intestinal alkaline phosphatase in bile: evidence for an ... · gut, 1981, 22, 493-498 intestinal...

Gut, 1981, 22, 493-498

Intestinal alkaline phosphatase in bile: evidence for anenterohepatic circulationT W WARNES,* P HINE, G KAY, AND A SMITH

From the Liver Section, University Department of Gastroenterology,Manchester Royal Infirmary, Manchester

SUMMARY The isoenzyme characteristics of alkaline phosphatase in human bile obtained frompatients with gallstones have been determined by means of electrophoresis on polyacrylamide gel,both before and after preincubation with neuraminidase, and also by means of inhibition tests withheat, urea, and L-phenylalanine. Bile alkaline phosphatase is shown to be partly secreted by theliver cell, but partly derived from the small intestine. The presence of small-intestinal alkalinephosphatase in bile implies an enterohepatic circulation of this high molecular weight glycoprotein.The significance of this finding is discussed in relation to the mechanism whereby large proteinmolecules may pass across the plasma membrane of the liver cell.

There are two main theories concerning theorigin of alkaline phosphatase in bile. The firstis that serum alkaline phosphatase, primarilyderived from bone, is excreted into bile; theother postulates that the alkaline phosphatase inbile is formed by the liver cells.' Most of theexperimental evidence suggests that the biliarytree is not an important excretory route forserum alkaline phosphatase and supports the con-cept of 'over-production' by the liver as themost important cause of the increased serumalkaline phosphatase in biliary obstruction.2 3The most convincing evidence has been providedby the many experiments in which infused homo-logous alkaline phosphatase has failed to accumu-late in the liver or be excreted in the bile in theexperimental animal.4- Furthermore, infusedplacental alkaline phosphatase is not excretedinto bile in the human.8 However, in the experi-mental animal, there is some evidence that theintestinal enzyme, unlike alkaline phosphatasederived frcm other organs, may be secretedinto bile.9 10 Serum alkaline phosphatase existsin the form of isoenzymes which can be differen-tiated by immunological means, by inhibitiontests with heat, urea and L7phenylalanine, andby electrophoresis.

*Address for correspondence: Dr T W Warnes, Liver SectionUniversity Department of Gastroenterology, Manchester RoyalInfirmary, Oxford Road, Manchester M13 9WL, England.

Received for publication 13 January 1981

A number of immunological studies of alka-line phosphatase isoenzymes have been made,but immunological cross-reactivity of the iso-enzymes has created problems. Boyerll showedcross-reactivity between intestinal and placentalalkaline phosphatase and also between both ofthese and a minor kidney component. More re-cently it has been shown that intestinal alkalinephosphatase consists of two isoenzymes, one ofwhich shows cross-reactivity with placental alka-line phosphatase and the other with liveralkaline phosphatase.12 Mulivor et al.13 haveshown that at least three genetic loci are in-volved in determining the various forms ofhuman alkaline phosphatase, one for the placen-tal form, at least one for the intestinal form,and at least one or more for the liver, bone,and kidney isoenzymes. Lehmann14 has shown,however, that it is possible to prepare a mono-specific antiserum against intestinal alkalinephosphatase by absorbing out other componentsand, therefore, immunological methods mayhave an important role to play in the identi-fication of intestinal alkaline phosphatase.The amount of small-intestinal alkaline phos-

phatase in serum is known to be geneticallydetermined, individuals who are blood group 0of B having higher levels than those of bloodgroup A.' In addition, there is an effect ofsecretor status. Thus, within any given bloodgroup, individuals who possess the glycoproteinsecretor substances in saliva show higher levels

493

on 24 March 2019 by guest. P

rotected by copyright.http://gut.bm

j.com/

Gut: first published as 10.1136/gut.22.6.493 on 1 June 1981. D

ownloaded from

http://gut.bmj.com/

494

of intestinal alkaline phosphatase in serum.On electrophoresis of serum, liver, bone, and

intestinal bands may be found and in serumfrom 78% of patients with liver disease an originband may be observed. This may represent acomplex of the liver isoenzyme with lipopro-tein.'5 The intestinal band can be easily dis-tinguished on electrophoresis, as it is inhibitedby L-phenylalaninel' and its electrophoreticmobility is unaltered after treatment with neur-aminidase."7

Electrophoretic separation of the bile alkalinepihosphatase isoenzymes has largely made use ofthe older techniques of paper or starch gel elec-trophores.is and conflicting results have beenobtained regarding the number and origin of thebands so obtained.17-2' Keiding,22 after electro-phoresis of human bile on starch gel, noted thepresence of a zone of alkaline phosphataseactivity which appeared to be identical with thatfound in intestinal fluid and lymph. Rosenberg19had previously made a similar observation,although he did not specifically suggest that itrepresented intestinal alkaline phosphatase. Thepresent study of the alkaline phosphatase iso-enzymes in human bile uses polyacrylamide gelelectrophoresis (PAGE), which is known to givea superior resolution of alkaline phosphataseisoenzymes,' to confirm and extend these earlierobservations.

Methods

Twen.ty-four samples of bile were obtained from17 patients either from the gall-bladder (10),hepatic duct (five), or common bile duct (one)at the time of operation, or from the commonbile duct by means of a T-tube drain after opera-tion (eight). One patient with carcinoma of thepancreas and one with chronic relapsing pan-creatitis were found to have no gallstones atoperation, while one patient with acute relapsingpancreatitis and one with chronic relapsing pan-creatitis were found to have gallstones. Theremaining 20 samples of bile were obtained atthe time of routine cholecystectomy in patientswith gall-stones.

Alikaline phosphatase was measured by a modi-fication of the King-Armstrong method23 andthe isoenzyme characteristics determined by in-activation with heat (56°C for 15 minutes) andIM urea, and by inhibition with 0.005ML-phenylalanine.24 Electrophoretic separation ofthe bile isoenzymes was performed by PAGEand the enzyme bands located by staining withbeta-naphthyl phosphate and Fast Blue BB.15 In

Warnes, Hine, Kay, and Smith

some experiments, L-phenylalanine was incor-porated into the staining mixture. Samples werealso treated with neuraminidase (type IV Cl.perfringens, Sigma Chemical Co.) before electro-phoresis, using a modification of the method ofRobinson and Pierce.'7 The enzyme was ex-tracted from human liver obtained at necropsy,and from small-intestinal mucosa obtained atoperation, by a modification of the butanolmethod of Morton.25The results were analysed statistically by

Schiffe's method of adjusted significance levels.Blood-groups in serum were determined bystandard serological techniques.

Results

Bile alkaline phosphatase was significantly dif-ferent from small-intestinal alkaline phosphatasein its sensitivity to heat (P<0-001), urea(P<0O001), and L-phenylalanine (p<0001) andalso from the liver isoenzyme after heat(P<0001), urea (P<0-001), and L-phenylalanineinhibition (P<0.01) (Table 1). Gall-bladder bilediffered from T-tube bile on heat inactivation(P<0001), urea inactivation (P<0-005), and L-phenylalanine inhibition (p<0-005) and also fromhepatic duct bile after heat inactivation(p<0005) and L-phenylalanine inhibition(P<0.02), though not after urea inactivation.Small-intestinal alkaline phosphatase was signi-ficantly different from liver alkaline phosphataseafter heat inactivation (P<0001), urea inactiva-tion (P<0001), and L-phenylalanine inhfibition(p<O0001).

After PAGE the main alkaline phosphatasebands found in the extracts of liver and smallintestine corresponded to the positions of theliver and small-intestinal bands found in serum.In 12 of the 19 samples of bile that were electro-phoresed a fast-running liver-type band (L) was

Table 1 Isoenzyme characteristics of bile, liver,and small-intestinal alkaline phosphatase

Heat Urea L-phenylalanine% remaining % remaining % inhibition

Hepatic duct (5) 24-5+ 8-6 29-8+17.3 9-6+ 5-4Gall bladder (10) 44.6±19-8 42.6+15-0 18.2+16.5T-tube (8) 26-1+ 8-3 26.6±14-9 9-2±10-2Common bile duct (1) 25-9 11-4 18-6All bile (24) 34-0±16-9 33.2+16.7 13.3±13.0Liver extract (3) 13-2±10-8 8-2± 6-7 8-3+ 1-4Small-intestinal

extract (5) 78.3±12-9 44.2± 7-6 78.9± 6-6

Samples were heated at 56°C for 15 minutes, inactivated with IM ureaand inhibited with 0.005M L-phenylalanine. Results are given ±S.D.



j.com/


ownloaded from

http://gut.bmj.com/

intestinal alkaline phosphatase in bile 495

Table 2 Electrophoretic bands observed in bile, liverand small-intestinal alkaline phosphatase

Liver Intestinal Origin Otherband band band

Hepatic duct (4) 2 2 4Gall bladder (9) 8 7 6 -

T-tube (5) 2 3 5 2Common bile duct (1) - 1 1 -

All bile (19) 12 13 16 2Liver extract (3) 3 - - -

Small-intestinalextract (5) 2* 5 2 1

*The liver band in small-intestinal extract was very faint.

Fig. 1 Polyacrylamide gel disc electrophoresisshowing the liver band (L), origin band (0), andintestinal band (I). A. Serum from a patient with liverdisease. B. Gall-bladder bile. C. Small-intestinalextract. D. Liver extract.

Fig. 3 The effect ofphenylalanine on theALP isoenzymes ofbile. A. T-tube bile-normal stain.B. T-tube bile-stainincorporatingL-phenylalanine.

Fig. 2 The effect of neuraminidase onelectrophoretic mobility. A. Liver extract.B. Small-intestinal extract. C. Hepatic-duct bile. The right-hand gel of each pair,superscript N, has been preincubatedwith neuraminidase.



j.com/


ownloaded from

http://gut.bmj.com/

496

seen, and 16 had an origin band (0) identicalwith that previously described15 in the serum ofpatients with liver disease (Fig. 1). In 13 samplesof bile a slower band (I) was found which wasidentical in position with both the main bandof small-intestinal extract and with the small-intestinal band in serum. When samples of biledemonstrating band I were incubated with neur-aminidase before electrophoresis the electro-phoretic mobility of this slower band, like thatof small-intestinal extract, was unaltered, whilethe main band L of bile was slowed in a similarmanner to that of liver extract alkaline phos-phatase (Fig. 2). When L-phenylalanine was in-cor,porated into the staining mixture the slowerband I decreased in intensity, while both the liver-type band (L) and the origin band (0) were un-altered (Fig. 3). The small-intestinal band wasseen in the majority of samples of bile, whetherthe sample was obtained from gall-bladder, hepa-tic duct, common bile duct, or by T-tubedrainage (Table 2). In two samples of bile anadditional minor band was found running aheadof the liver band.Blood group data were available for 14

patients; bile samples from 11 of these wereeJectrophoresed and the incidence of intestinalbands in these samples from patients of knownblood group is shown in Table 3.

Table 3 The incidence of intestinal bands in bilefrom patients of known blood group

Blood group No. Intestinal band

Present Absent

0 2 1 1A 6 4 2B 2 2 0AB 1 1 0

There was no significant difference in incidence of intestinal bandsbetween individuals in blood group A and those in blood groups0 and B combined (p= 1, Fischer's exact test).

Discussion

Substances secreted by the liver, absorbed by theintestine, and resecreted by the liver have anenterohepatic circulation, which may be portalas with bile salts or extraportal as in the case oflipids. Apart from phospholipids and choles-terol, a number of other substances such as bilepigments, folate, and certain steroid hormonesand antibiotics undergo this latter form of entero-hepatic circulation,'6 but there is no concreteevidence that high molecular weight substances

Warnes, Hine, Kay, and Smith

such as proteins undergo an enterohepaticcirculation.The isoenzyme characteristics of bile alkaline

phosphatase as determined in the present studyare similar to those previously reported byHorne et al.'7 As these lie intermediate betweenthose of liver and small intestinal alkaline phos-phatases, they are compatible with the thesisthat bile alkaline phosphatase represents a mix-ture of the other two isoenzymes. The inhibitiondata by themselves are, however, not conclusive.We have shown that on electrophoresis two

main isoenzymes of alkaline phosphatase arefound in bile. One is identical with that of liverextract and with the liver main band in serum,and this finding is in agreement with the reportof previous workers.28 The other appears to besmall-intestinal alkaline phosphatase and, likethe latter, is unaltered in electrophoretic mobilityby neuraminidase. The decrease in intensity ofthis band when L-phenylalanine is incorporatedinto the staining mixture (Fig. 3) confirms itssmall-intestinal origin.The higher levels of intestinal alkaline phos-

phatase in the serum of individuals in bloodgroups 0 or B are thought to be mainly dueto an increased synthesis or release of the iso-enzyme into the serum. If this were the over-riding factor, then one would expect also to findhigher levels of intestinal alkaline phosphatasein the bile of these patients. However, if thehigher levels were due to a slower rate of elimi-nation of the isoenzyme then one might expectto find lower levels in the bile of individuals inblood groups 0 or B, and correspondingly higherlevels in the bile of individuals in blood groupA. In the present study we found no significantdifference in the incidence of intestinal bands inbile between patients of blood group A andthose of blood groups 0 and B. However, inserum the differences are known to be quantita-tive rather than qualitative, and this may wellalso be the case with bile. If this is so, then animmunological method for measuring intestinalalkaline phosphatase in bile may clarify thispoint, providing the problems of cross-reactivitycan be overcome.Three possible explanations for the presence

of small-intestinal alkaline phosphatase in bilehave to be considered. The first is that it is dueto direct reflux from duodenal juice into bile.However, the 'small-intestinal' isoenzyme hasbeen identified in samples taken as high as thehe.patic ducts and radiological observationstaken at the time of endoscopic retrogradecholangiography suggest that free reflux of con-



j.com/


ownloaded from

http://gut.bmj.com/

Intestinal alkaline phosphatase in bile 497

trast into the common bile duct does not occurunder the conditions obtaining-that is, in theabsence of a previous sphincterotomy or of con-comitant administration of CCK-pancreozymin.The second possibility is that the 'small-intes-tinal' alkaline phospatase found in bile is, infact, derived from the gall-bladder epithelium.This is extremely unlikely in view of the factthat the isoenzyme was i;dentified in three outof five specimens of T-tube bile obtained afterremoval of the gall-bladder; in these samplesthe small-intestinal band was still present oneweek after cholecystectomy. The third, and, inour opinion most likely, possibility is that theintestinal isoenzyme arises in bile after directpassage through the hepatocyte. Thus, alkalinephosphatase in the bile of these patients appearsto consist largely of an isoenzyme secreted bythe hepatocyte, and another derived from thesmall-intestinal brush borders which has tra-versed the liver cell and been excreted into bile.Small-intestinal alkaline thosphatase is releasedinto duodenal juice by the hormones secretinand CCK-pancreozymin.29 Histological studieshave shown30 that, after feeding of long chaintriglycerides, the isoenzyme passes from the brushborder across the small intestinal epithelium andinto the lymphatic system.' Thus, in demonstrat-ing the presence of small intestinal alkaline phos-phatase in bile, the present study implies thatalkaline phosphatase undergoes an enterohepa-tic circulation.

Recent studies have elucidated the mechanismwhereby glycoproteins are able to traverse theplasma membrane of the hepatocyte. Ashwelland Morell31 have shown that sialic acid, whichoccurs only in the terminal portion of glycopro-tein carbohydrate side chains, plays an impor-tant role in determining the circulating life ofserum glycoproteins. Thus, removal of the termi-nal sialic acid groups and exposure of the pen-ultimate galactosyl group of the carbohydrateside chain of a glycoprotein results in rapidclearance from the circulation of many serumglycoproteins which can then be recovered inthe liver.32 33 It has also been demonstratedthat isolated hepatic cell plasma membranes con-taiin specific saturable binding sites for thedesialylated glycoproteins. It is therefore rel-evant that small-intestinal alkaline phosphataseis unique among the alkaline phosphatase iso-enzymes in lacking terminal sialic acid residues,as shown electrophoretically by its resistance toneuraminidase. Our study is consistent with pre-vious work in both man and the experimentalanimal in that, apart from the liver isoenzyme

which is presumably secreted into bile by thehepatocyte, no sialic acid-containing isoenzymesof alkaline phosphatase are present in bile.

Until very recently, no definite evidence wasavailable that galactosyl rece,ptors specific forsmall-intestinal alkaline phosphatase existed onthe plasma membrane of the hepatocyte,although our findings suggested that they shouldexist and might play a part in the uptake ofsmall-intestinal alkaline phosphatase into theliver. However, work has now been publishedby Scholtens et al.35 which demonstrates con-clusively that, in the experimental animal, agalactosyl receptor for small-intestinal alkalinephosphatase exists on the hepatocyte plasma mem-brane, strongly supporting the theory that small-intestinal alkaline phosphatase does undergo anenterohepatic circulation.

We wish to thank Professor HT Howat for hishelp and encouragement and Mr HB Torrancefor providing the samples of bile. Material onwhich this paper is based was presented to theEuropean Association for the Study of the Liver.36

References

'Warnes TW. Alkaline phosphatase. Gut 1972; 13:926-37.

2Polin SG, Spellberg MA, Teitelman L, et al. Theorigin of elevation of serum alkaline phosphatase inhepatic disease. Gastroenterology 1962; 42:431-8.

3Sebasta DG, Bradshaw FJ, Prockop DJ. Source ofthe elevated serum alkaline phosphatase activity inbiliary obstruction: studies utilising isolated liverperfusion. Gastroenterology 1964; 47:166-70.4Freeman S, Chen YP. The effect of jaundiced bloodupon normal dogs, with special reference to theserum phosphatase. J Biol Chem 1938; 123:23946.5Cantarow A, Miller L. Nonexcretion of jaundice-serum alkaline phosphatase in bile of normal dogs.Am J Physiol 1948; 153:444-6.6Wang CC, Grossman MI. Nonexcretion of serumalkaline phosphatase by the liver and the pancreasof normal dogs. Am J Physiol 1949; 156:256-60.

7Jersky J. Alkaline phosphatase in auxiliary livertransplantation. Gut 1974; 15:636-43.8Clubb JS, Neale FC, Posen S. The behaviour ofinfused human placental alkaline phosphatase inhuman subjects. J Lab Clin Med 1965; 66:493-507.9Maddock, S, Schmidt G, Thannhauser SJ. The effectof the intravenous injection of alkaline intestinalphosphatase on dogs following hepatectomy andtotal abdominal evisceration with and without hepa-tectomy. (Abstr.) Fed Proc 1942; 1:181.

°0Le Veen HH, Talbot LJ, Restuccia M, et al. Metab-olism and excretion of alkaline phosphatase:relation to liver function and determination of



j.com/


ownloaded from

http://gut.bmj.com/

498 Warnes, Hine, Kay, and Smith

maximal secretory rates of liver. J Lab Clin Med1950; 36:192-205.

11Boyer SH. Human organ alkaline phosphatases:discrimination by several means including starch gelelectrophoresis of antienzyme-enzyme supernatantfluids. Ann NY Acad Sci 1963; 103:938-55.

12Abdullah J Sh, Gowland E. An investigation intothe immunochemical properties of the isoenzymesof human alkaline phosphatase. Ann Clin Biochem1977; 14:212-17.

13Mulivor RA, Plotkin LI, Harris H. Differentialinhibition of the products of the human alkalinephosphatase loci. Ann Hum Genet Lond 1978; 42:1-13.

14Lehmann FG. Immunological relationship betweenhuman placental and intestinal alkaline phospha-tase. Clin Chim Acta 1975; 65:257-69.

15Warnes TW, Hine P, Kay G. Polyacrylamide geldisc electrophoresis of alkaline phosphatase iso-enzymes in bone and liver disease. J Clin Pathol1976; 29:782-7.

¢0Fishman WH, Green S, Inglis NI. L-phenylalanine:an organ specific, stereo-specific inhibitor of humanintestinal alkaline phosphatase. Nature (Lond) 1963;198:685-6.

'7Robinson JC, Pierce JE. Differential action ofneuraminidase on human serum alkaline phospha-tase. Nature (Lond) 1964; 204:472-3.

'8Rosenberg IN. Zone electrophoretic studies ofserum alkaline phosphatase. J Clin Invest 1959;38:630-44.

'9Chiandussi L, Greene SF, Sherlock S. Serumalkaline phosphatase fractions in hepato-biliary andbone diseases. Clin Sci 1962; 22:425-34.

20Pope CE II, Cooperband SR. Protein characteristicsof serum and bile alkaline phosphatase. Gastro-enterology 1966; 50:631-6.

21Hill PG, Sammons HG. An interpretation of theelevation of serum alkaline phosphatase in disease.J Clin Pathol 1967; 20:654-9.

22Keiding NR, The alkaline phosphatase fractions ofhuman lymph. Clin Sci 1964; 26:291-7.

23Kind PRN, King EJ. Estimation of plasma phos-phatase by determination of hydrolysed phenol withamino-antipyrine. J Clin Pathol 1954; 7:322-6.

24Warnes TW, Timperley WR, Hine P, et al. Pan-creatic alkaline phosphatase and a tumour variant.Gut 1972; 13:513-9.

25Morton RK. The purification of alkaline phospha-tases of animal tissues. Biochem J 1954; 57:595-603.

26Dowling RH. The enterohepatic circulation. Gastro-enterology 1972; 62:122-40.

27Horne M, Comish CJ, Posen S. Use of urea in theidentification of human alkaline phosphatase. J LabClin Med 1968; 72:905-15.

28Price CP, Hill PG, Sammons HG. The nature ofthe alkaline phosphatases of bile. J Clin Pathol1972; 25:149-54.

29Warnes TW, Hine P, Kay G. The action of secretinand pancreozymin on small-intestinal alkaline phos-phatase. Gut 1974; 15:39-47.

30Bluestein M, Malagelada J, Linscheer W, FishmanWH. Enzymorphology of alkaline phosphatase ofrat intestine following fatty acid infusion. J Histo-chem Cytochem 1970; 18:679.

31Ashwell G, Morell AG. A cryptic determinant ofglycoprotein catabolism. In: Jamieson GA,Greenwalt TJ, eds. Glycoproteins of blood cells andplasma. Philadelphia: Lipincott, 1971: 173-89.

32Van den Hamer CJA, Morell AG, Scheinberg IH,et al. Physical and chemical studies on caerulo-plasmin. IX. The role of galactosyl residues in theclearance of caeruloplasmin from the circulation.J Biol Chem 1970; 245:4397-402.

33Morell AG, Gregoriadis G, Scheinberg IH et al.The role of sialic acid in determining the survivalof glycoproteins in the circulation. J Biol Chem1971; 246:1461-7.

34Van Lenten L, Ashwell G. The binding of desialy-lated glycoproteins by plasma membranes of ratliver. Development of a quantitative inhibitionassay. J Biol Chem 1972; 247:4633-40.

35Scholtens HB, Meijer DKF, Koudstaal J, et al.The involvement of a galactosyl receptor in thein vitro binding and the in vivo clearance of intes-tinal alkaline phosphatase by rat liver. Abstracts ofthe 14th Meeting of the European Association forthe Study of the Liver, Dusseldorf, September, 1979.

36Warnes TW, Hine P, Kay G. The enterohepaticcirculation of alkaline phosphatase (AP) in patientswith gallstones. (Abstr.) Digestion 1977; 16:333-4.



j.com/


ownloaded from

http://gut.bmj.com/

intestinal alkaline phosphatase in bile: evidence for an ... · gut, 1981, 22, 493-498 intestinal...

Documents