Formation of ursodeoxycholic acid from chenodeoxycholic acid in the human colon: studies

Transcription

1 Formation of rsodeoxycholic acid from chenodeoxycholic acid in the hman colon: stdies of the role of 7ketolithocholic acid as an intermediate Hans Fromm, Rajendra P. Sarva, and Franco Bazzoli with the technical assistance of Ssan Ceiyak and Lawrence Mendelow Gastroenterology Unit, Montefiore Hospital, Department of Medicine, University of Pittsbrgh School of Medicine, Pittsbrgh, P 1 13 bstract The formation of rsodeoxycholic acid from chenodeoxycholic acid and the role of 7ketolithocholic acid as an intermediate in this biotransformation were stdied in vitro in fecal incbations as well as in vivo in the hman colon. [4 C]Labeled 7ketolithocholic and chenodeoxycholic acids were stdied at varios concentrations, and the biotransformation prodcts were analyzed by thinlayer chromatography, gasliqid chromatography, and mass spectrometry. There was rapid colonic conversion of 7ketolithocholic acid to rsodeoxycholic acid and, to a lesser extent, to chenodeoxycholic acid. The redction of 7ketolithocholic to rsodeoxycholic acid proceeded significantly faster anaerobically and at acid ph than nder aerobic and alkaline conditions. When chenodeoxycholic acid was incbated in vitro or instilled into the colon, varios amonts of 7ketolithocholic and rsodeoxycholic acids were formed. The formation of 7ketolithocholic acid was favored by alkaline conditins.l Isotope diltion stdies, in which trace amonts of laheled 7ketolithocholic acid were incbated with nlabeled chenodeoxycholic acid, indicate 7ketolithocholic acid to be the major intermediate in the intestinal bacterial conversion of chenodeoxycholic to rsodeoxycholic acidfmmm, H., R. P. Sarva, and F. Bazzoli. Formation of rsodeoxycholic acid from chenodeoxycholic acid in the hman colon: stdies of the role of 7ketolithocholic acid as an intermediate. J. Lipid Res Spplementary key words chenodeoxycholic acid biotransformation 7ketolithocholic acid formation intestinal bacterial 7ketolithocholic acid redction colonic rsodeoxycholic acid formation Chenodeoxycholic acid (CDC) and its 7/3epimer, rsodeoxycholic acid (UDC), show promise of being sefl in the treatment of cholesterol gallstones (11). CDC, which is synthesized in the liver from cholesterol, is a major bile acid in man (13). In contrast, UDC, which is thoght to be derived from CDC, is sally fond only in small concentrations in hman bile (14). One has assmed that UDC originates in the intestine, since it is absent in bile fistla bile (1 4). Previosly it has been shown in or laboratory that UDC can be formed in the liver from 7ketolithocholic acid (KLC), a ptative intermediate in the conversion reaction from CDC to UDC (1). Ths, one mode of UDC formation cold involve intestinal bacterial oxidation of CDC to KLC, which, in trn is absorbed and redced in the liver to UDC. nother possibility cold be that the entire biotransformation takes place in the colon withot participation of the liver. This has been sggested by in vitro experiments of Federowski et al. (16), in which UDC formation was shown to occr dring fecal incbation of CDC. However, in their stdies, in which they incbated [7/3 H] as well as [414C]labeled CDC, these athors did not identify KLC or any other intermediate (16), Fedorowski et al. (16) therefore conclded that KLC was not involved in the interconversion of CDC and UDC. Instead, they postlated the occrrence of an nsatrated intermediate sch as 6 or 7lithocholenic acid, since H label appeared in UDC after fecal incbation of [ 7/3 H]labeled CDC. lthogh this observation provides strong evidence for the existence of a pathway that involves an nsatrated intermediate, it does not exclde the possibility that varying portions of UDC are also formed via KLC. Several findings by other investigators are in spport of KLC being an important, thogh perhaps not the exclsive, intermediate in the biotransformation of CDC to UDC. First, Midtvedt and Norman (1 7) have shown that several bacterial species, which commonly inhabit the hman colon, are capable of oxidizing CDC to KLC. Secondly, KLC can bbreviations: CDC, chenodeoxycholic acid; UDC, rsodeoxycholic acid; KLC, 7ketolithocholic acid; LC, lithocholic acid; TLC, thinlayer chromatography; GLC, gasliqid chromatography; MS, mass spectrometry; C, cholic acid; DC, deoxycholic acid. ddress reprint reqests to: Dr. Hans Fromm, Montefiore Hospital, 349 Fifth vene, Pittsbrgh, P 113 Downloaded from by gest, on May 14, 18 Jornal of Lipid Research Volme 4,

2 be fond in conditions of bacterial overgrowth in the pper small bowel' and, physiologically, in the colon (1 8). The aims of the present stdy, therefore, were to examine, in vitro and in vivo I) the colon as the likely site of the formation of UDC from CDC and ) the role of KLC in this biotransformation reaction. Materials MTRILS ND MTHODS Nonradioactive CDC was spplied in the form of crystalline powder, from Tokyo Tanabe Co., Ltd. (Tokyo, Japan). The prity of this material was assessed by gasliqid chromatography (GLC). CDC was more than 99% pre, containing less than.1 % lithocholic acid (LC). [414C]CDC (sp act pci/mmol) was prchased from New ngland Nclear Corp. (Cambridge, M). It was more than 98% pre by thinlayer chromatography (TLC). Nonradioactive KLC was prepared by the oxidation of CDC with bffered potassim chromate or with Nbromosccinimide, as described by Fieser and Rajagopalan (1). The melting point was 13 C. KLC was more than 98% chemically pre by GLC. Confirmation of the strctre was obtained by nclear magnetic resonance (NMR) at the Hormel Institte, stin, MN, on a Varian CFT spectrometer, operating in the plsed Forier transform mode at 79.4 MHz. This material was labeled with I4C at the 4 position by halodecarboxylation followed by reaction with [ 14C]cyanide (). The specific activity was. mci/mmol. The synthesized [ 4I4C]KLC was prified by preparative TLC. The final prity was more than 99%. In vitro aerobic and anaerobic incbation stdies of labeled KLC and CDC Immediately after evacation, fresh stool specimens were obtained from for male and one female healthy volnteers, as well as from three female and two male patients with asymptomatic gallstones. Stool samples were homogenized with normal saline (approximately 1: 1 v/v). [414C]CDC and [414C]KLC, respectively, were incbated simltaneosly in different vials with the fresh stool homogenates at a temperatre of 37 C nder aerobic and anaerobic conditions in a Dbnoff incshaker (LabLine Instrments, Melrose Park, IL). The individal incbation reactions were terminated in the different vials after,., 1, 4 and 1 hr, respectively. The and Ihr incbation experiments were carried ot in dplicate. The maintenance of the aerobic and anaerobic conditions, respectively, was monitored with a disposable anaerobic indicator (GasPak, Becton, Dick ' Bolt, M. G., University of Chicago. Personal commnication. inson and Co., Cockeysville, MD) sspended in the vials above the incbation media. In control experiments, this indicator was fond to be reliable in discriminating between aerobic and anaerobic conditions. It trned ble if the caps of the incbation vials were loosened for the prpose of exposing the media to atmospheric oxygen (aerobic conditions), bt showed no color change if the samples were kept nder a nitrogen stream before the vials were tightly capped (anaerobic conditions). KLC and CDC were incbated at concentrations of.4 f.1 mm and.46 f.6 mm (mean k SM), respectively. These concentration figres encompass both the endogeneos and the exogenosly added nlabeled KLC and CDC, respectively. The endogenos fecal bile acid concentrations were as follows: CDC,.19 f.6 mm; UDC,.1 f.3 mm; Lc, 1.18 f.6 mm; DC, 1.89 f.4 mm; and c,.4 rf:.3 mm. No measrable qantity of endogeneos KLC was identified in any of the fecal samples. For these incbation experiments, stock soltions of [414C]CDC and [414C]KLC, respectively, were prepared. The respective isotope was dissolved in.1 M sodim pyrophosphate bffer at a concentration appropriate for the incbation experiment to be performed. erobic incbations Soltions of [14C]KLC and [14C]CDC, respectively, ( pl) were pipetted into sterile vials and mixed with 7 ml of sterile normal saline. This soltion was then mixed with 1 g of fresh stool homogenate. Normal saline was sed in order to dilte the samples and assre good mixing of the labeled bile acids with the fecal material. The vials were capped and the incbation media were then mixed in a testtbe stirrer (Vortex Genie, Scientific Indstries, Springfield, M). Sbseqently, the caps were loosened to expose the media to atmospheric oxygen. In order to stdy the inflence of ph on the biotransformation reactions, the incbations were carried ot at a ph ranging from.1 to 9.. The stool ph was adjsted with 1 N NaOH. Two types of incbation experiments were then performed. In one the stool ph was not readjsted dring the incbation period. In these experiments the ph dropped by.9 f.17 within 4 hr. In the second type of stdy the ph was kept constant throghot the incbation. The ph was monitored sing a gelfilled combination ph electrode (Orion Research, Cambridge, M, Model 91), which was placed into one of the incbation vials. If the ph changed, readjstment to the original vale was effected by addition of 1 N NaOH. The same amont of NaOH necessary to keep the ph constant in the phmonitored medim was added to the other incbation vials. The final ph in the different vials, which was recorded in all incbation stdies, showed only minor vari Downloaded from by gest, on May 14, Jornal of Lipid Research Volme 4, 1983

3 ations (mean difference from mean ph vale was.1 f.1). The incbations were terminated at the different time intervals by addition of concentrated HCI. naerobic incbations Homogenized fresh stool was kept nder a nitrogen stream. I4CLabeled KLC and CDC, respectively, were pipetted into sterile vials and mixed with 7 ml of preredced anaerobically steril.led chopped meatglcose medim (Scott Laboratories, Fiskeville, RI). This medim was then mixed with 1 g of fresh stool homogenate. These preparations of the incbation medim were carried ot nder a stream of nitrogen. Before capping, the vials were flshed with nitrogen. Termination of the incbations at the different time intervals was effected by addition of. ml of 33% KOH and 1 ml of absolte ethanol (16). No attempt was made in these anaerobic stdies to keep the ph constant. The ph vales at the beginning of the incbations ranged from In vitro isotope diltion stdies In order to stdy the role of KLC as an intermediate in the biotransformation of CDC to UDC, six series of isotope diltion stdies were performed. Trace amonts of [414C]KLC and nlabeled CDC were first mixed together and then mixed with fresh stool homogenate, as described above. The total concentration of CDC at the beginning of the incbations was. f.1 mm. The endogenos fecal bile acid concentrations in the fecal incbates were as follows: CDC,.1 1 f.1 mm; UDC,.8 f.4 mm; LC,. f.1 mm; DC,.6 f.19 mm; and C,.3 f.3 mm. One of the for sbjects from whom the fecal samples for the isotope diltion stdies were obtained, a healthy volnteer, showed nsally high endogenos concentrations of CDC and UDC. The latter represented 43% and IS%, respectively, of the total fecal bile acids. In the other three sbjects, CDC constitted always less than 1 % and UDC less than 7% of the fecal bile acids. erobic incbations were carried ot for,., 1, 3,, 7, 9, 11, and 4 hr with fecal samples of two female gallstone patients at native ph vales of 6. and 7.14, respectively. In for other experiments, fecal samples from two healthy male volnteers were incbated at a native ph of.4 and 6., respectively, and at ph vales that were adjsted to 7.1 and The specific activities of CDC, KLC, UDC, and LC were derived from the GLC measrements of the mass and the TLC determinations of the proportional radioactivity of the respective bile acids (vide infra). The precrsorprodct relationships were evalated sing specific activity time crves of these componds (3). In the calclations of the specific activities, the bile acids that were endogenosly present in the fecal incbates were considered. The calclation of the specific activities of CDC was based on the mass of both the endogeneos and exogenos CDC, since both can be expected to participate in a comparable fashion in the biotransformation reactions. For the comptation of the specific activities of KLC, UDC, and LC, the endogenos components were sbtracted from the respective total measrements. This treatment of the data was thoght to provide the best approximation for the specific activities of CDC, KLC, and UDC. However, the calclated specific activities of LC are probably slightly lower than the tre vales, since they also reflect the formation of LC from endogeneos UDC. In vivo stdies Two female and two male sbjects (Tables 1 and ) were admitted to or inhospital Cooperative Care facility. Rotine laboratory stdies, inclding liver fnction tests, were normal in these sbjects. n orocolonic tbe with a bag containing.7 cm3 of mercry at its tip was passed (4). In three of the for sbjects, a singlelmen tbe (Tygon R363, ID 1. mm, OD 3. mm (Norton Plastics and Synthetics Division, kron, OH) and, in the forth, a doblelmen tbe (ID 1. mm X, OD 4. mm) was sed (Table 1). The tip of the tbe was placed into the colon. second orointestinal tbe was passed with the tip in the second portion of Downloaded from by gest, on May 14, 18 TBL 1. In vivo biotransformation of 14Clabeled 7ketolithocholic acid dring steadystate perfsion of ascending colon with a doblelmen tbe Concentrations ph of Radioactive Metabolite (% on TLC) Bile cid Composition by GLC (mm) Sbject, of Infsed Collected Sex, ge KLC (mm) Sample KLC CDC UDC LC KLC CDC UDC LC C DC M.L., F, (Control) yr o ight ml per minte steadystate infsion of sodim KLC,., 1., and 1. mm, respectively; glcose, 1 mm and NaCI, 3 mm. Dring the infsion of each of the three KLC soltions, two 1min collections were obtained following a 1min eqilibration period. The samples were collected in the ascending colon 1 cm distal to the infsion site. Fromm, Saroa, and Baoli Formation of rsodeoxycholic acid in the colon 843

4 the dodenm. The position of both tbes was confirmed by an abdominal roentgenogram. The colonic ph vales ranged from 6. to 6.8. t 8:OO M of the day of the biotransformation stdy, 1 pci (1 mmol) of 14Clabeled sodim CDC and KLC, respectively, (1 mm soltion in normal saline) were instilled throgh the orocolonic tbe into the colon. In one of the for sbjects, the colonic instillation of the 1 mm soltion of [I4C]KLC was preceded by a steadystate perfsion stdy of three different concentrations of this compond (Table 1). The sbjects were fasting for 1 hr prior to the experiment. Following the colonic instillation of the labeled precrsor, the tbe was flshed with normal saline. Colonic aspirations were performed horly. When material was obtained from the colon, it was either freshly analyzed or immediately frozen at C for later analysis. Gallbladder contraction was effected by i.v. injection of. pg/kg of KinevacB, and approximately 1 ml samples of bile were obtained throgh the dodenal tbe. Individal stool collections were made over a period of 48 hr (Table ) and either analyzed freshly after each bowel movement or frozen immediately at C for later analysis. TLC analysis of chemical forms of radioactivity The bile acids in the in vitro and in vivo fecal samples were extracted by a previosly established method (4). s described in detail elsewhere, the bile acids in the in vitro fecal incbates were extracted with mberlite XD7 resin (Polysciences, Inc., Washington, P), prified by percolation throgh a Florisil colmn, and esterified sing ethereal diazomethane (4). The recovery (following these extraction and prification steps) of the total radioactivity incbated as [14C]CDC and [14C]KLC, respectively, was 9 f 1.%. The in vivo fecal samples (stool and colonic aspirates) were first sbjected to alkaline hydrolysis and then processed identically to the in vitro incbates (4). The biliary bile acids were analyzed as previosly described (1,, 6). The chemical form of the radioactivity of the extracted bile acid methyl esters was determined by TLC, sing chloroformacetonemethanol 7:3: (v/v) as a solvent system (1 ). The distribtion of the radioactivity on the TLC plates was determined by zonal scraping of the silica gel. Complete zonal scraping was carried ot on all TLC plates. The radioactive bile acid metabolites were identified by relating the distribtion of the radioactivity to that of pre reference standards (1 ). The recovery of the radioactivity from the TLC plates was 98 f 1.%. GLC and mass spectrometry (MS) analyses of bile acid composition The in vivo fecal samples and in vitro fecal incbates were also analyzed by GLC for nlabeled metabolites 844 Jornal of Lipid Research Volme 4, 1983 of CDC and KLC. In addition, in vitro fecal incbates were analyzed by MS. For GLC determinations, nordeoxycholic acid was sed as internal standard for bile acid qantification. fter the described methylation step, the bile acids were acetylated (1, 46). The methyl ester acetates were dissolved in dimethyl formamide and analyzed by GLC sing a flame ionization detector (Gas Chromatograph, Model 4 1, Packard Instrment, Downers Grove, IL): 1.&meter Ucolmns, mm ID, packed with 3% N6 on gaschrom Q 1 1 mesh (Spelco, Inc., Spelco Park, Bellefonte, P), (1, 46). In the in vitro stdies, correction was made for the endogenosly present bile acids. The mass of the bile acids present in the stool samples before addition of the respective precrsor was sbtracted from that measred afterwards at the different times of incbation. For preparation of the samples for MS, the individal bands representing CDC, KLC, and UDC were scraped off the TLC plates. The bile acids were elted from the silica gel with methanol. Following preparation of the bile acid methyl ester acetates, GLC/MS analysis was performed by Dr. rwin H. Mosbach, Director, Lipid Research Laboratory, Beth Israel Medical Center, New York, on a HewlettPackard Model 99B GLC/MS. Statistical analysis The paired comparison ttest was sed for the statistical evalation of the effect of different conditions (ph, aerobic verss anaerobic) on in vitro bile acid biotransformation. The correlation between fecal ph and in vitro KLC formation from CDC was statistically analyzed by least sqares regression. RSULTS Colonic biotransformation of KLC In vitro fecal incbation of labeled KLC. Varios proportions of 14Clabeled KLC were redced to both UDC and CDC in the aerobic as well as in the anaerobic fecal milie (Figs. 13). In most incbations more UDC than CDC was formed. The time crves of the reaction nder aerobic conditions at a native ph ranging from.1 to 6. are shown in Fig. 1. fter 4 hr of incbation, only abot 3% of the radioactivity was still present in KLC. fter 1 hr, this figre had decreased to approximately 13%. The peak of UDC formation, which averaged abot 3% of the original KLC radioactivity, occrred sally at 4 hr. s the radioactivity of KLC declined, there was progressive accmlation of LC (Fig. 1). The redction of KLC to UDC proceeded significantly faster at acid than at alkaline ph levels (Fig. ). The rate and pattern of biotransformation of KLC were similar in experiments in which the ph of the incbation medim Downloaded from by gest, on May 14, 18

5 1 I l z * c >_ I :! n a LL ac a HOURS OF INCUBTION Fig. 1. erobic in vitro fecal biotransformation of ''Clabeled KLC at native ph ranging from.1 to 6. in three control and one gallstone sbjects (mean f SM). total of 11 incbation stdies of KLC were carried ot at a concentration of.4 f.1 mm. was only initially adjsted to higher levels and in those in which the higher ph was maintained throghot the incbations. naerobic conditions were significantly more condcive than aerobic conditions to the redction of KLC to UDC, both at lower and higher fecal ph vales (Fig. 3). The identity of KLC, UDC, and CDC bands separated by TLC was confirmed by GLC/MS. No attempt was made to identify the strctre of componds that elted from the TLC plates with KLC, CDC, and UDC, bt they constitted less than 1% of the respective total bile acid. These componds possibly represented 38 hydroxy epimers (7, 8). In vivo infsion oflabeled KLC into colon. The biotransformation of KLC shown in vivo after infsion into the colon of two hman sbjects (Table 1 and Table ) was comparable to that fond in the in vitro incbation experiments (Figs. 13). The proportion of KLC converted in vivo in the colon to UDC within 1 hr ranged from abot 6% to 77%. This rate of UDC formation in vivo (Tables 1 and ) resembled the rate observed in vitro nder anaerobic conditions more closely than that nder aerobic conditions (Fig. 3). The reslts of the stdies obtained by TLC analysis of the radioactive precrsor and metabolites were congrent with those of the GLC analyses of the corresponding nlabeled componds. Colonic biotransformation of CDC In vitrofecal incbation oflabeled CDC. In the anaerobic fecal incbation experiments, abot 9% of CDC was metabolized to LC within 1 hr (Fig. 4). In addition, varios proportions of CDC were transformed to KLC and UDC (Fig. 4). The correlation between percent of radioactivity as KLC on TLC and fecal ph was significant for the aerobic incbation stdies (n = 3, r =.679, P <. 1) as well as the aerobic and anaerobic incbations combined (n = 33, r =.98, P <.1). However, the correlation between fecal ph and KLC formation was not significant if only the data of the ten anaerobic incbation series were sed for the calclation. The reslts of the TLC and GLC analyses were, again, confirmed by MS. In vitro incbation of nlabeled CDC with trace amonts of labeled KLC (isotope diltion stdies). s shown in Fig., there was rapid biotransformation of significant Downloaded from by gest, on May 14, 18 Fromm, Sara, and Bamli Formation of rdeoxycholic acid in the colon 84

6 W PH *. NUMBR OF INCUBTION SRIS: 8 i I ** L Dc LC CDC 4 a HOURS OF INCUBTION LC KLC CDCUDC LC Fig.. ffect of alkalinization on in vitro aerobic fecal biotransformation of ''Clabeled KLC in ten comparative incbation series of samples from one control and one gallstone sbject (mean f SM). The KLC concentration at the beginning of the incbations was.41 f.1 mm. *, Significant at P <.; **, significant at P <.. amonts of nlabeled CDC to KLC, UDC, and LC in the six series of isotope diltion stdies sing incbations of fecal samples from for sbjects. The biotransformation of the I4Clabeled KLC, which was incbated in trace amonts in these stdies, was similar to that shown in other experiments in Fig. 1. fter 1 hr, 36 f 3.6% of the radioactivity incbated as KLC appeared in UDC, 16 f.% in CDC, 4 f.8% in LC, and 4 k 3.8% in the original compond. The corresponding percentages of radioactivity after hr of incbation were 6 f 3. in UDC, 8 f 3. in CDC, 6 f.9 in LC, and 1 f 3. in KLC. The time crves of the specific activities of CDC, KLC, UDC, and LC in these stdies are shown in Fig. 6. rapid decline in the specific activity of KLC was accompanied by a rise in that of UDC, CDC, and LC. In every stdy, both at acid and alkaline fecal ph levels, the KLC intersected with the UDC and/or LC crves. In vivo infsion oflabeled CDC into the colon. Formation of KLC and UDC from CDC was also observed in vivo in the two sbjects in whom [14C]CDC was infsed into the transverse colon (Table ). It is of interest that one of the two sbjects (W.D., who had severe bile acid malabsorption de to a distal ileal resection) was fond, prior to the infsion of CDC, to have not only this bile acid bt also KLC and UDC in considerable concentrations in the stool. The formation of KLC and UDC in the colon was then frther docmented by fecal analysis of the radioactive metabolities after colonic infsion of 14Clabeled CDC. In addition, after CDC was instilled into the colon, there was a marked rise in the colonic concentration of nlabeled KLC and UDC. In both sbjects, in the patient with ileal resection as well as in the one with asymptomatic gallstones, 1% and 17%, respectively, of [14C]CDC were converted to [14C]UDC in hr. The corresponding figres for fecal KLC for Downloaded from by gest, on May 14, Jornal of Lipid Research Volme 4, 1983

7 ph 6. *** t ROBIC I WROBIC *** T i' 1 1 KLC COC UOC LC KLC CDC UDC LC KLC COC UDC LC ph 7.9 *** *** 1 *** T Downloaded from by gest, on May 14, 18 KLC COC UDC LC KLC CDC UOC LC KLC COC ric LC. I 4 HOURS OF INCUBTION Fig. 3. In vitro comparison of aerobic and anaerobic biotransformation of trace amonts of '4Clabeled KLC at a ph of 6. and 7.9, respectively, in a fecal sample of a gallstone sbject. t both ph levels, aerobic and anaerobic incbation series were carried ot at CDC concentrations of.17,.33,.49 and.6 mm, respectively. *, Significant at P <.; ***, significant at P <.1. mation dring the same time interval were 7% and %, respectively. (Table ). Biliary bile acids following colonic infsion of KLC and CDC, respectively t 1. hr after colonic infsion of [I4C]KLC in sbject M.L., 3% of the radioactivity appearing in dodenal bile was fond in CDC, 4% in UDC, 7% in LC, and 16% nchanged in KLC (Table ). The corresponding figres for the chemical form of the radioactivity in bile, 1 hr after colonic infsion of [14C]CDC in sbject W.D., were 77% for CDC, 6% for UDC, 1% for LC, and % for KLC (Table ). similar distribtion of the radioactivity in biliary bile acids was fond in the third sbject, from whom dodenal bile was obtained after co Fromm, Sara, and Bamli Formation of rsodeoxycholic acid in the colon 847

8 Q.e L C 8 I C M : W z x I m e c x i m " + t B 3 I 3 Lr C I : $ c! B? > C oi crl cl m < b L *cc (n : 1 + x= =z.i : B p S y 9.c 3 x : a a e, d 8 L* cc 3 g 3 3 g ZG 3 st.g s 3 r gs 8 i! & %,8 UP. $; U c =. e.o 8.3 nzgz m L % 39? gm ' P f Q<% %c 4.g $ b : 3. f v ) v) mt I I I l l I I l l $ I I Ill+" ""9 zz 1 1 s m( ".1 mmm m a e m * i? x z m I I 199 IF c: momd=j maw om li, 14'9 Jlm ooozz e mwm I I I $gt % I I Idd Z Z % ; M GVmz mwm.. wmw mo ;:.1mm.1 "?;% q;; mm X"ZZ.14 h 3 Y no m m a a m w a mmm m : mt I IO&+ amp mw 8% c?qf e.^ t w mow w o gn mwm a em m.iq, mm=?a :G% lnm mww I e"? e YC?? '9 m* e ooozz II.3 I I Im I I 1. 1 I l l I l l I I z I: j._ z '. mmt omm m 4 n IZ I t m O.l t LC 1 1: Immmt mm I W Z m w w a 1. 1 b? ormt mm.1 lm.1 t IZZ 1 loo ; 1 C jz B rnz mt I I%% w Imm It Izz%= zmm I'.l i.z Y a c Q tg L v La x $ m l I 1 e4 I I l t l ' 4r 3 I Immn 1 1 IC.( z I X C 3". *. zz. dj?c???? e?? 44 G zzz zz Dppp rmc 1 : h " " = " e g e I e e e cw c.?o g c g???w c??m * I mmc.( owm om gn C Y C C C 1 B m m L ; $ 4 UP) g <z rzz z 4 Z B c z <z 3. p s a bgi x n Gi n ; Y P % % n $._ L 8 L U M X P m x M O ms Q : rc %.1 x G.; rz; % 3s <. :$ Z p t' e 3: U cr 3 p Y. 9 sb, $ :,sa z < ".yc yje g g m n.y 3 L B x oz I._ Zd I.G I. cr C Y p. 8 x e m c. T e, L e, c t' c % rnm & M p '= c W c % j, $3.s W y ; c a z 34 L CLr O.4!.p ~+.o W " cnw I sa a. C S M l 9 >m Downloaded from by gest, on May 14, Jornal of Lipid Research Volme 4, 1983

9 IO. I 4 HOURS OF INCUBTION Fig. 4. In vitro anaerobic fecal biotransformation of ''Clabeled CDC in eight sbjects, in whom a total of ten incbation series were carried ot. The CDC concentration at the beginning of the incbation was.46 f.6 mm. lonic instillation of a labeled bile acid. In this sbject, C.C. (Table ), dodenal bile was obtained 4 hr after instillation of [ 14C]CDC into the colon. Seventyseven percent of the ''Clabel in bile appeared in CDC, 8% in UDC, 13% in LC, and % in KLC. DISCUSSION Previosly it has been shown in or laboratory that KLC, a ptative intermediate in the conversion of CDC to UDC, is absorbed in the small intestine and redced in the liver to CDC and, to a lesser degree, to UDC (1). In the present stdy, this biotransformation reaction was, for the first time, also fond to take place in vivo in the hman colon. KLC was rapidly biotransformed both in in vitro fecal incbations and in vivo after colonic instillations. It is of note that, in contrast to the reaction in the liver (1), more KLC was transformed to UDC than to CDC in the colon. These findings are in agreement with in vitro anaerobic fecal incbation stdies by Higashi, Setogshi, and Kazki (9). One reason for the appearance of more UDC than CDC dring the colonic biotransformation of KLC cold be that intestinal bacterial enzymes preferentially catalyze a 76redction of this compond. nother reason cold be that CDC is degraded more rapidly than UDC to LC. Previos stdies in or laboratory indicate the former possibility to be the more likely one, since we showed that the 7dehydroxylations of CDC and UDC to LC are, in most cases, very similar (3). There are few sorces of information in the literatre that relate to or observations regarding the effects of aerobic conditions and changes in fecal ph on the biotransformation of KLC. Both factors significantly inflenced the reaction; more UDC was formed at an acid ph (ph.1 to 6.) and in an anaerobic milie than nder alkaline (ph 7. to 8.7) and aerobic conditions. On the other hand, the formation of KLC from CDC was favored by an alkaline fecal ph. These findings are consistent with stdies by Macdonald and Roach (31), which showed the ph optimm to be higher for the 7ahydroxysteroid dehydrogenase than for the 76h~droxysteroid dehydrogenase. 1 Downloaded from by gest, on May 14, 18 Fromm, Sara, and Baoli Formation of rsodeoxycholic acid in the colon 849

10 * It a X CDC * c b X KLC UDC LC 4 * 4 4 x o X CDC KLC UOC LC CDC KLC UDC LC CDC KLC UDC LC I 9 4 HOURS OF INCUBTION Fig.. Time crves of the fecal biotransformation of nlabeled CDC to KLC, UDC, and LC in the isotope diltion stdies, which were performed in six incbation series on fecal samples of two male control sbjects and two female asymptomatic gallstone patients. The fecal CDC concentrations at the beginning of the incbations ranged from.13 to. mm. The stdies in the two control sbjects were carried ot at both acid (open sqares and triangles) and alkaline (filled sqares and triangles) ph levels. Similar to Salen et al. (3), we had previosly observed significant increases of biliary UDC in gallstone patients treated with CDC (). In several of the patients, the UDC content exceeded % of the total bile acid pool. This sggested intestinal bacterial transformation of CDC as the major sorce of UDC formation. The liver is only involved in the formation of that portion of biliary UDC that is prodced by hepatic redction of KLC. This portion can only be sm.all, since the majority of KLC is converted in the liver to CDC and less than 1% to UDC (1 ). The present stdy also represents, to or knowledge, the first in vivo demonstration of the biotransformation of CDC to UDC and its effect on biliary bile acid composition. The biotransformation of CDC, after infsion of this compond into the colon, differed considerably between the two sbjects stdied (Table ). Consistent with or previos observations in patients with diarrhea de to ileal resection and bile acid malabsorption (4), the fecal bile acids in sbject W.D. contained a relatively small pro portion of lithocholic acid. It is of interest that, in spite of this apparent depression of 7dehydroxylation, other bacterially catalyzed reactions, sch as the dehydrogenation of the aoh sbstitent at C7 and the stereospecific redction of the newly formed oxo moiety to a 7POH grop, were active, as evidenced by the presence of sbstantial amonts of KLC and UDC in feces. Therefore, marked increases in the fecal CDC concentrations, as they were present in this patient to a level of 1.7 mm, do not appear to sppress bacterial 7a or 78hydroxysteroid dehydrogenase activity. These in vivo observations are consistent with or in vitro incbation stdies, in which changes in the fecal CDC concentration from.3 to.93 mm did not reslt in any noticeable depression of KLC or UDC formation. lso, the deconjgation reaction was apparently normal since the fecal bile acids were present in the deconjgated form. The second sbject stdied, who except for having asymptomatic gallstones was healthy, showed both a normal rate of LC formation and the appearance 34 X Downloaded from by gest, on May 14, 18 8 Jornal of Lipid Research Volme 4, 1983

11 ~ pns. C QH.4 pm6.. *...*... * MI 3? 9 II 4 I S 9 4 I S 9 B 'OoO I DM 14 Fig. 6. Time crves of specific activities of '"Clabeled KLC, UDC, CDC, and LC after in vitro fecal incbation of trace amonts of "Clabeled KLC with nlabeled CDC (isotope diltion stdies). ach set of crves represents the reslts of one incbation series. total of six incbtion series was carried ot on fecal samples of for sbjects (see also legend of Fig. for incbation conditions). The biotransformation rate and pattern of the nlabeled CDC for each stdy is shown in Fig.. The solid line represents the ["CIKLC, the dotted line the ["C]cDC. the dashed line the ["CILC, and the dotteddashed line the ['"CIUDC specific activity. The specific activity (DPM X mg' X IO') is deplcted on the yaxes and hors of incbation on the xaxes. Downloaded from by gest, on May 14, 18 of sizable amonts of UDC. However, no KLC was identified in the fecal samples, ths indicating a very rapid redction of KLC to UDC as well as CDC and/ or the involvement of an intermediate other than KLC. The composition of radioactively labeled bile acids in dodenal bile after colonic infsion of [ I4C]KLC and [ ''C]CDC, respectively, is consistent with previos stdies in or laboratory (1 ). The radioactive bile acids in bile represent the prodct of bacterial biotransformation, colonic absorption, hepatic biotransformation and biliary excretion of the respective labeled bile acid instilled into the colon. fter colonic infsion of [I4C]KLC, the percentage of the label appearing in CDC was higher in bile than in the colon (Table Z), since a major portion of KLC that is absorbed is converted in the liver predominantly to CDC (1). The proportion of the I4Clabel fond in biliary CDC was also, as expected, very high when [I4C]CDC was infsed into the colon. In this instance, the percentage of the biliary radioactivity fond in CDC is determined by the rate of both the colonic absorption and the intracolonic biotransformation of ['4C]CDC (CDC is not altered at the steroid ncles dring its passage throgh the liver (1 )). Or in vivo observations are in agreement with or previos stdies of the hepatic metabolism of KLC, CDC, and UDC (1), as well as with the in vitro data Fromm, Sara, and Baoli Formation of rsadeorycholic acid in the colon 81

12 of or own stdy and those of most other investigators who have stdied the bioconversion of CDC in fecal incbation systems or intestinal bacterial cltres (1 6, 17, 6, 7, 33). However, neither or GLC nor TLC systems allowed a separation of 3Phydroxy epimers from the corresponding normally occrring Sahydroxy bile acids. 3PHydroxy epimers, sch as isolithocholic acid, have been isolated in feces by other investigators (7, 8). ccording to Hirano, Masda, and Oda (7), isolithocholic acid may constitte p to 1% of the fecal metabolites of CDC and UDC. Or nearly complete recovery of the total radioactivity on the TLC plates in the LC, KLC, CDC, and UDC bands is consistent with the observation on GLC systems that the corresponding 3a and 3Phydroxy epimers elte closely to each other (7). In other words, isolithocholic acid and other 3Phydroxy epimers, if they occr, wold be measred together with the respective Sahydroxy componds. The latter does, however, not affect the main prpose of or stdies, namely the evalation of the role of KLC as an intermediate in the conversion of CDC to UDC. The ability of several anaerobic as well as aerobic intestinal bacterial species to oxidize CDC to KLC was first reported by Midtvedt and Norman (1 7). More than a decade later, Fedorowski et al. (16) showed that the entire biotransformation of CDC to UDC can take place in the mixed bacterial milie of anaerobically incbated stool samples. lthogh these athors did not identify any KLC dring the formation of UDC, we and other athors (6, 7, 33) showed this ptative intermediate to be formed in the corse of this reaction. However, in agreement with Fedorowski et al. (16), we initially sspected that KLC is neither the only, nor necessarily the most important, intermediate in the conversion of CDC to UDC (6). This sspicion grew ot of in vitro and in vivo biotransformation stdies of labeled as well as nlabeled CDC, in which the amont of KLC formed was observed to be considerably smaller than that of UDC. However, in the interpretation of this finding, one has to consider the complex interaction of the mltiple enzymatic reactions that are involved in the intestinal bacterial metabolism of CDC. The reversible reactions CDC = KLC * UDC are not only inflenced by the comparative availability and reactivity of specific dehydrogenases and redctases in the colon, bt also by the velocity of the irreversible bacterial dehydroxylation of CDC and UDC to LC. The isotope diltion experiments carried ot in this stdy, in which trace amonts of [I4C]KLC were incbated with nlabeled CDC, show a precrsorprodct relationship between CDC, KLC, UDC, and LC. In every experiment, the specific activity crve of [ 14C]KLC intersected with that of [I4C]UDC and/or [I4C]LC. n intersection of the [I4C]KLC and [14C]UDC specific activity crves can be expected to occr if the formation of KLC proceeds at a rate similar to that of its transformation to UDC and LC. In contrast, the specific activity crve of [I4C]KLC intersects only with that of [I4C]LC if the velocity of the KLC UDC LC is higher than the CDC + KLC reaction. It is noteworthy that the precrsorprodct relationship between CDC, KLC, and UDC was evident at acid as well as at alkaline fecal ph levels. These findings are consistent with KLC being the major intermediate in the intestinal bacterial conversion of CDC to UDC, regardless of whether fecal ph is in the acid or alkaline range.l This paper was presented in part at the nnal Meeting of the merican Federation for Clinical Research in Washington, 198, and at the nnal Meeting of the merican Gastroenterological ssociation in Salt Lake City, 198. The stdy was spported in part by grant M19689 from the National Instittes of Health. Dr. Fromm was the recipient of a Research Career Development ward, M9, from the National Institte of rthritis, Metabolism and Digestive Diseases. The athors are indebted to Dr. rwin H. Mosbach, Jr., Director, Lipid Research Laboratory, Beth Israel Medical Center, New York, for the mass spectrometric analyses, and to Dr. Robert. Yee, Gradate School of Pblic Health, University of Pittsbrgh, Dr. Gerald L. Carlson, Dermal Research Department, S. C. Johnson and Sons, Inc., Racine, WI, and Dr. Warren F. Diven, Departments of Pathology and Biochemistry, University of Pittsbrgh, for helpfl discssions. The skillfl technical assistance of Praflla min and Yvonne Korica is grateflly acknowledged. Manscript received 7 Jne 198 and in reoised form 18 Janary RFRNCS Danzinger, R. G.,. F. Hofmann, L. J. Schoendfield, and J. L. Thistle Dissoltion of cholesterol gallstones by chenodeoxycholic acid. N. ngl. J. Med Thistle, J. L., and. F. Hofmann fficacy and specificity of chenodeoxycholic acid therapy for dissolving gallstones. N. ngl. J. Med. 89: 669. her, J. H., R. H. Dowling, H. Y. I. Mok, and G. D. Bell Chenodeoxycholic acid treatment of gallstones: a followp report and analysis of factors inflencing response to therapy. N. ngl. J. Med. 93: Fromm, H.,. schler, D. Tollner, H. Canzler, and F. W. Schmidt In vivo dissolving of gallstones: the effect of chenodeoxycholic acid. Dtsch. Med. Wochenschr Coyne, M. J., G. G. Bonorris,. Chng, L. I. Goldstein, D. Lahana, and L. J. Schoenfield Treatment of gallstones with chenodeoxycholic acid and phenobarbital. N. ngl. J. Med. 9: Barbara, L.,. Roda,. Roda, C. Sama, D. Festi, G. Mazzella, and R. ldini The medical treatment of cholesterol gallstones: experience with chenodeoxycholic acid. Digestion. 14: Dowling, R. H.,. F. Hofmann, and L. Barbara Workshop on Ursodeoxycholic cid. MTP Press Limited, Lancaster. Downloaded from by gest, on May 14, 18 8 Jornal of Lipid Research Volme 4, 1983

13 8. Fromm, H Ursodeoxycholic acid for gallstone dissoltion: the emergence of a new therapetic application for an old bile acid. In Gallstones. M. M. Fisher, C.. Goresky,.. Shaffer, and S. M. Strasberg, editors. Plenm Press, New York Nakagawa S., I. Makino, T. Ishizaki, and I. Dohi Dissoltion of cholesterol gallstones by rsodeoxycholic acid. Lancet 11: Maton, P. N., G. M. Mrphy, and R. H., Dowling Ursodeoxycholic acid treatment of gallstones. Doseresponse stdy and possible mechanism of action. Lancet. 11: Salen G.,. Colalillo, D. Verga,. Bagan, G. S. Tint, and S. Shefer ffect of high and low doses of rsodeoxycholic acid on gallstone dissoltion in hmans. Gastroenterology Nakayama, F Oral cholelitholysischeno verss rso: Japanese experience. Dig. DiS. Sci. : Danielsson, H., and T. T. Tchen Steroid metabolism. In Lipids, Steroids, and Carotenoids. D. M. Greenberg, editor. cademic Press, New York Hellstrom, K., and J. Sjovall On the origin of lithocholic and rsodeoxycholic acids in man. cta Physiol. Scand. 1: Fromm, H., G. L. Carlson,. F. Hofmann, S. Farivar, and P. min Metabolism in man of 7ketolithocholic acid: a precrsor of chenodeoxycholic and rsodeoxycholic acid. m. J. Physiol. 39 G161Gl Fedorowski, T., G. Salen, G. S. Tint, and. Mosbach Transformation of chenodeoxycholic acid and rsodeoxycholic acid by hman intestinal bacteria. Gastroenterology. 77: Midtvedt, T., and. Norman Bile acid transformation by microbial strains belonging to genera fond in intestinal contents. cta Pathol. Microbiol. Scand. 71: li, S. S.,. Kksis, and J. M. R. Beveridge xcretion of bile acids by three men on a fatfree diet. Can. J. Bwchem. 44: neroth, P., B. Gordon, R. Ryhage, and J. Sjovall Identification of mono and dihydroxy bile acids in hman feces by gasliqid chromatography and mass spectrometry. J. Lipid Res. 7: Knodell, R. G., D. Kinsey,. C., Boedeker, and D. P. Collin Deoxycholate metabolism in alcoholic cirrhosis. Gastroenterology. 71: Fieser, L. F., and S. Rajagopalan Selective oxidation with Nbromosccinimide. I. Cholic acid. J. m. Chem. SOL. 71: Carlson, G. L., and H. Fromm Preparation of radiolabelled 3ahydroxy7keto&cholanic acid and its glycine and tarine conjgates. J. Labelled Compd. Radwpharm. XVI, 3: Reiner, J. M Recent advances in moleclar pathology. Isotopic analyses of metabolic systems. xp. Mol. Pathol McJnkin, B., H. Fromm, R. P. Sarva, and P. min Factors in the mechanism of diarrhea in bile acid malabsorption: fecal pha key determinant. Gastroenterology Fromm, H., H. C. rbler,. schler, and F. W. Schmidt lterations of bile acid metabolism dring treatment with chenodeoxycholic acid. Stdies of the role of the appearance of rsodeoxycholic acid in the dissoltion of gallstones. Klin. Wochenschr. 4: Sarva, R. P., H. Fromm, G. L. Carlson, L. Mendelow, and S. Ceryak Intracolonic conversion in man of chenodeoxycholic acid (CDC) to rsodeoxycholic acid (UDC) with and withot formation of 7ketolithocholic acid (KLC) as an intermediate. Gastroenterology. 781 (bstract). 7. Hirano, S., N. Masda, and H. Oda In vitro transformation of chenodeoxycholic acid and rsodeoxycholic acid by hman intestinal flora, with particlar reference to the mtal conversion between the two bile acids. J. Lipid Res. : Shefer, S., G. Salen, S. Haser, B. Dayal, and. K. Batta Metabolism of isobile acids in the rat.j Bid. Chem. 7: Higashi, H., T. Setogshi, and T. Kazki Interconversion between chenodeoxycholic acid and rsodeoxycholic acid in anaerobic cltres of intestinal bacteria, and redction of 7ketolithocholic acid to both bile acids. cta Hepatol. Jpn Bazzoli, F., H. Fromm, R. P. Sarva, R. F. Sembrat, and S. Ceryak Comparative formation of lithocholic acid from chenodeoxycholic and rsodeoxycholic acids in the colon. Gastroenterology. 83: Macdonald I.., and P. D. Roach Bile salt indction of 7a and 7&hydroxysteroid dehydrogenases in Clostridim absonm. Biochim. Bwphys. cta Salen, G., G. S. Tint, B. liav, N. Deering, and. H. Mosbach Increased formation of rsodeoxycholic acid in patients treated with chenodeoxycholic acid. J. Clin. Invest. 3: Macdonald, I.., D. M. Htchison, and T. P. Forrest Formation of rso and rsodeoxycholic acids from primary bile acids by Clostridim absonm. J. Lipid Res. : Downloaded from by gest, on May 14, 18 Fromm, Sara, and Baoli Formation of rsodeoxycholic acid in the colon 83