ANNALS OF CLINICAL AND LABORATORY SCIENCE, Vol. 6, No. 1 Copyright © 1976, Institute for Clinical Science

Analysis of the Conjugated Bile Acid Distribution in Human Intestinal Aspirates using Gas Liquid Chromatography * L.J. HAEFFNER, Ph.D., S. J. GORDON, M.D.t, S. STRUM, M.D., M. ELLIOT, B.S., AND O. D. KOWLESSAR, M.D. Department of Medicine, Division of Gastroenterology, Jefferson Medical College of Thomas Jefferson University, Philadelphia, PA 19107

ABSTRACT The conjugated bile acid pattern was evaluated in intestinal aspirates of five normal subjects and five patients known to have an abnormal bile acid dis­ tribution. The glycine/taurine (G/T) ratios for total bile acids were determined by gas liquid chromatography (GLC) and by enzymatic assay using 3-hydroxysteroid dehydrogenase (STDH). By both methods G/T ratios in normal samples ap­ proximated 3.0 as previously reported. A discrepancy between ratio values ob­ tained by the two methods was found in patient samples. It is suggested that the presence of keto bile acids in patient aspirates correlates with this discrepancy. The G/T ratios for individual bile acids were obtained by GLC. The G/T ratio for cholic acid was higher than previously reported. Recovery studies showed that a loss of taurocholic acid in the extraction procedure employed did not account for the high G/T-cholic acid.

Introduction that G/T ratios for individual bile acids de­ Analysis of total conjugated bile acids of termined by paper chromatography approxi­ intestinal aspirates in normal subjects has mate the G/T ratio for total bile acids.13 revealed a G/T ratio approaching 3.0 in The present study was undertaken to com­ published studies.1’2'3,91012131415 This ratio pare the total bile acid G/T ratio of intes­ has been derived utilizing a number of tech­ tinal samples from normal subjects niques including paper chromatography or employing TLC followed by enzyme assay thin layer chromatography (TLC) followed (STDH) with that obtained by TLC by chemical determinations or enzymatic followed by GLC, a technique not pre­ assay. Furthermore, it has been suggested viously utilized for this purpose. The G/T “This work was supported in part by the following ratios for individual bile acids were also grants: Training Grant 2T01 AM5572, National In­ assessed by GLC. In addition, the same stitute of Arthritis, Metabolism and Digestive Disease, methods were applied to intestinal aspirates Bethesda, MD. Contract N00014-68-A-0516, Office of Naval Research, Project Themis and Wallace Labora­ of patients having an abnormal bile acid dis­ tories, Cranbury, NJ. tribution characterized by absent deoxy- t Address reprint requests to Susan J. Gordon, M.D. cholic acid.

15 16 HAEFFNER, ET AL. Materials and Methods solvent system propionic acid: iso-amyl acetate: water: n-propanol (30:40:10:20).7 SAMPLE COLLECTION Standards of cholic acid and glycocholic Intestinal samples were obtained from five acid were simultaneously chromatographed normal subjects and five patients previously and bile acids were detected with iodine found to have an abnormal total bile acid vapor. distribution characterized by the absence of The regions corresponding to the glycine- deoxycholic acid. After an overnight fast, a and taurine-conjugated bile acids were modified Cantor tube was positioned under removed from the plate with a scintered fluoroscopic guidance at the ligament of glass suction apparatus,“ eluted with meth­ Treitz. Thirty minutes after a test meal, anol and dried on a rotary evaporator. The consisting of one slice of buttered toast and a solutions containing glycine- and taurine- cup of tea with sugar, a duodenal sample was conjugated bile acids were brought to 10 and collected in two subjects, while in four sub­ 5 ml, respectively, with methanol. jects the tube was allowed to pass into the distal jejunum where samples were obtained. DETERMINATION OF G/T RATIO All samples were collected over ice, im­ STDH assay. Precisely 0.1 ml and 0.2 ml mediately centrifuged and 2.0 ml aliquots of the glycine- and taurine- conjugated bile were stored at —20° until processed. acid solutions, respectively, were assayed by the method of Iwata and Yamasaki,8 SAMPLE PREPARATION modified to include a blank containing A modification of the STDH assay of sample solution, buffer, NAD and water. Iwata and Yamasaki8 was used to determine The G/T ratio was calculated directly from the bile acid content of an aliquot of intes­ the change in absorbance at 340 nm (aA ) as tinal sample. Using this determination, the follows: total number of aliquots containing 4 to 5 itmoles of bile acid was further processed. AA glycine conjugates x 100 For each 2.0 ml aliquot, the proteins were ACA taurine conjugates x 25 precipitated with 20 ml of ethanol by GLC assay. Exactly 9 ml and 4 ml of the heating at 90° for ten minutes. Exactly 4.6 solutions containing the glycine- and ml N NaOH was added to the supernatant taurine-conjugated bile acids, respectively, which was then extracted with 20 ml petro­ were evaporated to dryness and reconsti­ leum ether. The ether was washed with a tuted with 1.5 ml methanol. The bile acids mixture of 1.2 ml water and 2.2 ml N NaOH were hydrolyzed with 3.5 ml of 2.7 N NaOH in 50 percent ethanol. The initial aqueous- in an autoclave for three hours (122°, 15 psi). ethanol layer and the wash were combined, The free bile acids were recovered by four adjusted to pH 1.1 and extracted se­ chloroform extractions, dried and the quentially with 35, 50, 35, 50 ml of chloro­ methyl-trifluoroacetate derivatives prepared form. The chloroform extracts of all 2.0 ml for GLC (vide infra). The G/T ratios were aliquot preparations from the same sample calculated from the areas on the chromato­ were then combined and evaporated to graph, using the formula: dryness. A methanolic solution of each sample was quantitatively applied to a thin 10/9 (GL + GD + GCDC + GC) layer plate. 5/4 (TL + TD + TCDC + TC) The glycine- and taurine-conjugated bile in which L, D, CDC, and C represent the acids were separated by TLC on Silica Gel G plates 1 mm thick (20 x 20 cm) with the * Lurex Corporation. ANALYSIS OF CONJUGATED BILE ACID DISTRIBUTION 17 areas of measurable lithocholic acid, deoxy- ica Gel G.7 The purified 14C-TC was cholic acid, chenodeoxycholic acid and brought to 10 ml in benzene: ethanol (3:2). cholic acid of the glycine (G) or taurine (T) After protein precipitation of intestinal sam­ conjugated bile acid solution. ples from one normal subject (sample 1) and one patient (sample 2), 0.2 ml of the 14C-TC CONJUGATED BILE ACID DISTRIBUTION (GLC) was added to the supernatant solutions. The GLC analyses of the methyl-trifluo- After mixing, aliquots were removed for roacetate derivatives were performed on STDH and scintillation counting and the glass columns (4 mm id x 4 ft) packed with 2 values obtained were considered to repre­ percent QF-1 on 100/200 Gas Chrom Q.° sent 100 percent. After removal of the neu­ The operating temperatures were: injection tral fats and steroids with petroleum ether, port, 260°; column oven, 245°; hydrogen aliquots were again taken for STDH and flame detector, 260°. Helium was used as the counting. Finally, aliquots of both the carrier gas at a flow rate of 60 ml per min. chloroform and aqueous layers were counted The area of each bile acid on the and STDH assays were performed on metha- chromatograph was determined by triangu­ nolic solutions of the dried chloroform ex­ lation and expressed as area per /¿I solution tracts. Radioactivity was measured by liquid injected. Four injections per sample were scintillation spectrometry using both assayed and the values averaged. The indi­ Aquasol (3.5 ml water and 15 ml Aquasol) vidual bile acids were expressed as a and PPO (5 ml ethanol and 15 ml 2,5- percentage of the total measurable bile acid diphenyloxazole-5.5 g per liter toluene). content. Results RADIOACTIVE SODIUM TAUROCHOLATE (TC) The G/T ratios from intestinal aspirates of RECOVERY the five normal subjects as determined by Exactly 1.5 mg of sodium-taurocholate- the STDH and GLC assays are reported in (Carbonyl 14C)t was purified by TLC on Sil- table I. Statistical analysis comparing results obtained by the two methods reveals a p ° Applied Science Laboratories. value of 0.05. The mean G/T ratios for intes­ tMallinckrodt. tinal samples were 3.1 ± 0.9 by STDH and

T ABLE I

Conjugated Bile Acids in Control Subjects: G/T Ratios

Sex, Age G/T-CDC G/T-C Subject Race Sample G/T STDH G/T GLC G/T-D GLC GLC GLC

LL M, 32, W Duodenal 3.6 3.7 3.4 2.2 10.5 OB (1) M, 45-, W Duodenal 2.4 2.1 1.5 1.2 5.0 (2) Distal jejunal 2.6 1.9 2.7 1.7 1.2 FH M, 23, W Distal jejunal 2.0 2.3 2.1 1.9 2.9 AS M, 21, W Distal jejunal 4.3 3.6 2.1 1.6 8.8 VW F , 26, B Distal jejunal 3.8 2.0 0.9 0.5 6.2

Mean ± S.D. 3.1 ± 0.9 2.6 ± 0.8 2.1 ± 0;9 1.5 ± 0.6 5.8 ± 3.5

STDH - indicates data obtained by the 3-hydroxysteroid dehydrogenase assay. GLC - indicates data from gas liquid chromatographs calculated by triangulation. Aliquots from the same preparations were used for both assays (G - glycine conjugated; T - taurine conjugated; D - deoxycholic acid; CDC - chenodeoxycholic acid; C - cholic acid). 18 HAEFFNER, ET AL.

T A B L E II conjugated bile acids in all normal subjects Percent Distribution of Conjugated accounting for the high G/T-C ratio noted. Bile Acids in Normal Subjects The G/T ratios from samples of the five patients are shown in table III. A striking dis­ Subject GDGCDC GC TD TCDC TC crepancy between G/T by STDH and G/T LL 29.3 13.7 37.6 8.6 6.3 4.5 by GLC was noted in four samples, while the OB (1) 14.9 16.7 37.4 10.0 13.6 7.5 (2) 23.0 29.8 13.6 8.6 17.6 8.2 ratios for I.B. and V.T. were in good FH 12.9 19.5 37.9 6.1 10.2 13.4 agreement. With the exception of these two AS 11.1 14.4 54.0 5.4 9.0 6.1 VW 2.8 10.9 53.9 3.2 20.4 8.8 patients, G/T ratios from all patient samples exceeded the normal ranges obtained by The samples correspond to those in Table I. The both assay methods. The G/T-CDC and G/ glycine conjugates were separated from the taurine conjugates by thin layer chromatography, hydrolyzed T-C from each patient sample are also and assayed by gas liquid chromatography. The TC levels were corrected as described in the Discussion. reported in this table. For these bile acid GD - glycine conjugated deoxycholic acid. ratios, it may be seen that no sample was GCDC - glycine conjugated chenodeoxycholic acid. GC - glycine conjugated cholic acid. within the normal range. TD - taurine conjugated deoxycholic acid. The total bile acid pattern determined by TCDC - taurine conjugated chenodeoxycholic acid. TC - taurine conjugated cholic acid. GLC6 of intestinal samples from each of the patients is presented in table IV. The presence or absence of ursodeoxycholic acid, 2.6 ± 0.8 by GLC. In table I are shown the 3-hydroxy-7-keto-cholanic acid and 3,12- G/T ratios for deoxycholic acid (D), dihydroxy-7-keto-cholanic acid on the chenodeoxycholic acid (CDC) and cholic chromatographs of these samples is noted. acid (C). The mean G/T-C (5.8 ± 3.5) was These unusual bile acids were not mea­ higher than that calculated for G/T-D (2.1 surable in any of the samples from the ± 0.9) and G/T-CDC (1.5 ± 0.6). The dis­ normal subjects studied. Only glycine tribution of the conjugated bile acids in the conjugates of ursodeoxycholic acid and keto same samples is presented in table II. These bile acids were noted in all patients except data show that the predominant conjugated I.B. where taurine conjugates of these bile bile acid was glycocholic acid. The acids were also present. dihydroxy bile acids were predominant A radioactive recovery experiment among the taurine conjugates. Taurocholic employing labelled sodium taurocholate is acid comprised 13.4 percent or less of reported in table V. The recoveries of 14C-

TABLE III

Glycine/Taurine Ratios in Patients With Absent Deoxycholic Acid

Subject Diagnosis G/T (STDH) G/T (GLC) G/T-CDC G/T-C

BC Jejunal-ileal bypass (1) 5 years post surgery 10.6 42.9 28.7 >100 (2) 5 years 7 months post surgery 23.6 48.0 24.2 >600 IB Bacterial overgrowth, unknown etiology 2.7 2.0 0.4 11.2 VT Secondary biliary cirrhosis 0.6 0.4 0.1 0.8 WS Short bowel (ileal resection) 4.7 17.5 5.1 >40 SC Liver disease with colitis 4.9 7.2 3.7 11.3

STDH - indicates data obtained by the 3-hydroxysteroid dehydrogenase assay. GLC - indicates data from gas liquid chromatographs calculated by triangulation. Aliquots from the same preparations were used for both assays (G - glycine conjugated; T - taurine conjugated; CDC - chenodeoxycholic acid; C - cholic acid). ANALYSIS OF CONJUGATED BILE ACID DISTRIBUTION 19

TABLE IV GLC are in general agreement with pre­ Percent Distribution of Total Bile Acids In viously reported values. '•2-3'9'10-12-13>14-15 In Patients With Absent Deoxycholic Acid (GLC) spite of the fact that in these reported studies a number of variables were pres­ Subject L D CDC C U K ent,—sex, age, race, food vs hormonal stimulation, duodenal, jejunal or ileal tube BC (1) <1 0 49 51 + + (2) 1 0 43 56 + + positioning and laboratory assay procedures, IB 3 0 54 43 + + VT 1 0 41 58 0 0 a G/T ratio approaching 3.0 was consistently WS 10 0 44 46 + + found and, therefore, can be accepted as SC <1 0 33 67 + + normal. In the present study, there is a borderline L - lithocholic acid. D - deoxycholic acid significant difference between the G/T ratios CDC - chenodeoxycholic acid. obtained by the STDH and GLC assay C - cholic acid. U - ursodeoxycholic acid. methods (p = 0.05) in normal subjects. In K - 3-hydroxy-7-keto-cholanic acid and/or evaluation of the patient data (table III) this 3,12-dihydroxy-7-keto-choianic acid. These bile acids were not quantified since standard discrepancy is even more apparent, with four keto bile acids were only 60 to 70 percent as reactive as the other reported bile acids on of six samples showing marked differences. gas liquid chromatography. Possible explanations for this discrepancy have been considered. Calculation of the total bile acid G/T TC in the pre-TLC chloroform extracts were ratio by GLC was based only on additive 58 percent and 45 percent. However, STDH values for lithocholic acid (when mea­ assays on separate aliquots taken at the same surable), deoxycholic acid, chenodeoxy­ time showed total bile acid recoveries of 97 cholic acid and cholic acid. This ratio is percent and 95 percent, respectively. presumed to be valid for normal subjects, since these samples did not contain addi­ Discussion tional measurable bile acids. In contrast, the The G/T ratios from intestinal aspirates of gas liquid chromatographs of samples from normal subjects determined by STDH and four of the five patients studied showed unusual bile acids, ursodeoxycholic acid and keto bile acids (table IV). When keto bile TABLE V acids were quantitatively injected onto the Comparison of Percent Taurocholate Recoveries GLC column, they were found to be 60 to 70

Radioactive percent as reactive as a deoxycholic acid STDH Assay TC Assay standard in the flame ionization detector Sample Sample Sample Sample Preparatory Stage 1 2 1 2 (table IV). Furthermore, the longer reten­ tion times of these bile acids produced broad Supernatant 100 100 100 100 peaks on the chromatographs which were After petroleum ether extraction 103 93 100 93 less accurately measured by triangulation. Chloroform layer 97 95 58 45 For these reasons, it was difficult to quantify Aqueous layer ND ND 42 55 accurately the individual keto bile acids,

STDH - indicates data obtained by the 3-hydroxy- thereby preventing their inclusion in the cal­ steroid dehydrogenase assay. culation of the G/T ratio by GLC. TC - sodium taurocholate. Taurocholate-was added to each subject's The STDH assay is specific for the 3 sample. Aliquots were taken simultaneously for hydroxyl of the steroid nucleus. Bile acids STDH assay and for scintillation counting at different stages of the workup for thin layer containing a 3 keto group should not react in chromatography. By STDH assay, there was no loss of bile acids during thin layer chromatog­ the enzyme assay, but if present a potential raphy and elution from the plate. error could occur in G/T ratios calculated by 2 0 HAEFFNER, ET AL. this method. However, no 3 keto bile acids difference between the two assay systems. were observed on the chromatographs of V.T. demonstrated no keto bile acids and samples from the patients studied. I.B. was the only patient who demonstrated Since the STDH assay is based on the measurable keto bile acids in both taurine formation of NADH from NAD, any extra and glycine fractions. Under the latter cir­ source of hydrogen in solution could cumstance, any effect of the keto bile acids produce falsely high values, as could any on the enzyme assay could occur in both the substance absorbing at 340 nm. Interaction glycine and the taurine fractions yielding a with the alcoholic medium, especially at low normal ratio. It appears that there is a cor­ bile acid concentrations, has been dem­ relation between the presence of keto bile onstrated.4 For these reasons, the sample acids and the discrepancy of the G/T ratios blank (consisting of sample solution, buffer, calculated by STDH and GLC. NAD, but with water replacing the enzyme Although an earlier study utilizing paper solution) was prepared and assayed si­ chromatography suggested that G/T ratios multaneously with each glycine- or taurine- for each individual bile acid were similar to conjugated bile acid solution. Usually this the G/T ratios for total conjugated bile acids blank gave a low reading, but on occasion in normal subjects,13 more recent studies the blank of solutions containing taurine- with labelled bile acids have reported conjugated bile acids was quite high. In this conflicting results using TLC. Garbutt et al5 investigation, utilization of this blank was found equivalent G/T ratios for C (2.8 ± employed to correct for any false positive 0.8) and D (3.4 ± 0.4) while Morris et al11 readings, however, this correction has not reported G/T-C ratios in ten normal subjects been applied in other reported studies. ranging from 1.8 to 8.2. No data are cur­ All patient samples with abnormally high rently available using GLC techniques to G/T ratios showed a G/T by GLC in excess calculate G/T ratios. With the methods here of the G/T by STDH. All of these samples reported, normal subjects had G/T-D (2.1 ± had detectable glycine-conjugated keto bile 0.9) in agreement with earlier studies;1,5 G/ acids. It is, therefore, possible that keto bile T-CDC (1.5 ± 0.6) lower than previously acids have an inhibitory effect within the reported;1 and G/T-C (5.8 ± 3.5) higher enzyme assay system. If these bile acids oc­ than previously noted.1,5 cur only in the glycine fraction, a lower G/T High G/T-C ratios reported in this study ratio would be found by STDH than by may be due to high values for glycocholic GLC. Therefore, the reactivity of 3-hydroxy- acid or low values for taurocholic acid. For 7-ket'o-cholanic acid and 3,12-dihydroxy-7- this reason, the total conjugated bile acid keto-cholanic acid in the STDH procedure distribution was analyzed. Little quanti­ was evaluated. These specific keto bile acids tative information is currently available on were selected since they were identified in the normal distribution of the conjugated patient samples. Preliminary studies in our bile acids, but two radioactive pool studies laboratory have shown that additive results have reported such data.114 Calculation of were obtained when glycine-conjugated the data presented by Abaurre et al1 shows keto bile acid was added to glyco- that the mean distribution of conjugated bile chenodeoxycholic acid in a ratio of glyco- acids in intestinal samples from four normal keto/GCDC (0.03 to 0.09). subjects after cholecystokinin stimulation Although an inhibitory effect was not was TC-13 percent, GC-42 percent, TCDC-8 demonstrated in these experiments, an percent, GCDC-27 percent, TD- 2 percent interaction of keto bile acids within the and GD-8 percent. When expressed as enzyme system is still possible. Only two percentages, the mean values recorded by patient samples did not show a marked Stahl and Arnesjo14 give a distribution of ANALYSIS OF CONJUGATED BILE ACID DISTRIBUTION 21 TC-18 percent, GC-24 percent, TCDC-14 material, while GLC afforded the op­ percent, GCDC-19 percent, TD-10 percent, portunity for complete separation of this bile GD-15 percent in nine samples after test acid. meal stimulation. Both of these studies em­ ployed TLC to separate the glycine and References taurine conjugates from each other and sub­ 1. Abaurre, R., Gordon, S. G., Mann, J. G. and Kern, sequently used chemical and/or enzymatic F., Jr.: Fasting bile salt pool size and composition after ileal resection. Gastroenterology 58:679-688, assays. Both studies showed cholic acid to be 1969. the major bile acid for both glycine and 2. Bruusgaard, A. and Thaysen, E. H.: Increased ratio taurine conjugates. In the present study of glycine/taurine conjugated bile acids in the early diagnosis of terminal ileopathy. Acta Med. Scand. where TLC separation was followed by GLC 188:547-548, 1970. (table II), glycocholic acid was found to be 3. Encrantz, J. C. and Sjovall, J.: Bile acids in newborn the predominant conjugated bile acid in all and adult humans. Acta Chem. Scand. ii:1093, 1957. but one sample. However, taurocholic acid 4. Engert, R. and Turner, M. D.: Problems in the values were lower than expected. As a conse­ measurement of bile acids with 3 a hydroxysteroid quence of this observation, the recovery of dehydrogenase. Anal. Biochem. 51:399-407, 1973. 5. Garbutt, J. T., Heaton, K. W., Lack, L., and Tyor, taurocholic acid was investigated. M. P.: Increased ratio of glycine to taurine Labelled sodium taurocholate added to conjugated bile salts in patients with ileal disorders. samples with normal and abnormal bile acid Gastroenterology 56:711-720, 1969. 6. Haeffner, L. J., Gordon, S. J., Strum, S., Elliot, M., distribution (table V) yielded final recov­ and Kowlessar, O. D.: Analysis of the total bile acid eries of 58 percent and 45 percent, respec­ distribution in human intestinal aspirates by gas tively. As a result of this experiment, ob­ liquid chromatography. Ann. Clin. Lab. Sci. 5:507- 510, 1975. served taurocholic acid levels were 7. Hofmann, A. F.: Thin layer absorption chromatog­ considered to represent 50 percent of the raphy of free and conjugated bile acids on silicic content of the starting material and acid. J. Lipid Res. 3:127-129, 1962. 8. Iwata, T. and Yamasaki, K.: Enzymatic determi­ mathematical corrections were made to the nation and thin layer chromatography of bile acids distributions and ratios reported. STDH in blood. J. Biochem. 56:424-431, 1964. assays of these same samples revealed total 9. Mallory, A., Kern, F., Jr., Smith,J., and Savage, D.: Patterns of bile acids and microflora in the human bile acid recoveries of 97 percent and 95 per­ small intestine. 1. Bile acids. Gastroenterology cent, and there was no further loss of bile 64:26-33, 1973. acids during TLC and elution from the 10. Mallory, A., Smith, J., Kern, F., Jr., and Savage, D.: Patterns of bile acids and microflora in the human plate. Therefore, the entire loss of small intestine. J. Lab. Clin. Med. 78:829-830, taurocholic acid represented 5 percent or 1971. less of the total bile acid content. It is ap­ 11. Morris, J. S., Low-Beer, T. S. and Heaton, K. W.: Bile salt metabolism and the colon. Scand. J. parent that selective loss of taurocholic acid Gastroent. 8:425-431, 1973. during chloroform extraction does not ex­ 12. Pomare, E. W. and Heaton, K. W.: The effect of plain either the low taurocholic acid levels cholecystectomy on bile salt metabolism. GUT 14:753-762, 1973. found by GLC, or the observation that the 13. Sjovall, J.: Bile acids in man under normal and predominant taurine conjugates are dihy­ pathological conditions. Bile acids and steroids 73. droxy bile acids. Clin. Chim. Acta 5:33-41, 1960. 14. Stahl, E. and Arnesjo, B.: Taurocholate metabolism It seems reasonable to conclude that pre­ in man. Scand. J. Gastroent. 7:559-566, 1972. vious techniques measured taurocholic acid 15. Turnberg, L. A. and Grahame, G.: Bile salt secre­ in the presence of some additional reactive tion in cirrhosis of the liver. Gut ii.126-133, 1970.