Clinical Science and Molecular Medicine (1 917) 52, 5 1-65.
Evidence for renal control of urinary excretion of bile acids and bile acid sulphates in the cholestatic syndrome J . A. SUMMERFIELD, JULIA CULLEN, S. BARNES BARBARA H. BILLING
AND
Medical Unit, Royal Free Hospital, London
(Received 14 June 1976; accepted 18 August 1976)
variable (9-50%). However, in urine, sulphate esters accounted for a large proportion of the total bile acids (33-72%). 7. The output of bile acid sulphate in the urine was related to the urine total bile acid output but the serum concentration of bile acid sulphate remained relatively constant. Consequently, in contrast to the non-sulphated bile acids, whose renal clearancewas relatively constant, the renal clearance of sulphated bile acids was directly related to the urine total bile acid output. This finding is inconsistent with the earlier hypothesis that their predominance in urine was due to a high renal clearance. It may indicate renal synthesis of some of the bile acid sulphates in the urine and/or inhibition of active renal tubular reabsorption of sulphated bile acids by non-sulphated bile acids.
summary
1. The bile acids and bile acid sulphates in the urine, serum and bile of eight cholestatic patients were studied quantitatively by gasliquid chromatography and gas-liquid chromatographylmass spectrometry. 2. The primary bile acids (cholic acid and chenodeoxycholic acid) comprised on average 94% of the total bile acids in bile, 70% in the serum and 64% in urine. 3. The percentage composition of bile acids in bile was relatively constant and was not influenced by the degree of cholestasis. In contrast, in the serum only the primary bile acids were increased, the concentrations of the secondary bile acids (deoxycholic acid and lithocholic acid) and the minor bile acids remaining constant. 4. The data do not support the hypothesis that monohydroxy bile acids accumulate in cholestasis and are related to the pathogenesis of this syndrome. 5. The pattern of bile acid urinary excretion was similar to that in the serum. But in one patient, 3a,7&12a-trihydroxy-5&cholan-24-0ic acid was a principal urinary bile acid, although very low concentrations of the compound were found in that patient’s serum, suggesting that some of the minor bile acids in urine may originate by epimerization in the kidney. 6.In bile only a small proportion of the bile acids was sulphated (range 2.146%) and in serum the degree of sulphation was very
Key words : bile acids, cholestaticjaundice, renal clearance of bile acids, sulphate esters of bile acids. Introduction In the cholestatic syndrome, the entero-hepatic circulation of bile acids is interrupted and bile acids accumulate in the serum, resulting in an increased urinary excretion. Both the primary bile acids, cholic acid (3a,7a,l2a-trihydroxy-58cholan-24-oic acid) and chenodeoxycholic acid (3~,7a-dihydroxy-5~-cholan-24-oic acid), and the secondary bile acids, deoxycholic acid (3a, 12a-dihydroxy-5~-cholan-24-oic acid) and lithocholic acid (3a-hydroxy-5&cholan-24-oic acid), have been identified in urine and serum (Carey
Correspondence: Dr J. A. Summerfield, Medical Unit, Royal Fsee Hospital, Pond Street, Hampstead, London NW3 2QG.
51
52
J . A. Summerfield et al.
& Williams, 1965; Sandberg, Sjovall, Sjovall & Turner, 1965; Gregg, 1967; Norman & Strandvik, 1971; Makino, Sjovall, Norman & Strandvik, 1971), in addition to other bile acids (Makino et al., 1971; Back, 1973; Summerfield, Billing & Shackleton, 1976a). Although many have found cholic acid and chenodeoxycholic acid to be the predominant bile acids in the urine and serum there have been recent reports of high concentrations of monohydroxy bile acids in the serum (Murphy, Ross & Billing, 1972; Williams, Kaye, Baker, Hurwitz & Senior, 1972) and urine (Makino et al., 1971). We have studied the prevalence of raised concentrations of these bile acids and the role of sulphation. Sulphate esters of bile acids (here subsequently referred to as bile acid sulphates), initially identified in low concentrations in bile by Palmer & Bolt (1971), are known to play an important part in bile acid metabolism in cholestasis. Although low concentrations of bile acid sulphates are found in serum, large amounts are excreted in urine (Makino, Hashimoto, Shinozaki, Yashino & Nakagawa, 1975; Stiehl, 1974). Bile acid sulphates are presumed to be derived from the liver, the discrepancy between the low serum concentration and the large urine output of bile acid sulphates being ascribed to a greater renal clearance of bile acid sulphates than of non-sulphated bile acids (Makino et al., 1975; Stiehl, 1974). This evidence is, however, inconclusive and as the rat kidney is capable of synthesizing bile acid sulphates (Summefield, Gollan & Billing, 1976b), data obtained from cholestatic patients were re-examined in order to define more closely the role of kidney in bile acid metabolism. We now describe the bile acid and bile acid sulphate patterns in simultaneously collected samples of urine, serum and bile from eight patients with various degrees of either intrahepatic or extrahepatic cholestasis. We suggest that the kidney may control the urinary bile acid excretion.
Materials and methods Patients Eight patients (aged 38-61 years) with the cholestatic syndrome were studied (Table l), five with primary biliary cirrhosis and three with
extrahepatic obstructive jaundice. Diagnosis was based on the clinical presentations, liverfunction tests, and appearance of needle liver biopsy. Serum mitochondria1 antibodies were present in all the five patients with primary biliary cirrhosis. The diagnosis in the three patients with extrahepatic obstructive jaundice was established at laparotomy. Renal function was normal, as judged by the blood urea concentration, and none had received antibiotics for at least 10 days before the study or had a urinary tract infection. Starting at 10.00 hours, urine was collected for 24 h in a container with 40 ml of NaOH (1 mol/l) to inhibit bacterial growth. At the end of this collection, serum and bile were sampled from the fasting patients. In the patients with extrahepatic obstructive jaundice, bile was obtained at percutaneous transhepatic cholangiography before laparotomy, and in those with primary biliary cirrhosis a duodenal tube was positioned under fluoroscopic control and bilerich duodenal juice collected after stimulation with 100 units of Pancreozymin-Boots (The Boots Co. Ltd, Nottingham, U.K.). No bile sample was obtained from the duodenal intubation of patient no. 4. All samples were stored at -20°C. Informed consent for this study was obtained from all the patients. Reagents and methods In addition to the reagents described previously (Summerfield et al., 1976a, b), sodium [24-14C]lithocholate (54 pCi/pmol) was purchased from Amersham-Buchler G.m.b.H. and Co. KG (D-3300 Braunschweig, West Germany) and sodium [24-14C]taurocholate (58 pCi/pmol) from The Radiochemical Centre (Amersham, Bucks., U.K.). The labelled bile salts were purified by t.1.c. on 0.25 mm Silica gel/CT plates (Reeve Angel Scientific Ltd, London, SE1 6BD). Sodium [24-14C]lithocholate was run in the solvent system waterlpropan-1-ol/propanoic acid/2-methylbutyl acetate (1 :2: 3 :4, by vol.) (Hofmann, 1962) and sodium [24' 4C]taurocholate was run in butan-1-ol/acetic acidlwater (10:1:1, by vol.) (Ganshirt, Koss & Morianz, 1960), with appropriate standards. The compounds were detected by spraying with a methanolic solution of iodine (3.5 g/100 ml), which was then allowed to sublime. The silica gel containing the labelled bile acids was
Normal values
54F
56F
58F
46M
57M
61F
38F
57F
Patient Age and no. sex
a 19
u 19
4
5
illness (Yam)
of
Duration
Primary biliary cirrhosis Primary biliary cirrhosis Primary biliary cirrhosis Rimary biliary cirrhosis Primary biliary cirrhosis Adenocarcinoma of bile duct Benign biladuct stricture Metastases, carcinoma ovary
Diagnosis
+++
No
++
(+I
No
(+I
No
+++
Pruritus
2-14
496
522
428
241
87
308
63
24
457
445
393
216
75
287
51
17
73 62-80
4-15
3-13
66
57
75
65
51
81
77
Total
36-50
33
35
26
31
37
26
33
50
Albumin
Plasma proteins (g/I)
35
22
57
75
40
250
42
37
Aspartate transaminase (i.u./l)
48
122
64
56
94
88
48
128
Alkaline phosphatase (KA units/ Total Conj. 100ml)
(Irmol/l)
Bilirubin
3.8-6.7
7.8
6.7
19.2
5.3
7.6
4.1
5.3
14
Cholesterol (mmol/l)
2,246
2.2
3.3
2.8
4.2
5.3
6.6
3.3
7.3
urea (mmol/l)
Blood
Negative
Negative
Negative
1:40
> 1:40
> 1:40
> 1:60
1:40
Mitochondria1 antibody titre
TABLE 1. Clinical data for the eight patients The chemical pathological data were obtained within 3 days of the collection of samples for bile acid analysis.
Extrahepatic cholestasis Extrahepatic cholestasis Extrahepatic cholestasis
Cirrhotic
Slight fibrosis
Cirrhotic
Cirrhotic
Extensive fibrosis
Liver biopsy diagnosis
v,
w
*
P
9
$: 3'
54
J. A . Summerfield et al.
scraped off the plates and eluted with methanol. Samples of urine (100 ml) were adjusted to pH 10 with NaOH (1.0 mol/l) and placed in an ultrasonic bath for 5 min to eliminate protein binding of bile acids. The resulting flocculent precipitate was removed by centrifugation. Samples of serum (2-10 ml) and bile (1-5 ml) were adjusted to pH 10 with NaOH (1.0 mol/l), made up to 100 ml with water, and placed in an ultrasonic bath for 5 min. N o flocculent precipitate appeared with serum or bile. To each sample was added a tracer dose (approx. 2.5 x lo5 d.p.m.) of either sodium [24-14C]lithocholate or sodium [24-14C]taurocholate to assess the recovery of bile acids during the subsequent procedures. The analytical methods have been described before (Summerfield et al., 1976a, b). The bile acids were extracted with Amberlite XAD-7 [Rohm and Haas (U.K.) Ltd, Croydon CR9 3NB, U.K.], with chromatography on Sephadex LH20 (Pharmacia, Uppsala, Sweden) to separate the sulphate esters (Sjovall & Vihko, 1966). The sulphate fraction was hydrolysed by solvolysis (Burstein & Lieberman, 1958). Both fractions were deconjugated with clostridial cholylglycine hydrolase (Nair, Gordon, Gordon, Reback & Mendeloff, 1965) and the free bile acids extracted and methylated. Sa-Cholestane (Applied Science Laboratories, State College, Pennsylvania, U S A . ) was added to the samples, as an internal standard, and they were methylated in either 2 ml of methanol/ acetyl chloride (19:1, vlv) overnight or freshly prepared ethereal diazomethane foi 15 min and evaporated to dryness under a stream of N,. The 0-trimethylsilyl ethers were prepared by the method of Makita & Wells (1963). Gas-liquid chromatography (g.1.c.) was performed on 2.7 m x 2 mm (internal diam.) glass columns packed with 1% Hi-Eff 8BP on Gas-Chrom Q (100-120 mesh) in a Pye 104 series, model 64, gas chromatograph with a heated flame ionization detector. The operating conditions were: nitrogen carrier flow rate 20 rd/min, hydrogen flow rate 20 ml/min, column temperature 235"C, detector oven and injector temperature 250°C. The retention times of the bile acids were calculated relative to methyl cholate trimethylsilyl ether. The bile acids were quantified by comparing the peak area, corrected for the cholestane peak area, with the peak areas from equimolar bile acids standards
analysed on the same day. The total bile acid concentration was expressed as the sum of the g.1.c. results. For bile acids with no available standards, peak areas were compared with a standard with a similar number of hydroxyl groups (cholic acid for trihydroxy bile acids; chenodeoxycholic acid for dihydroxy bile acids ; lithocholic acid for monohydroxy bile acids). Since the relative peak area response was similar for all the standard bile acids this error was small. A further source of error arose from the nonhomogeneous nature of some peaks containing C2, steroids as well as bile acids (Summerfield ef al., 1976a). Confirmation of the identity of the bile acids was obtained by g.l.c./mass spectrometry on a Varian Mat 731 instrument in some of the samples by mass fragmentography. In the urine samples, chromatography on Sephadex LH20 separated the trihydroxy bile acids (nonsulphate fraction) from dihydroxy and monohydroxy bile acids (largely in the sulphate fraction). For example, chenodeoxycholic acid and 3a,7~,12a-lrihydroxy-5~-choIan-24-oic acid were difficult to separate on g.l.c., but when obtained from urine they were almost completely separated by their differing sulphation patterns (Table 5 ) : in the serum and bile neither of these bile acids was sulphated to any degree, making separation and measurement of the smaller component (3a,78,12a-trihydroxy58-cholan-24-oic acid) impossible (Tables 3 and 4). The presence of low concentrations of this trihydroxy bile acid in the chenodeoxycholic acid peak in samples of serum and bile was confirmed by mass fragmentography. The major components of the bile acid mixture, cholic acid and chenodeoxycholic acid, were consequently slightly contaminated. The identification of the bile acids in these samples has been described in detail (Summerfield et al., 1976a). Recovery experiments
A mixture of free and conjugated cholic acid, chenodeoxycholic acid, deoxycholic acid and lithocholic acid and sulphate esters of lithocholic acid, glycolithocholic acid and taurolithocholic acid (4-12 lrmol) were added to normal urine, serum and bile and analysed by the above method. The total bile acid concentration was determined by the method of Murphy, Billing & Baron (1970), in the starting material and
Urinary bile acids in cholestasis
after extraction of the free bile acids from the enzymic deconjugation step. The recoveries in single samples of urine, serum and bile were: for free and conjugated bile acids, 85%, 78% and 88 % respectively, and for monohydroxy bile acid sulphates, 93 %, 74% and 69% respectively. In addition, a tracer dose (approx. 2.5 x los d.p.m.) of either sodium [24-14C]lithocholateor sodium [24-14C]taurocholatewas added to each of the samples in this study. Sodium [24-14C]lithocholate was selected to assess recovery of non-polar bile acids and sodium [24-14C]taurocholate to assess recovery of the polar compounds,and also to indicatewhether the enzymic deconjugationwas complete. TWO1 ml aliquots of thestarting material weresolubilizedwith1 ml of NCS tissue solubilizer (Amersham/Searle, Arlington Heights, Illinois, U.S.A.) at 45°C for 20 min and decolourizedwith 250 pl of hydrogen peroxide (100 vol.) at 45°C for 20 min. Acetic acid (50p l ) was then added. A portion (10 ml) of a scintillant, containing 2,5-diphenyloxazole(4 g), 1:4-di-2-(5-phenyloxazolyl)-benzene(0.05 g) and Triton X-100 [Rohmand Haas (U.K.) Ltd] (500 ml) in toluene (1 litre) was added. The free bile acids of both the non-sdphate and sulphate fractions, extracted from the enzymic deconjugation step, were made up to 1 ml in methanol and two 10 p1 aliquots were added to 10 ml of scintillant. Counting of 14C radioactivity was performed in a Packard Tricarb liquid-scintilla-
55
tion spectrometer. Corrections for quenching were made by the channels ratio method. The radioactivity from the labelled bile acids was found in the non-sulphate fractions only. The mean recoveries of the labelled bile acids were: for sodium [24-14C]lithocholateadded to the samples of five patients, urine 91 f4.3% ( ~ s E M ) , serum 84+1.2%, bile 91+2.1%; for sodium [24-' 4C]taurocholate added to the samples from three patients, urine 97%, serum 94%, bile 79%.
Results The bile acids identified and measured by g.1.c. are listed in Table 2 with their retention times relative to methylcholate trimethylsilyl ether (RRT,). The major components of the bile acids in urine, serum and bile, cholic acid, deoxycholic acid, chenodeoxycholic acid and lithocholic acid were present in sufficient concentration to be identified by g.1.c.lma.s spectrometry in most samples. The concentrations of the other bile acids varied widely and those samplescontaining an appreciableamount of these minor components were submitted to g.l.c./mass spectrometry for confirmation of their identity. In other samples, particularly bile, where the concentration of the minor components was usually very small, identification depended on the retention time of the
TABLE 2. Bile acids and their relative retention times A-M are letters used to identify the bile acids in Tables 3, 4 and 5 . Retention times are relative to methylcholate trimethylsilyl ether ( = 1). Relative retention time A
B C D E
F
Systematic name
1 1$20
3a,7a,12a-Trihydroxy-S&cholan-24-oate
1.25 1.40 1.51
230
3~,7a-Dihydroxy-5/kholan-24-oate 3a,6a.7a-Trihydroxy-S~-cholan-24-oate 3a,l2a-Dihydroxy-S~cholan-24-oate 3a,7/?,12a-Trihydroxy-S&cholan-24-oate 3a,7a-Dihydroxy-S/$~holan-24-oate 3a-Hydroxy-Sacholan-24-oate or 3/I-hydroxy-S~-cholan-24-oate 3a,6a-Dihydroxy-Sj&cholan-24-0ate 3a-Hydroxy-Sjkholan-24-oate
2-37 2.56 2.83
3a,7~-Dihydroxy-5&cholan-24-oate 3&Hydroxychol-5-en-24-oate
G
1.53 1.64
H
1.77
I J K L M
2.00
Trivial name Cholate Trihydroxy Hyocholate Deoxycholate Chenodeoxycholate Allo-lithocholate or iso-lithocholate Hyodeoxycholate Li thocholate Dihydroxy Ursodeoxycholate
TABLE 3. Biliary bile acids and bile acid sulphates in cholestasis
Mean+ SD
8
4 5
7
3
2
6
1
Patient no.
339 (3.2) 540 (0.7)
355 (2.5) 1397 (0.6) 5
nd
10 (9.1)
(nd)
(nd)
2
nd
(nd)
(nd)
5
12
21
(W
(nd)
27 (44)
42
nd
nd
nd
nd
nd
(nd)
(nd)
15 (nd) nd
(nd)
19
(nd)
6
(nd) nd
23
nd
11
(2.8) 850 (0.7)
nd
nd
nd
3185 (1.3) 2631
134 (18) 33
C
F
B
A
D
Trihydroxy
(nd) nd
19
(nd) nd
4
267 (3.7) 97 (4.1)
1979 (2.2) 1838 (1.3)
(2.5) 670 (5.1)
(100) 26 (17)
863 (3.2) 1637
G
63 (16) 5
E
2 (46)
3 (45)
nd
27
(nd)
2 (100)
10 (100)
Not available 10 5 (40) (nd)
6
1 (100)
nd
nd
L
(nd)
(nd)
(29) 13
58
(100)
13
K
2 (100) 29 (6.6)
nd
6 (100)
(100)
7 (100) 6
I
Dihydroxy
1 (100)
nd
(100)
21
nd
(100)
3
nd
nd
H
12 (27) 4 (33)
22 (25) 37 (17)
(35) 47 (54)
84
101 (34)
J
M
1 (100)
nd
nd
11 (54)
1 (100)
nd
50 (18)
Monohydroxy
660
674
341 1
2390
1686
4488
4416
Total
3.3k0.9
2.4
4.0
2.1
2.8
4.6
3.7
3.8
sulphated
% of total
Results of the analyses of individual bile acids (pnol/l) are shown for seven patients. The figures in parentheses indicate the percentage of the individual bile acids present as sulphate ester. A-M refer to the bile acids listed in Table 2. nd = not detected.
%
0
&
s
B3
+
4
nd
8
Meanf SD
nd
5
nd nd
nd
4
7
nd
1.5 (nd) 4.1 (33) 1.o (56) nd
nd
L
nd
K
3
I
nd
G
nd
nd
nd
nd
E
2
C
F
D
nd
B
Dihydroxy
nd
A
Trihydroxy
6
1
Patient no.
H J
Monohydroxy
TABLE 4. Serum bile acids and bile acid sulphates in cholestasis Results of the analyses of individual bile acids (,umol/l) are shown for eight patients. See also Table 3.
nd
M 17.0
Total
46
sulphated
% of total
5
58
J. A . Summerfield et al.
compound on g.1.c. This group of compounds will be referred to collectively as the minor bile acids; it includes: hyocholic acid, 3a,7/?,12atrihydroxy-5/?-cholan-24-oic acid, an unidentified trihydroxy bile acid (RRT, 1.20), 3/?,7adihydroxy-5/?-cholan-24-oic acid, hyodeoxycholic acid, ursodeoxycholic acid, one unidentified dihydroxy bile acid (RRT. 2.37), a monohydroxy bile acid (allo- or iso-lithocholic acid) and 3~-hydroxychol-5-en-24-oic acid. Total bile acids
the much larger chenodeoxycholic peak, making measurement impossible (Table 4). The sum of the concentrations of cholic acid and chenodeoxycholic acid was highly correlated with the serum total bile acid concentration (r = 0.99, P < 0.001). However, in contrast to the bile, the sum of the lithocholic acid and deoxycholic acid concentrations ( r = -0.04) and the sum of the minor bile acid concentrations (r = 0.41) was not significantly related to the serum total bile acid concentration and remained very low. Consequently, whereas the molar percentage of cholic acid and chenodeoxycholic acid in the serum increased linearly with the serum total bile acid concentration ( r = 0.94, P < 0.001), the molar percentages of lithocholc acid and deoxycholic acid (r = -0.88, P < 0.01) and the minor bile acids (r = -0.79, Pc 0.05) decreased
Bile. The biliary bile acid analyses of seven patients (Table 3) show that the biliary total bile acid concentration varied between 661 and 4480 ,umol/l. The predominant bile acids were cholic acid (53 f22%; mean percentage composition fSD) and chenodeoxycholic acid (41 ? 23%); in only two instances was chenodeoxycholic acid the major bile acid. Small amounts of lithocholic acid (1.6 +0.8%) and deoxycholic acid (0.90k 1.1 %) were present. The minor bile / acids comprised only 3.7f1.4% of the total. a / 8o / Although a small amount of 3a,7/?,12a-trihya’ droxy-5/?-cholan-24-oicacid was identified in the two samplesexamined by mass fragmentography, it was hidden in the much larger chenodeoxycholic acid peak on g.l.c., making measurement impossible (Table 3). The concentration of total primary bile acids was highly correlated with the biliary total bile acid concentration (r = 0.999, P
Evidence for renal control of urinary excretion of bile acids and bile acid sulphates in the cholestatic syndrome J . A. SUMMERFIELD, JULIA CULLEN, S. BARNES BARBARA H. BILLING
AND
Medical Unit, Royal Free Hospital, London
(Received 14 June 1976; accepted 18 August 1976)
variable (9-50%). However, in urine, sulphate esters accounted for a large proportion of the total bile acids (33-72%). 7. The output of bile acid sulphate in the urine was related to the urine total bile acid output but the serum concentration of bile acid sulphate remained relatively constant. Consequently, in contrast to the non-sulphated bile acids, whose renal clearancewas relatively constant, the renal clearance of sulphated bile acids was directly related to the urine total bile acid output. This finding is inconsistent with the earlier hypothesis that their predominance in urine was due to a high renal clearance. It may indicate renal synthesis of some of the bile acid sulphates in the urine and/or inhibition of active renal tubular reabsorption of sulphated bile acids by non-sulphated bile acids.
summary
1. The bile acids and bile acid sulphates in the urine, serum and bile of eight cholestatic patients were studied quantitatively by gasliquid chromatography and gas-liquid chromatographylmass spectrometry. 2. The primary bile acids (cholic acid and chenodeoxycholic acid) comprised on average 94% of the total bile acids in bile, 70% in the serum and 64% in urine. 3. The percentage composition of bile acids in bile was relatively constant and was not influenced by the degree of cholestasis. In contrast, in the serum only the primary bile acids were increased, the concentrations of the secondary bile acids (deoxycholic acid and lithocholic acid) and the minor bile acids remaining constant. 4. The data do not support the hypothesis that monohydroxy bile acids accumulate in cholestasis and are related to the pathogenesis of this syndrome. 5. The pattern of bile acid urinary excretion was similar to that in the serum. But in one patient, 3a,7&12a-trihydroxy-5&cholan-24-0ic acid was a principal urinary bile acid, although very low concentrations of the compound were found in that patient’s serum, suggesting that some of the minor bile acids in urine may originate by epimerization in the kidney. 6.In bile only a small proportion of the bile acids was sulphated (range 2.146%) and in serum the degree of sulphation was very
Key words : bile acids, cholestaticjaundice, renal clearance of bile acids, sulphate esters of bile acids. Introduction In the cholestatic syndrome, the entero-hepatic circulation of bile acids is interrupted and bile acids accumulate in the serum, resulting in an increased urinary excretion. Both the primary bile acids, cholic acid (3a,7a,l2a-trihydroxy-58cholan-24-oic acid) and chenodeoxycholic acid (3~,7a-dihydroxy-5~-cholan-24-oic acid), and the secondary bile acids, deoxycholic acid (3a, 12a-dihydroxy-5~-cholan-24-oic acid) and lithocholic acid (3a-hydroxy-5&cholan-24-oic acid), have been identified in urine and serum (Carey
Correspondence: Dr J. A. Summerfield, Medical Unit, Royal Fsee Hospital, Pond Street, Hampstead, London NW3 2QG.
51
52
J . A. Summerfield et al.
& Williams, 1965; Sandberg, Sjovall, Sjovall & Turner, 1965; Gregg, 1967; Norman & Strandvik, 1971; Makino, Sjovall, Norman & Strandvik, 1971), in addition to other bile acids (Makino et al., 1971; Back, 1973; Summerfield, Billing & Shackleton, 1976a). Although many have found cholic acid and chenodeoxycholic acid to be the predominant bile acids in the urine and serum there have been recent reports of high concentrations of monohydroxy bile acids in the serum (Murphy, Ross & Billing, 1972; Williams, Kaye, Baker, Hurwitz & Senior, 1972) and urine (Makino et al., 1971). We have studied the prevalence of raised concentrations of these bile acids and the role of sulphation. Sulphate esters of bile acids (here subsequently referred to as bile acid sulphates), initially identified in low concentrations in bile by Palmer & Bolt (1971), are known to play an important part in bile acid metabolism in cholestasis. Although low concentrations of bile acid sulphates are found in serum, large amounts are excreted in urine (Makino, Hashimoto, Shinozaki, Yashino & Nakagawa, 1975; Stiehl, 1974). Bile acid sulphates are presumed to be derived from the liver, the discrepancy between the low serum concentration and the large urine output of bile acid sulphates being ascribed to a greater renal clearance of bile acid sulphates than of non-sulphated bile acids (Makino et al., 1975; Stiehl, 1974). This evidence is, however, inconclusive and as the rat kidney is capable of synthesizing bile acid sulphates (Summefield, Gollan & Billing, 1976b), data obtained from cholestatic patients were re-examined in order to define more closely the role of kidney in bile acid metabolism. We now describe the bile acid and bile acid sulphate patterns in simultaneously collected samples of urine, serum and bile from eight patients with various degrees of either intrahepatic or extrahepatic cholestasis. We suggest that the kidney may control the urinary bile acid excretion.
Materials and methods Patients Eight patients (aged 38-61 years) with the cholestatic syndrome were studied (Table l), five with primary biliary cirrhosis and three with
extrahepatic obstructive jaundice. Diagnosis was based on the clinical presentations, liverfunction tests, and appearance of needle liver biopsy. Serum mitochondria1 antibodies were present in all the five patients with primary biliary cirrhosis. The diagnosis in the three patients with extrahepatic obstructive jaundice was established at laparotomy. Renal function was normal, as judged by the blood urea concentration, and none had received antibiotics for at least 10 days before the study or had a urinary tract infection. Starting at 10.00 hours, urine was collected for 24 h in a container with 40 ml of NaOH (1 mol/l) to inhibit bacterial growth. At the end of this collection, serum and bile were sampled from the fasting patients. In the patients with extrahepatic obstructive jaundice, bile was obtained at percutaneous transhepatic cholangiography before laparotomy, and in those with primary biliary cirrhosis a duodenal tube was positioned under fluoroscopic control and bilerich duodenal juice collected after stimulation with 100 units of Pancreozymin-Boots (The Boots Co. Ltd, Nottingham, U.K.). No bile sample was obtained from the duodenal intubation of patient no. 4. All samples were stored at -20°C. Informed consent for this study was obtained from all the patients. Reagents and methods In addition to the reagents described previously (Summerfield et al., 1976a, b), sodium [24-14C]lithocholate (54 pCi/pmol) was purchased from Amersham-Buchler G.m.b.H. and Co. KG (D-3300 Braunschweig, West Germany) and sodium [24-14C]taurocholate (58 pCi/pmol) from The Radiochemical Centre (Amersham, Bucks., U.K.). The labelled bile salts were purified by t.1.c. on 0.25 mm Silica gel/CT plates (Reeve Angel Scientific Ltd, London, SE1 6BD). Sodium [24-14C]lithocholate was run in the solvent system waterlpropan-1-ol/propanoic acid/2-methylbutyl acetate (1 :2: 3 :4, by vol.) (Hofmann, 1962) and sodium [24' 4C]taurocholate was run in butan-1-ol/acetic acidlwater (10:1:1, by vol.) (Ganshirt, Koss & Morianz, 1960), with appropriate standards. The compounds were detected by spraying with a methanolic solution of iodine (3.5 g/100 ml), which was then allowed to sublime. The silica gel containing the labelled bile acids was
Normal values
54F
56F
58F
46M
57M
61F
38F
57F
Patient Age and no. sex
a 19
u 19
4
5
illness (Yam)
of
Duration
Primary biliary cirrhosis Primary biliary cirrhosis Primary biliary cirrhosis Rimary biliary cirrhosis Primary biliary cirrhosis Adenocarcinoma of bile duct Benign biladuct stricture Metastases, carcinoma ovary
Diagnosis
+++
No
++
(+I
No
(+I
No
+++
Pruritus
2-14
496
522
428
241
87
308
63
24
457
445
393
216
75
287
51
17
73 62-80
4-15
3-13
66
57
75
65
51
81
77
Total
36-50
33
35
26
31
37
26
33
50
Albumin
Plasma proteins (g/I)
35
22
57
75
40
250
42
37
Aspartate transaminase (i.u./l)
48
122
64
56
94
88
48
128
Alkaline phosphatase (KA units/ Total Conj. 100ml)
(Irmol/l)
Bilirubin
3.8-6.7
7.8
6.7
19.2
5.3
7.6
4.1
5.3
14
Cholesterol (mmol/l)
2,246
2.2
3.3
2.8
4.2
5.3
6.6
3.3
7.3
urea (mmol/l)
Blood
Negative
Negative
Negative
1:40
> 1:40
> 1:40
> 1:60
1:40
Mitochondria1 antibody titre
TABLE 1. Clinical data for the eight patients The chemical pathological data were obtained within 3 days of the collection of samples for bile acid analysis.
Extrahepatic cholestasis Extrahepatic cholestasis Extrahepatic cholestasis
Cirrhotic
Slight fibrosis
Cirrhotic
Cirrhotic
Extensive fibrosis
Liver biopsy diagnosis
v,
w
*
P
9
$: 3'
54
J. A . Summerfield et al.
scraped off the plates and eluted with methanol. Samples of urine (100 ml) were adjusted to pH 10 with NaOH (1.0 mol/l) and placed in an ultrasonic bath for 5 min to eliminate protein binding of bile acids. The resulting flocculent precipitate was removed by centrifugation. Samples of serum (2-10 ml) and bile (1-5 ml) were adjusted to pH 10 with NaOH (1.0 mol/l), made up to 100 ml with water, and placed in an ultrasonic bath for 5 min. N o flocculent precipitate appeared with serum or bile. To each sample was added a tracer dose (approx. 2.5 x lo5 d.p.m.) of either sodium [24-14C]lithocholate or sodium [24-14C]taurocholate to assess the recovery of bile acids during the subsequent procedures. The analytical methods have been described before (Summerfield et al., 1976a, b). The bile acids were extracted with Amberlite XAD-7 [Rohm and Haas (U.K.) Ltd, Croydon CR9 3NB, U.K.], with chromatography on Sephadex LH20 (Pharmacia, Uppsala, Sweden) to separate the sulphate esters (Sjovall & Vihko, 1966). The sulphate fraction was hydrolysed by solvolysis (Burstein & Lieberman, 1958). Both fractions were deconjugated with clostridial cholylglycine hydrolase (Nair, Gordon, Gordon, Reback & Mendeloff, 1965) and the free bile acids extracted and methylated. Sa-Cholestane (Applied Science Laboratories, State College, Pennsylvania, U S A . ) was added to the samples, as an internal standard, and they were methylated in either 2 ml of methanol/ acetyl chloride (19:1, vlv) overnight or freshly prepared ethereal diazomethane foi 15 min and evaporated to dryness under a stream of N,. The 0-trimethylsilyl ethers were prepared by the method of Makita & Wells (1963). Gas-liquid chromatography (g.1.c.) was performed on 2.7 m x 2 mm (internal diam.) glass columns packed with 1% Hi-Eff 8BP on Gas-Chrom Q (100-120 mesh) in a Pye 104 series, model 64, gas chromatograph with a heated flame ionization detector. The operating conditions were: nitrogen carrier flow rate 20 rd/min, hydrogen flow rate 20 ml/min, column temperature 235"C, detector oven and injector temperature 250°C. The retention times of the bile acids were calculated relative to methyl cholate trimethylsilyl ether. The bile acids were quantified by comparing the peak area, corrected for the cholestane peak area, with the peak areas from equimolar bile acids standards
analysed on the same day. The total bile acid concentration was expressed as the sum of the g.1.c. results. For bile acids with no available standards, peak areas were compared with a standard with a similar number of hydroxyl groups (cholic acid for trihydroxy bile acids; chenodeoxycholic acid for dihydroxy bile acids ; lithocholic acid for monohydroxy bile acids). Since the relative peak area response was similar for all the standard bile acids this error was small. A further source of error arose from the nonhomogeneous nature of some peaks containing C2, steroids as well as bile acids (Summerfield ef al., 1976a). Confirmation of the identity of the bile acids was obtained by g.l.c./mass spectrometry on a Varian Mat 731 instrument in some of the samples by mass fragmentography. In the urine samples, chromatography on Sephadex LH20 separated the trihydroxy bile acids (nonsulphate fraction) from dihydroxy and monohydroxy bile acids (largely in the sulphate fraction). For example, chenodeoxycholic acid and 3a,7~,12a-lrihydroxy-5~-choIan-24-oic acid were difficult to separate on g.l.c., but when obtained from urine they were almost completely separated by their differing sulphation patterns (Table 5 ) : in the serum and bile neither of these bile acids was sulphated to any degree, making separation and measurement of the smaller component (3a,78,12a-trihydroxy58-cholan-24-oic acid) impossible (Tables 3 and 4). The presence of low concentrations of this trihydroxy bile acid in the chenodeoxycholic acid peak in samples of serum and bile was confirmed by mass fragmentography. The major components of the bile acid mixture, cholic acid and chenodeoxycholic acid, were consequently slightly contaminated. The identification of the bile acids in these samples has been described in detail (Summerfield et al., 1976a). Recovery experiments
A mixture of free and conjugated cholic acid, chenodeoxycholic acid, deoxycholic acid and lithocholic acid and sulphate esters of lithocholic acid, glycolithocholic acid and taurolithocholic acid (4-12 lrmol) were added to normal urine, serum and bile and analysed by the above method. The total bile acid concentration was determined by the method of Murphy, Billing & Baron (1970), in the starting material and
Urinary bile acids in cholestasis
after extraction of the free bile acids from the enzymic deconjugation step. The recoveries in single samples of urine, serum and bile were: for free and conjugated bile acids, 85%, 78% and 88 % respectively, and for monohydroxy bile acid sulphates, 93 %, 74% and 69% respectively. In addition, a tracer dose (approx. 2.5 x los d.p.m.) of either sodium [24-14C]lithocholateor sodium [24-14C]taurocholatewas added to each of the samples in this study. Sodium [24-14C]lithocholate was selected to assess recovery of non-polar bile acids and sodium [24-14C]taurocholate to assess recovery of the polar compounds,and also to indicatewhether the enzymic deconjugationwas complete. TWO1 ml aliquots of thestarting material weresolubilizedwith1 ml of NCS tissue solubilizer (Amersham/Searle, Arlington Heights, Illinois, U.S.A.) at 45°C for 20 min and decolourizedwith 250 pl of hydrogen peroxide (100 vol.) at 45°C for 20 min. Acetic acid (50p l ) was then added. A portion (10 ml) of a scintillant, containing 2,5-diphenyloxazole(4 g), 1:4-di-2-(5-phenyloxazolyl)-benzene(0.05 g) and Triton X-100 [Rohmand Haas (U.K.) Ltd] (500 ml) in toluene (1 litre) was added. The free bile acids of both the non-sdphate and sulphate fractions, extracted from the enzymic deconjugation step, were made up to 1 ml in methanol and two 10 p1 aliquots were added to 10 ml of scintillant. Counting of 14C radioactivity was performed in a Packard Tricarb liquid-scintilla-
55
tion spectrometer. Corrections for quenching were made by the channels ratio method. The radioactivity from the labelled bile acids was found in the non-sulphate fractions only. The mean recoveries of the labelled bile acids were: for sodium [24-14C]lithocholateadded to the samples of five patients, urine 91 f4.3% ( ~ s E M ) , serum 84+1.2%, bile 91+2.1%; for sodium [24-' 4C]taurocholate added to the samples from three patients, urine 97%, serum 94%, bile 79%.
Results The bile acids identified and measured by g.1.c. are listed in Table 2 with their retention times relative to methylcholate trimethylsilyl ether (RRT,). The major components of the bile acids in urine, serum and bile, cholic acid, deoxycholic acid, chenodeoxycholic acid and lithocholic acid were present in sufficient concentration to be identified by g.1.c.lma.s spectrometry in most samples. The concentrations of the other bile acids varied widely and those samplescontaining an appreciableamount of these minor components were submitted to g.l.c./mass spectrometry for confirmation of their identity. In other samples, particularly bile, where the concentration of the minor components was usually very small, identification depended on the retention time of the
TABLE 2. Bile acids and their relative retention times A-M are letters used to identify the bile acids in Tables 3, 4 and 5 . Retention times are relative to methylcholate trimethylsilyl ether ( = 1). Relative retention time A
B C D E
F
Systematic name
1 1$20
3a,7a,12a-Trihydroxy-S&cholan-24-oate
1.25 1.40 1.51
230
3~,7a-Dihydroxy-5/kholan-24-oate 3a,6a.7a-Trihydroxy-S~-cholan-24-oate 3a,l2a-Dihydroxy-S~cholan-24-oate 3a,7/?,12a-Trihydroxy-S&cholan-24-oate 3a,7a-Dihydroxy-S/$~holan-24-oate 3a-Hydroxy-Sacholan-24-oate or 3/I-hydroxy-S~-cholan-24-oate 3a,6a-Dihydroxy-Sj&cholan-24-0ate 3a-Hydroxy-Sjkholan-24-oate
2-37 2.56 2.83
3a,7~-Dihydroxy-5&cholan-24-oate 3&Hydroxychol-5-en-24-oate
G
1.53 1.64
H
1.77
I J K L M
2.00
Trivial name Cholate Trihydroxy Hyocholate Deoxycholate Chenodeoxycholate Allo-lithocholate or iso-lithocholate Hyodeoxycholate Li thocholate Dihydroxy Ursodeoxycholate
TABLE 3. Biliary bile acids and bile acid sulphates in cholestasis
Mean+ SD
8
4 5
7
3
2
6
1
Patient no.
339 (3.2) 540 (0.7)
355 (2.5) 1397 (0.6) 5
nd
10 (9.1)
(nd)
(nd)
2
nd
(nd)
(nd)
5
12
21
(W
(nd)
27 (44)
42
nd
nd
nd
nd
nd
(nd)
(nd)
15 (nd) nd
(nd)
19
(nd)
6
(nd) nd
23
nd
11
(2.8) 850 (0.7)
nd
nd
nd
3185 (1.3) 2631
134 (18) 33
C
F
B
A
D
Trihydroxy
(nd) nd
19
(nd) nd
4
267 (3.7) 97 (4.1)
1979 (2.2) 1838 (1.3)
(2.5) 670 (5.1)
(100) 26 (17)
863 (3.2) 1637
G
63 (16) 5
E
2 (46)
3 (45)
nd
27
(nd)
2 (100)
10 (100)
Not available 10 5 (40) (nd)
6
1 (100)
nd
nd
L
(nd)
(nd)
(29) 13
58
(100)
13
K
2 (100) 29 (6.6)
nd
6 (100)
(100)
7 (100) 6
I
Dihydroxy
1 (100)
nd
(100)
21
nd
(100)
3
nd
nd
H
12 (27) 4 (33)
22 (25) 37 (17)
(35) 47 (54)
84
101 (34)
J
M
1 (100)
nd
nd
11 (54)
1 (100)
nd
50 (18)
Monohydroxy
660
674
341 1
2390
1686
4488
4416
Total
3.3k0.9
2.4
4.0
2.1
2.8
4.6
3.7
3.8
sulphated
% of total
Results of the analyses of individual bile acids (pnol/l) are shown for seven patients. The figures in parentheses indicate the percentage of the individual bile acids present as sulphate ester. A-M refer to the bile acids listed in Table 2. nd = not detected.
%
0
&
s
B3
+
4
nd
8
Meanf SD
nd
5
nd nd
nd
4
7
nd
1.5 (nd) 4.1 (33) 1.o (56) nd
nd
L
nd
K
3
I
nd
G
nd
nd
nd
nd
E
2
C
F
D
nd
B
Dihydroxy
nd
A
Trihydroxy
6
1
Patient no.
H J
Monohydroxy
TABLE 4. Serum bile acids and bile acid sulphates in cholestasis Results of the analyses of individual bile acids (,umol/l) are shown for eight patients. See also Table 3.
nd
M 17.0
Total
46
sulphated
% of total
5
58
J. A . Summerfield et al.
compound on g.1.c. This group of compounds will be referred to collectively as the minor bile acids; it includes: hyocholic acid, 3a,7/?,12atrihydroxy-5/?-cholan-24-oic acid, an unidentified trihydroxy bile acid (RRT, 1.20), 3/?,7adihydroxy-5/?-cholan-24-oic acid, hyodeoxycholic acid, ursodeoxycholic acid, one unidentified dihydroxy bile acid (RRT. 2.37), a monohydroxy bile acid (allo- or iso-lithocholic acid) and 3~-hydroxychol-5-en-24-oic acid. Total bile acids
the much larger chenodeoxycholic peak, making measurement impossible (Table 4). The sum of the concentrations of cholic acid and chenodeoxycholic acid was highly correlated with the serum total bile acid concentration (r = 0.99, P < 0.001). However, in contrast to the bile, the sum of the lithocholic acid and deoxycholic acid concentrations ( r = -0.04) and the sum of the minor bile acid concentrations (r = 0.41) was not significantly related to the serum total bile acid concentration and remained very low. Consequently, whereas the molar percentage of cholic acid and chenodeoxycholic acid in the serum increased linearly with the serum total bile acid concentration ( r = 0.94, P < 0.001), the molar percentages of lithocholc acid and deoxycholic acid (r = -0.88, P < 0.01) and the minor bile acids (r = -0.79, Pc 0.05) decreased
Bile. The biliary bile acid analyses of seven patients (Table 3) show that the biliary total bile acid concentration varied between 661 and 4480 ,umol/l. The predominant bile acids were cholic acid (53 f22%; mean percentage composition fSD) and chenodeoxycholic acid (41 ? 23%); in only two instances was chenodeoxycholic acid the major bile acid. Small amounts of lithocholic acid (1.6 +0.8%) and deoxycholic acid (0.90k 1.1 %) were present. The minor bile / acids comprised only 3.7f1.4% of the total. a / 8o / Although a small amount of 3a,7/?,12a-trihya’ droxy-5/?-cholan-24-oicacid was identified in the two samplesexamined by mass fragmentography, it was hidden in the much larger chenodeoxycholic acid peak on g.l.c., making measurement impossible (Table 3). The concentration of total primary bile acids was highly correlated with the biliary total bile acid concentration (r = 0.999, P
