Differential distribution of molecular forms of cholecystokinin in human and porcine small intestinal mucosa

Regulatory Peptides, 8 (1984) 9-19

9

Elsevier RPT 00254

Differential distribution of molecular forms of cholecystokinin in human and porcine small intestinal mucosa P.N. Maton, A.C. Selden and V.S. Chadwick Gastroenterology Unit, Department of Medicine, Royal PostgraduateMedical School, Du Cane Road, London WI2 0HS, U.K.

(Received 6 July 1983; revised manuscript received 17 October 1983; accepted for publication 18 October 1983)

Summaw To examine the distribution of cholecystokinins (CCKs) along the small intestine we examined the nature of CCKs in samples of jejunum,,mid-intestine and ileum from human and porcine intestine. CCKs in intestinal mucosa were extracted by boiling in both neutral and acid conditions, and subjected to high pressure liquid chromatography (HPLC) to separate the forms of C C K followed by radioimmunoassay of separate fractions. In neutral extracts of human intestine C C K immunoreactivity totalled 119.4, 22.9 and < 1 n g / g in jejunum, mid-intestine and ileum, whilst in acid extracts the corresponding values were 65.3, 47.4 and < 1 n g / g . Amounts of C C K extracted from porcine mucosa were of similar magnitude. In neutral extracts material co-chromatographing on H P L C with synthetic porcine C C K 8 predominated, whilst in acid extracts material co-chromatographing with CCKs 33/39 was the major form. These forms of human and porcine CCKs extracted from the mucosa behaved similarly to C C K 8 and C C K 33/39 standards on HPLC, in the radioimmunoassay and on molecular exclusion chromatography - suggesting marked similarity of the C C K s in the two species. In both species there was a marked change in the ratios of C C K 8 : C C K 3 3 / 3 9 down the intestine from 1 : 0.8 in human jejunum to 1 : 5.6 in mid-intestine and from 1 : 1.5 in porcine jejunum to 1 : 5.8 in mid-intestine. These observations may explain the changing patterns of CCKs in circulation with time after ingestion of a fat meal and the greater impairment of C C K 8 than C C K 33/39 release observed in coeliac disease.

Address for correspondence: Dr. P.N. Maton, Digestive Diseases Branch, Room 9C-103, Building 10,

National Institutes of Health, Bethesda, Maryland 20205, U.S.A. Telephone: 010-1-301-496-4201. 0167-0115/84/$03.00 © 1984 Elsevier Science Publishers B.V.

10 intestine; cholecystokinin; molecular forms; radioimmunoassay; high performance liquid chromatography

Introduction

Cholecystokinin (CCK) was first isolated from porcine intestinal mucosa as a 33 amino acid residue peptide in the 1960's [1]. Since then immunochemical evidence has accumulated suggesting the existence of other forms with 4, 8, 12, 39 and 58 [2-5] residues, and some of these molecules have been extracted from tissues in pure form. All these forms of CCK have a common carboxyl terminal sequence, and all those molecules of 7 or more amino acid residues include a sulphated tyrosine residue essential for full cholecystokinin-like biological activity [6]. However, the peptides differ in their spectrum and potency of biological activities [7]. In the 1920's Ivy demonstrated that CCK bioactivity was situated in the upper small intestine [8] and immunohistochemical techniques have confirmed that most CCK-containing cells are in that region [9,10]. Several studies have examined the forms of CCKs present in the intestinal mucosa in various species including man. In contrast to early reports it is now clear that a large proportion of CCK present in the upper small intestine is CCK 8 with CCK 33 being the second important form [2-4]. In none of these studies, however, was the distribution of the molecular forms along the length of the intestine examined. Our previous studies of release of CCK into the circulation after a fat meal suggested that both CCK 8 and CCK 33/39 were released [11]. Ho~vever, peak concentrations of CCK 8 occurred earlifr than peak concentrations of CCK 33/39, raising the ffossibility that these two forms could be differentially distributed down the intestine. Accordingly, we wished tO study the distribution of the molecular forms of CCK along the gut. Furthermore since all available ~ C K standards are porcine forms and all antisera are raised against these molecules we wished to compare the physico-chemical properties of human and porcine CCKs.

Methods

Human and porcine intestinal samples Intestinal tissue was obtained from three levels of a single human intestine obtained within 40 min of death in a patient who died of cirrhosis. Tissue was taken from the jejunum 2.5 cm distal to the ligament of Treitz, the mid-point of the small intestine, and 15 cm proximal to the ileocaecal valve. The samples of intestine were opened, washed with ice-cold 0.155 mol. 1-1 sodium chloride, and stored at - 2 0 ° C prior to further processing. Samples of porcine intestine were obtained from 2 animals within 5 min of death. Samples were obtained from sites corresponding to the sites used in the human study, and treated in the same manner.

Extraction and chromatography of tissue samples Samples (ca. 100 mg) of human or porcine intestine were scraped to separate the mucosa from the muscle layer. The mucosal fraction was quickly weighed and boiled in neutral or acid conditions. Neutral solutions preferentially extract small forms of CCK, whilst acid conditions preferentially extract the larger forms [12]. In neutral extractions the mucosa was boiled in 1.0 ml of distilled water for 1 min and the tissue homogenised in a Dounce homogeniser. The suspension was briefly reboiled, and then centrifuged at 3000 x g for 10 min and the supernatant removed. In the acid extraction procedure, tissue was boiled as for the neutral extraction but then glacial acetic acid was added in sufficient amount to render the solution 0.5 mol. 1-1 with respect to acetic acid (pH 2.5). This mixture was homogenised, reboiled, transferred to a 4°C refrigerator for 40 min to cool and then after centrifugation the supernatant was removed. Supernatants from the tissue extracts were then applied to a pre-equilibrated Seppak (Waters Inc., U.K.). The Seppak was prewashed with 10 ml each of 0.155 mol- 1-1 sodium chloride pH 2.1 (Solution A) and acetonitrile : water 3 : 2 (Solution B) in a ratio 1:1, followed by Solution B, pure acetonitrile, Solution B, Solution A : B , 1:1 and finally 0.155 mol-1-1 saline. Tissue extracts were applied to the Seppak, washed through with 10 ml of Solution A, and eluted with 10 ml of acetonitrile. This eluate was reduced in volume by 30% by blowing down under nitrogen, 3 ml of Solution A added and the resulting solution applied to a gradient elution high performance liquid chromatograph (HPLC) as previously described [11]. This system separates CCKs from each other (though CCK 33 and 39 are only incompletely separated) and from gastrins. Recovery of CCK 8 and CCK 33 added to tissues prior to boiling and processed through the Seppak and HPLC was not significantly different from 100%.

Radioimmunoassays Column fractions from the HPLC were analysed using a radioimmunoassay directed against the common carboxyl terminal portion of porcine CCKs. Nonsulphated [125I]CCK 8 was used as tracer and pure synthetic porcine CCK 8 (Squibb Inc., London) was used as standard. This assay has previously been described in detail [11]. Molar cross-reactivity of various peptides CCK 8 : 33 : 39 : 4 : Gastrin 17 : Gastrin 34 were 1 : 0.64 : 0.46 : 0.005 : 0.5 : 0.5. There was no significant cross-reactivity with motilin, vasoactive intestinal polypeptide, glucose-dependent insulinotropic peptide, somatostatin, or glucagon. The within- and between-assay coefficients of variations were 10.1% and 14.2% respectively. Column fractions were also analysed for gastrin immunoreactivity using a relatively specific gastrin radioimmunoassay with 14% cross-reactivity with CCKs [11].

Characterisation of human CCKs Much CCK immunoreactivity extracted from human intestinal mucosa co-eluted with porcine CCK 8 and CCK 33 standards on the HPLC. To check on the identity

12 of these putative human CCKs two other studies were performed. Human material co-chromatographing with porcine CCK 8 was assayed in the radioimmunoassay using differing dilutions of the fractions to assess the degree of parallelism with the standard curve. Similar studies were performed with the human material which co-chromatographed on HPLC with CCK 33. The molecular sizes of the major fractions from the HPLC were also measured. The HPLC fraction containing human material corresponding to porcine CCK 8 was applied to a 1 × 50 cm Sephadex G50SF column equilibrated with ammonium bicarbonate 0.1 mol. 1-1 pH 7.8 and previously calibrated with porcine standards, dextran blue and [125I]sodium iodide. Column fractions were analysed for CCK immunoreactivity by radioimmunoassay. The HPLC fraction containing human material co-chromatographing with porcine CCK 33 was also subjected to Sephadex G50SF chromatography but on this occasion equilibrated in 0.5 mol • 1- i acetic acid. Again, column fractions were analysed by radioimmunoassay.

Results H u m a n and porcine intestinal mucosa Neutral extracts

A complete HPLC chromatogram of a neutral extract of human jejunal mucosa is shown in Fig. 1. CCK-like immunoreactivity was only detected in a limited part of the chromatogram, much of it co-chromatographing with synthetic porcine CCK 8. No material co-chromatographing with CCK 4 was detected though' the cross-reactivity of this molecule in our radioimmttnoassay is very low, but a small amount of material Co-dhromatographing with gastdns was detected. Analysis of these same fractions with the relatively specific gastrin antiserum confirmed that material co-chromatographing with porcine CCKs was CCK-like rather than gastrin-like. The CCK imt~unoreactivity obtained from neutral extracts of a single human, and a single porcine intestine is shown in Figs. 2 and 3. In both species a similar pattern of results was observed. There was a gradient of total CCK-like immunoreactivity down the intestine with the highest concentrations being present in jejunal extracts. In human extracts CCK immunoreactivity totalled 119.4 n g / g in jejunum, 22.9 n g / g in mid-intestine, and < 1 n g / g in ileum. Corresponding values in 2 porcine intestines were 51.7 and 77.6 n g / g in jejunum, 9.2 and 9.8 n g / g in mid-intestine and < 1 n g / g in ileum. In both species most CCK immunoreactivity in neutral extracts co-chromatographed with the synthetic porcine CCK 8 standard used to calibrate the column. A c i d extracts

The results of the radioimmunoassay of HPLC fractions from acid extracts of human and porcine mucosa are shown in Figs. 4 and 5. As with the neutral extracts in both species, there was a gradient of total CCK immunoreactivity down the intestine. In human intestine total immunoreactivity fell from 65.3 ng/g in the

13

CCK 4

8 33

50-

40-

oo

,~

G 34 I I

30-

o c

E

20 -

.s c L-

,o

10oo ,.1. ~-~

)

0 --

I

I

I

I

I

I

I

I

0

lO

20

30

40

50

60

lO

HPLC Fraction number Fig. 1. High performance liquid chromatograph of a neutral extract of h u m a n jejunal mucosa, with column fractions analysed using an antibody to c o m m o n carboxyl terminal of C C K s and gastrins ( ) and a relatively gastrin-specific antibody ( . . . . . ). The arrows indicate the retention times of s y n t h e t i c porcine C C K 4, and C C K 8, 99% pure C C K 33 and gastrin 17-I and 34.

8

34 G ll

CCK 33139

+

60

+

+

40

>,

20

._>

c

20

E_

~

jejunum

mid

0 10

0]

ileum

;5' ;~'

;~

i

I

41

I

I

4~

I

4~ ' ~

HPLC Fraction number

Fig. 2. CCK-like immunoreactivity ( n g / g ) in a restricted portion of the HPLC, of neutral extracts of h u m a n intestinal mucosa taken from jejunum, mid-small intestine and ileum. The arrows have the same significance as in Fig. 1.

14 CCK 33139

8 --

___>"

40-

20-

Z

jejunum

o-

E

E

20 1

mid

10 0]

ileum

34, G ]7

L.;

3

7

9

41

45

43

47

HPLC Fractionnumber Fig. 3. CCK-like immunoreactivity in fractions from the H P L C of neutral extracts of porcine jejunal, mid-intestinal and ileal mucosa.

jejunum to 47.4 ng/g in mid-intestine to < 1 ng/g in ileum. In the 2 porcine intestines jejunal CCK immunoreactivity totalled 64.9 and 91.8 ng/g, in mid-intestine 33.4 and 34.0 ng/g and in ileum < 1 ng/g. In contrast to neutral extracts, only a small amount of material co-chromatographed with porcine CCK 8, but a greater proportion of the immunoreactivity co-chromatographed with pure CCK 33/39. Immunoreactivity was present in more fractions following aci~t extraction than neutral extraction probably indicating the presence of more molecular forms.

8

_-° ,0]

CCK

33/39

34 G 17

jejunum

0

'G .~

40 1 20

.E

0

mid

ileum

0 I

I

35

I

I

37

I

I

39

I

I

41

HPLC Fraction

I

I

43

I

I

45

I

I

47

number

Fig. 4. CCK-like immunoreactivity in fractions from the H P L C of acid extracts of human intestinal mucosa.

15 8

_~ c

-

20 1

CCK 33•39

34

G ll

jejunum

0 40

~o

20

mid

c E

E

0

0

]

ileum !

t

35

,

I

37

[

I

39

I

I

41

I

I

43

I

I

"1 41

I

45

HPLC Fraction numbar Fig. 5. CCK-like immunoreactivity in fractions from the HPLC of acid extracts of porcine mucosa.

intestinal

Characterisation of CCKs extracted from human intestinal mucosa

CCK immunoreactive material in extracts of human jejunum was further characterised by examining its behaviour in the radioimmunoas~ay. When dilutions of samples co-chromatographing with porcine CCK 8 or CCK 33/39 standard were assayed in the radioimmunoassay the dilution curves were parallel to porcine standards (Fig. 6) indicating homology at the carboxyl terminal end of the CCKs with respect to antibody binding.

80

60_ B/Bo~ 40-

20--

O-

I

i

i

10

100

1000

/

10000

CCK81 tube (fg) Sample vol (p.IxlO) Fig. 6. Standard curve representing inhibition of binding of [125I]CCK 8 at graded amounts of standard porcine CCK 8 (@), and with graded amounts of putative human CCK 8 (11) and CCK 33/39 ( 0 ) in fractions from the HPLC. synthetic

16 purcine

.o

CCK

33/39

, ~

acetic acid

100

i>_ -

~

porcine

CCK

8

20o

c

E

ammonium bicarbonate

100

1

I

I

I

1

0

20

40

60

80

Percent

elution

blue

dextran

I

I

to

100 125

120 I

Fig. 7. CCK-like immunoreactivity in fractions from Sephadex G50 superfine columns. The top panel indicates the elution of immunoreactivity when putative h u m a n C C K 33 from the HPLC was applied to a column equilibrated in 0.5 mol. 1- i acetic acid. The lower panel indicates the elution of immunoreactivity when putative h u m a n C C K 8 was applied to a column equilibrated in a m m o n i u m bicarbonate. The elution positions of porcine C C K 8 and porcine C C K 33/39 are shown "for comparison.

The molecular dimensions of the putative human CCK 8 and CCK 33/39 were assessed by Sephadex chromatography. The putative CCK 8 co-chromatographed with synthetic porcine CCK 8 and putative CCK 33/39 co-chromatogralbhed with 99% pure CCK 33/39 (Fig. 7) indicating similar molecular size. Thus the major forms of human intestinal CCKs behaved like porcine CCK 8 and CCK 33/39 with respect to hydrophobicity (HPLC), molecular size (Sephadex chromatography) and immunoreactivity (radioimmunoassay). Distribution of molecular forms of CCK in human and porcine intestine

By combining the results of acid and neutral extracts, and correcting for the molar cross-reactivity of CCK 33/39 in the radioimmunoassay (0.5) the amounts of TABLE I Distribution of C C K 8 and C C K 33/309 along h u m a n and porcine small intestinal mucosa

Jejunum Human Porcine Mid-intestine Human Porcine

CCK 8 (pmol/g)

C C K 33/39 (pmol/g)

Ratio (CCK 8 : 33/39)

60.5 54.6 48.2

47.1 56.0 46.1

1 : 0.8 1 : 1.0 1 : 1.0

12.4 8.0

68.9 44.1

1.7

5.3

1 : 5.6 1 : 8.8 1 : 3.1

17 CCK 8 and CCK 3 3 / 3 9 extracted from each sample could be determined and the ratio of the molecular forms at each level of the intestine calculated. As shown in Table I, although the total amounts of CCKs extracted varied, all studies demonstrated that CCK 8 and CCK 33/39 were of similar concentration in the jejunum, but that in the more distal intestine the larger forms predominated.

Discussion These data demonstrate the similarity of human and porcine intestinal mucosal CCKs with respect to hydrophobicity, molecular size and immunoreactivity. In addition they confirm the previous findings of a gradient of total mucosal CCK down the intestine. Interestingly they show for the first time that there is a change in ratio of CCK 8 to CCK 33/39 down the gut with the small forms predominating proximally and the larger forms predominating distally. Previous methods of measuring intestinal CCKs have depended on separation of the various forms of CCK by molecular exclusion chromatography. In this study we used HPLC which provides superior recovery of peptides. It may be that these differences in methodology account for differences in ratios of CCKs observed in the upper small intestine in the various studies. In duodenum Rehfeld [2] found the molar ratio of CCK 8 : 33/39 to be c 1 : 0.5 and Calam et al. [12] found a ratio of c 1 : 0.35, whereas we found a ratio of c 1 : 1. Our aim in this study was to define the distribution in the gut of the forms of CCK that we were able to identify in the circulation after a fat meal [11] - CCK 8 and CCK 33/39. We did not identify CCK 4 but our antibody cross-reacts very weakly with this molecular form. Similarly we did not identify any major amount of immunoreactivity which corresponded to a large molecular weight form [12,25] but our studies do not exclude the possibility that such a form was present in the samples in small amount. Clearly, human CCKs are very similar in structure to porcine CCKs and may be identical. The behaviour in the radioimmunoassay suggests conservation of the biologically active carboxyl-terminal of human and porcine CCKs whilst the results of Sephadex chromatography suggest that the CCKs of the two species are of similar size. The HPLC demonstrates that the ratios of hydrophobic and hydrophilic amino acid residues are similar [13]. Nevertheless, there is indirect evidence for species differences between human and porcine CCK 33/39 at the amino terminal portion of the molecule [14]. The reason for the differential distribution of CCKs is not clear but a similar phenomenon has been noted in the case of gastrin, where gastrin 17 predominates in the antrum and gastrin 34 predominates in the duodenum [15]. These changes presumably represent different degrees of post-translational modification of a large precursor form. Such differences in different organs are well recognised - nearly all brain CCK consists of small forms [16]. The distribution of CCKs in intestinal mucosa may provide an explanation for our findings of post-prandial elevation in plasma CCK concentrations, whereby

18

peak CCK 8 concentrations occur earlier than peak CCK 33/39 concentrations [11]. Furthermore such a distribution is consistent with our findings in patients with coeliac disease in which mucosal damage is most severe proximally and release of CCK 8 is impaired more than release of CCK 33/39 [17].

Acknowledgements We are grateful to Dr. V. Mutt for generous gifts of 99% pure CCK 33 and 39, to Squibb Inc. for gifts of sulphated and non-sulphated CCK 8 and to Dr. V.L.W. Go for anti CCK antibody. Mrs. Jean De Luca typed the manuscript. This work was supported by the Medical Research Council. P.N.M. was the Srrfith, Kline & French British Society of Gastroenterology Research Fellow.

References 1 Jorpes, J.E. and Mutt, V., In Jorpes, J.E. and Mutt, V. (Eds.) Secretin Cholecystolduin - Pancreozymin and Gastrin. Handbook of Experimental Pharmacology, volume 34. Springer-Verlag, Berlin, 1973, 1-179. 2 Rehfeld, J., Immunochemical studies on cholecystokinin 11 distribution and molecular heterogeneity in the central nervous system and small intestine of man and hog. J. Biol. Chem., 253 (1978) 4022-4027. 3 Dockray, G.J., Immunoreactive component resembling cholecystokinin octapeptide in intestine. Nature, 270 (1977) 359-361. 4 Ryder, S., Eng, J., Straus, E. and Yalow, R.S., Alkaline extraction and characterisation of cholecystokinin - immunoreactivity from rat gut. GastroenterologY, 81 (1981) 267-275. 5 Eysselein, V.E., Reeve, J.R. Jr., Shively, J.E., Hawke, D. and Walsh, J.H., Partial structure of a large canine cholecystokinin (CCK58) amino acid sequence. Peptides, 3 (1982) 687-691. 6 Johnson, L.R., Stening, G.R. and Grossman, M.I., Effect of sulfation on the gastrointestinal actions of caerulein. Gastroenterology, 58 (1970) 208-216. 70ndetti, M.A., Rubin, B., Engel, S.L., Pluscec, J. and Sheehan, J.T., Cholecystokinin-pancreozymin: recent developments. Am. J. Dig. Dis., 15 (1970) 149-156. 8 Ivy, A.C., Kloster, G. and Lueth, H.G., On the preparation of "cholecystokinin". Am. J. Physiol., 91 (1929) 336-344. 9 Polak, J.M., Pearse, A.G.E., Bloom, S.R., Identification of cholecystokinin-secreting cells. Lancet, 2 (1975) 1016-1018. 10 Buffa, R., Solcia, E. and Go, V.L.W., Immunohistochemical identification of the cholecystokinin cell in the intestinal mucosa. Gastroenterology, (1976) 528-532. 11 Maton, PIN., Selden, A.C. and Chadwick, V.S., Large and small forms of cholecystokinin in human plasma: measurement using high pressure liquid chromatography and radioimmunoassay. Regul. Peptides, 4 (1982) 251-260. 12 Calam, J., Ellis, A. and Dockray, G., Identification and measurement of molecular variants of cholecystokinin in duodenal mucosa and plasma: diminished concentrations in patients with coeliac disease. J. Clin. Invest., 69 (1982) 218-225. 13 Molnar, I. and Horvath, C., Separation of amino-acids and peptides on non-polar stationary phases by hlgh-performance liquid chromatography, J. Chromatogr., 140 (1977) 623-640. 14 Go, V.L.W., Ryan, R.J. and Summerskill, W.H.J., Radioimmunoassay of porcine cholecystokinin-pancreozymin. J. Lab. Clin. Med. (1971) 684-689.

19 15 Malstrom, J., Stadil, F. and Rehfeld, J.F., Gastrins in tissue. Concentration and component pattern in gastric duodenal and jejunum mucosa of normal human subjects and patients with duodenal ulcer. Gastroenterology, 70 (1976) 697-703. 16 Dockray, G.J., Immunochemical evidence of cholecystokinin-like peptides in brain. Nature (London), 264 (1976) 568-570. 17 Maton, P.N., Selden, A.C., Fitzpatrick M.L. and Chadwick, V.S., A reversible defect of gallbladder emptying and plasma cholecystokinin release in coeliac disease. Gut, 24 (1983) A494.