenodeoxycholate as the major analyte present. Figure 7 shows the separation of the taurine bile acids isolated from serum of the same individual with taurocholate as the major form present. In addition to the reagent peaks there are also apparently other components of serum which react with BMC but these are also generally well separated from the bile acids. Again our method is superior to the other published procedures using BMC. The chromatogram from Okuyama et al. (1979), who used methanol + water for the mobile phase, showed incomplete separation of C D C and DC. Andreolini et al. (1988) achieved good separation using acetonitrile + water as the mobile phase but their run-time required 4 h instead of 50 min as in our method

2.64

7 2.61 L

Chromatography

0

Complete separation of four glycine conjugates and four free bile acids was achieved using acetonitrile methanol + water (15 : 13.8 : 71.2) with a convex gradient from 0% to 63% acetonitrile over 47 min at 1.0 mL/min (Fig. 4). A faster elution of the four taurine conjugates was achieved using the same mobile phase with a convex gradient from 0% to 60% acetonitrile over 40min at 1.2 mL/rnin (Fig. 5 ) . Some trace levels of other compounds, arising from breakdown of the reagent or contaminants in the methanol, appeared in the

+

138 BIOMEDICAL CHROMATOGRAPHY, VOL. 4, NO. 4, 1990

2.58

minutes

Figure 5. Chromatogram of a standard mixture of taurine conjugated bile acids ranging from 2.2 to 3.3 pmol for amounts injected. For the abbreviations, see text.

@ John Wiley & Sons Limited, 1990

ASSAY O F B I L E ACIDS B Y HPLC

3.80-

0

Table 2. Recovery and reproducibility for the determination of free bile acids and their conjugates in spiked serum ( n =6)

I

I

3.40-

25

3'5

30

43

4b

SEA

Mean ( % ) i S D

GC GCDC GDC CA UDC G LC CDC DC TC TUDC TCDC TDC

8516.1 87 f 6.0 87110.5 83 f 9.2 96 f 4.9 88 * 9.0 921 7.8 86k8.1 81 5.8 75 5.7 7216.6 83 f 10.4

CV%

7.2 6.9 12.1 11.1 5.1 10.2 8.5 9.4 7.2 7.6 9.2 12.5

* *

'

minutes

Figure 6. Chromatogram of free and glycine conjugated bile acids in serum from an exposed worker For the abbreviations, see text.

the 4 h run time is required and Goto et al. (1983), whose method requires 3 separate c ~ r o m a t o g r a p ~runs. y

Recovery, reproducibility and sensitivity

Determination of bile acids in serum

The recoveries of 6 normal sera spiked with the mix of bile acids are shown in Table 2. Amounts can be quantitated using either peak height or area under the curve. Although comparison of values using each of the alternatives provided essentially the same numbers, use of area under the curve is preferred. Detection of each of the bile acids was found to be linear over a range of 0.5 to 300 pmol injected onto the HPLC. The mean recoveries of free and glycine-conjugated bile acids ranged from 83 - 96% with coefficient of variations of 5.1 - 12.1%. Analysis of taurine conjugates required a more complex procedure but recoveries were still good with a range of 72 - 83% and coefficient of variations of 7.2 - 12.5%. The detection limits (at a signal to noise ratio of 3) corresponded to 0.05 PM for GC, GCDC, G D C and UDC and 0.08 PM for the remainder. The only other methods to achieve the same degree of sensitivity as ours were Andreolini et al. (1988) where

The method has been used for the determination of bile acids in sera from workers and from rats exposed to chemicals (unpublished) and children with hepatobiliary diseases, in order to provide examples of actual determinations in different settings. Some examples of these results are presented in Table 3, while Figs 6 and 7 show typical chromatograms of SBA in an exposed worker. In conclusion, we have amalgamated aspects from various chromatographic methods used to assay bile acids. The method that has been developed is more sensitive than most other procedures, having a detection limit of 0.05 - 0.08 nmol/mL, provides much improved separation of peaks and is less time consuming. Recovery and reproducibility are at least the equivalent of other published methods. Finally, we have demonstrated that the procedure can be applied to sera from both normal individuals and diseased patients.

Table 3. Concentrations of SBA in different subjects"(Fmol/L)

22

2b

, . , , , . . , . , ,

, , ,

, , , 26

28

30

32

. . , , , , . I .

34

36

A

B

C

D

E

F

0.75 1.78 0.38 0.12

0.73 3.03 0.23 0.20 0.24

0.73 2.12 0.17 0.37 0.46

42.62 90.50 0.36 0.41

46.20 18.72 0.72

67.65 33.45

ND

ND

0.22

ND ND ND

ND ND

0.29

0.19

ND

ND ND

0.13

0.13 0.21 0.24

1.41 0.18 0.15 0.50 0.60

ND

ND

ND

ND

3.35

5.23

6.69

134.91

ND

0.46 ND

0.27

ND

0.12 ND

0.14 0.13 1.07 ND ND

1.50 68.60

N D ~ ND ND ND

0.24 ND

4.35 ND

10.56 ND

116.25

38

minutes

Figure 7. Chromatogram of taurine conjugated bile acids in serum from an exposed worker. For the abbreviations, see text.

@ John Wiley & Sons Limited, 1990

SEA

GC GCDC GDC CA UDC GLC CDC DC TC TUDC TCDC TDC Total

a A, B and C are healthy subjects. D, E and Fare children with biliary atresia. ND, not detectable.

BIOMEDICAL CHROMATOGRAPHY, VOL. 4, NO. 4, 1990 139

G. W A N G E T A L

REFERENCES Andreolini, F., Beade, S. C. and Novotny, M. (1988). J. High Resolut. Chromarogr. 11. 20. Edling C. and Tagesson C. (1984). Br. J. Ind. Med. 41, 257. Franco G., Fonte R., Ternpini G., and Candura F. (1986). lnt. Arch. Occup. fnviron. Health 58, 157. Goto J., Saito M., Chikai T., Goto N. and Narnbara T. (1983). J. Chromatogr. 276, 289. Hernanz A . and Codoceo R. (1985). Clin. Chim. Acta 145, 197. Karnada S., Maeda M. and Tsuji A. (1983). J. Chromatogr. 272, 29. Linnet, K., Kelbaek and Frandsen, P. (1982). Scand. J. Gastroenterol. 17, 263. Liss, G. M. Greenberg, R. A., and Tarnburro, C. H. (1985). Am. J. Med. 78. 68.

140 BIOMEDICAL CHROMATOGRAPH^, VOL. 4, NO. 4, 1990

Nakayarna F. and Nakagaki M. (1980). J. Chromatogr. 183, 287. Okuda, H., Obata, H., Nakanishi, T., Hisarnitsu. T., Matsubara, K. and Watanabe, H. (1984). Hepato-gastroenterol 31, 168. Okuyarna, S., Uernura, D. and Hirata, Y. (1979). Chem Lett (Japan), 461. Ruben, A., Th and Van Berge-Henegouwen G. P. (1982). Clin. Chim. Acta 119, 41. Skrede, S., Solberg, H. E., Blornhoff, J. P. and Gjone, E. (1978). Clin. Chem. 24, 1095. Street, J. M., Trafford, D. J. H. and Makin, H. L. J. (1985). J. Chromatogr. 343, 259. Received 22 September 1989; accepted (revised) 5 February 1990.

@ J o h n Wiley & Sons Limited, 1990