0021-972X/01/$03.00/0
The Journal of Clinical Endocrinology & Metabolism
Copyright © 2001 by The Endocrine Society
Vol. 86, No. 1
Printed in U.S.A.
The Predominant Cholecystokinin in Human Plasma and
Intestine Is Cholecystokinin-33*
J. F. REHFELD, G. SUN, T. CHRISTENSEN,
AND
J. G. HILLINGSØ
Department of Clinical Biochemistry, Rigshospitalet, University of Copenhagen, 2100
Copenhagen, Denmark
ABSTRACT
Cholecystokinin (CCK) occurs in multiple molecular forms; the
major ones are CCK-58, -33, -22, and -8. Their relative abundance in
human plasma and intestine, however, is debated. To settle the issue,
extracts of intestinal biopsies and plasma from 10 human subjects
have been examined by chromatography, enzyme cleavages, and measurements using a library of sequence-specific RIAs. Plasma samples
were drawn in the fasting state and at intervals after a meal. The
abundance of the larger forms varied with the 8 C-terminal assays in
the library, as 2 assays overestimated and 3 underestimated the
amounts present. One assay, however, measured carboxyamidated
and O-sulfated CCKs with equimolar potency before and after tryptic
cleavage. This assay showed that the predominant plasma form is
CCK-33, both in the fasting state (⬃51%) and postprandially (⬃57%),
whereas CCK-22 is the second most abundant (⬃34% and 30%, respectively). In contrast, CCK-58 is less abundant in human intestines
(⬃18%) and plasma (⬃11%). Its predominance in feline intestines,
however, was confirmed. Hence, the results show a significant species
variation and emphasize the necessity of highly specific and well
characterized assays in molecular studies of CCK. (J Clin Endocrinol
Metab 86: 251–258, 2001)
C
HOLECYSTOKININ (CCK) is an important gut hormone that regulates gallbladder contraction, pancreatic enzyme secretion, hormone secretion, and growth. Moreover, CCK influences intestinal motility and satiety (for
reviews, see Refs. 1 and 2). The hormonal effects are due to
CCK in plasma that originates almost entirely from I cells in
the small intestinal mucosa (3, 4). CCK peptides are also
expressed in large quantities in neurons (5, 6), but neuronal
CCK contributes negligibly to CCK in plasma.
CCK was first identified in extracts of porcine small intestine as a carboxyamidated and tyrosyl O-sulfated peptide
containing 33 amino acid residues (CCK-33) (7). Subsequent
studies have shown that pro-CCK is processed at several
mono- and dibasic sites to release bioactive forms of different
lengths. To date these include CCK-83, CCK-58, CCK-39,
CCK-33, CCK-22, and CCK-8, that are all carboxyamidated
and O-sulfated (5, 6, 8 –13) and, hence, ligands for the CCK-A
receptor. There is general agreement that the predominant
form of CCK in central and peripheral neurons is CCK-8 (for
reviews, see Refs. 14 and 15). In contrast, there have been
remarkably diverging opinions about the main form in
plasma and intestinal tissue. The suggestions have varied
from CCK-8 to CCK-58 (16 –23).
To understand the function of a hormone, it is essential to
know the molecular form in which it circulates. For heterogeneous hormones at least the predominant form in plasma
should be known. In this regard it is amazing that consensus
still remains to be reached about the molecular nature of CCK
in plasma. Much of the problem appears to be due to variations in technology and assay quality (for review, see Ref.
24). Two laboratories have repeatedly reported that CCK-58
is the major hormonal form of CCK in several mammals,
including man (22, 25, 26). Their claim has gained widespread acceptance even though it is at odds with observations in other laboratories (18, 20, 21, 27). To settle the discrepancy, we have now studied the molecular pattern of
CCK in biopsies of human small intestine and in plasma
sampled in fasting and fed normal human subjects using a
new assay that measures the bioactive forms of CCK with
equimolar potency. Moreover, this assay does not react with
the homologous hormone, gastrin (28).
CCK nomenclature
Human pro-CCK is a protein of 95 amino acid residues.
The earliest posttranslational modification of pro-CCK appears to be O-sulfation of Tyr77, Tyr91, and Tyr94 by sulfotransferases in the trans-Golgi network of the I cells. Sequence 83– 86 (Phe-Gly-Arg-Arg) constitutes the amidation
site, which requires processing by prohormone convertases,
carboxypeptidase E, and the amidation enzyme complex to
release bioactive carboxyamidated CCK peptides (Fig. 1).
The largest bioactive form is CCK-83, which corresponds to
the amidated sequence 1– 83 of pro-CCK (13). This sequence
is cleaved at five or more monobasic sites to release CCK-58,
-39, -33, -22, and -8, all of which have the same C-terminal
heptapeptide amide sequence (-Tyr(SO3⫺)-Met-Gly-TrpMet-Asp-PheNH2), that also is the minimal epitope necessary for receptor binding. Binding to CCK-A receptors requires that the tyrosyl residue of the heptapeptide amide is
O-sulfated, whereas CCK-B receptors do not discriminate
sulfated from nonsulfated CCK peptides or CCK from gastrin peptides (for review, see Ref. 29).
Received May 16, 2000. Revision received September 5, 2000. Accepted September 15, 2000.
Address all correspondence and requests for reprints to: Dr. J. F.
Rehfeld, Department Clinical Biochemistry (KB 3011), Rigshospitalet,
DK-2100 Copenhagen, Denmark. E-mail: [email protected].
* This work was supported by grants from the Danish Medical Research Council, the Danish Biotechnology Program for Peptide Research, the Danish Cancer Union, the Gangsted Foundation, and the
Vissing Foundation.
251
252
REHFELD ET AL.
JCE & M • 2001
Vol. 86 • No. 1
FIG. 1. Scheme of pro-CCK processing
in I-cells of the small intestine.
Although pro-CCK may be processed to more than the
bioactive peptides mentioned above, the present study deals
only with the four major components that are resolved by the
chromatography employed. These are CCK-58, -33, -22, and
-8. Such distinction allows the best overview of the molecular
heterogeneity, but does not exclude that the chromatographic CCK-33 peak also contains a small amount of
CCK-39 and/or that the CCK-8 peak contain traces of CCK-7
(9). Metabolism of these variant forms, however, is similar to
that of the corresponding main form (i.e. CCK-39 vs. CCK-33,
and CCK-7 vs. CCK-8).
Materials and Methods
Tissue extracts
Biopsies of human jejunal or ileal mucosa were obtained from the
Department of Surgical Gastroenterology, and feline jejunal mucosa
from anesthetized cats was obtained from the Department of Experimental Pathology, Rigshospitalet (Copenhagen, Denmark). The biopsies
of human small intestine were histologically normal jejunal or ileal
mucosa from resections for midgut carcinoid tumors. The tissue samples
were immediately frozen in liquid nitrogen. The biopsy collection was
approved by the local ethics committee. The frozen tissue was cut into
pieces of a few milligrams, boiled in water (10 mL/g tissue) for 20 min,
homogenized, and centrifuged at 10,000 ⫻ g for 30 min at 4 C. The
supernatant was decanted (neutral extract), and the pellet was then
redissolved in ice-cold acetic acid (10 mL/g), left at room temperature
for 20 min, and centrifuged as described above (acid extract). The extracts were stored at ⫺40 C until analysis.
Plasma for CCK quantitations
Blood samples were collected into chilled tubes containing 3.9 ␮mol
ethylenediamine tetraacetic acid (EDTA)/mL blood. Within 30 min, the
samples were centrifuged at 3000 ⫻ g at 4 C for 10 min. The plasma was
stored at ⫺20 C until extraction, which was performed as follows. One
volume of plasma (usually 1.0 mL) was mixed with 2 vol 960 mL/L
ethanol on a whirlmixer for 10 s. The mixture was then centrifuged for
30 min at 1200 ⫻ g, and the supernatant was decanted and evaporated
at 37 C in a Speed-Vac concentrator (SVC 200 H, Savant Instruments).
The dried extracts were then reconstituted to the original volume with
assay buffer and assayed. The basal and postprandial concentrations in
plasma were measured in five healthy females and males (mean age, 36
yr) after an overnight fast. The meal consisted of an omelet (two eggs
mixed with 10 g flour, 25 mL cream, salt, and pepper) with two slices
of bacon, 250 mL orange juice, 250 mL milk, 250 mL yogurt, and two
slices of toasted bread with butter and cheese, i.e. 1470 calories, of which
CCK-33 IS THE MAJOR HORMONAL CCK IN MAN
45% was fat, 37% was carbohydrates, and 18% was protein. Blood
samples were taken from each of the subjects from 60 min before to 125
min after ingestion of the meal.
253
were calibrated with human CCK-33, CCK-22, and CCK-8 as well as with
[125I]albumin and 22NaCl to indicate void (Vo) and total (Vt) volumes.
Enzyme analysis
Plasma for chromatography
To measure the immunoreactivity of the larger endogenous forms of
CCK, chromatographic fractions of the jejunal tissue extracts were also
measured after tryptic cleavage. Each fraction was incubated with trypsin (100 mL/L trypsin-l-p-tosylamino-2-phenylethyl chloromethyl ketone; Worthington Biochemical Corp., Freehold, NJ) for 30 min at 20 C.
The cleavage was terminated by boiling the fraction for 10 min. Similarly, tissue extracts were cleaved with trypsin before chromatography
to ensure their molecular nature. The principles and performance of the
tryptic analysis were reported in detail previously (31).
Four healthy persons (two of each sex; mean age, 31 yr) ingested a
meal as described above. Blood samples (200 mL) were drawn from an
arm vein immediately before the meal as well as 30, 90, and 150 min
postprandially. Because it has been suggested (30) that acidification is
necessary to prevent in vitro degradation, one half of each blood sample
was drawn into EDTA tubes as described above (neutral sample), and
the other half was drawn into EDTA tubes containing 1 mL 0.5 mol/L
sodium acetate buffer (pH 3.6)/5 mL blood (acid sample). Immediately
after centrifugation, 50 mL of the neutral plasma samples were extracted
directly on Sep-Pak (Waters Corp., Milford, MA) cartridges, and 50 mL
of the acidified plasma were poured slowly into 150 mL 20 g/L trifluoroacetic acid (TFA) under constant stirring (23, 28). This mixture was
extracted on Sep-Pak C18 cartridges prewashed with 10 mL 960 mL/L
ethanol followed by 10 mL of a 13 mmol/L solution of TFA. Ten milliliters of ice-cooled plasma were then loaded on each cartridge at a flow
rate of 1 mL/min. After the cartridge was washed with 10 mL 13 mmol
TFA/L, the CCK peptides were eluted by 2 mL 800 mL/L ethanol
containing 13 mmol TFA/L. Evaporation of the eluates was performed
as previously described. All steps were performed consecutively without freezing the plasma or extracts.
CCK measurements
The measurements of CCK in tissue extracts, plasma, and chromatographic fractions were performed with RIA using a panel of eight high
titer CCK antisera specific for the C-terminal ␣-amidated sequence of
CCK as well as an antiserum specific for the N-terminus of human
CCK-22 and an antiserum specific for CCK glycine-extended at the
C-terminus. Characteristics of these antisera are presented in Tables 1
and 2. O-Sulfated CCK-8 and CCK-22 were used as standards and in
monoiodinated form as tracer labeled by the Bolton-Hunter technique
(32). Details of antibody production and assay characteristics were previously reported (28, 32–35). It should be noted that the antiserum library
includes Ab 92128. As previously demonstrated (28), Ab 92128 displays
a unique specificity, comprising equimolar binding of all CCK-A receptor ligand forms of CCK regardless of size and length, and at the same
time it does not bind any gastrins. The CCK assay using Ab 92128
consequently quantitates the different molecular forms of CCK accurately. Also the use of Ab 89009 (directed against the N-terminus of
Chromatography
One milliliter of tissue extract or plasma concentrate was applied to
a Sephadex G-50 superfine column (10 ⫻ 1000 mm) and eluted with
either 125 mmol/L NH4HCO3, pH 8.2, or 20 mmol/L sodium veronal,
pH 8.4, containing 0.6 mmol/L thiomersal and 1 g/L BSA at 4 C with
a flow rate of 4 mL/h. Fractions of 1.0 mL were collected. The columns
TABLE 1. Characteristics of antisera specific for the ␣-amidated C-terminus of bioactive CCKs
a
b
Antiserum
no.
Antibody
titera
Assay
sensitivityb
2609
2717
8007
92127
92128
92132
92136
92138
1:200,000
1:150,000
1:3,200,000
1:500,000
1:1,500,000
1:1,600,000
1:4,000,000
1:800,000
2.0
5.0
1.5
4.0
⬍0.5
⬍1.0
2.0
2.0
Reactivity with CCK/gastrin peptides (%)
CCK-8
(sulfated)
CCK-8
(non-sulfated)
CCK-33
(sulfated)
Gastrin-17
(non-sulfated)
100
100
100
100
100
100
100
100
25
63
40
31
⬍0.01
30
16
32
56
100
64
92
103
85
100
91
100
166
71
82
⬍0.01
368
16
248
Final antiserum dilution in RIA.
Lowest concentration of CCK-8(s) (picomoles per L) different from 0 pmol/L within 95% confidence limit under standard conditions.
TABLE 2. Concentrations of CCK-58-like peptides in eight chromatographic pools of extracts of human jejunal mucosa (picomoles per L)
as measured with RIAs using different C-terminal specific antisera
Extract no.
Antiserum
no.
1
2
3
4
5
6
7
8
2609
2717
8007
92127
92128a
92132
92136
92138
2604b
4
10
14
37
16
13
9
26
7
13
9
23
29
15
15
8
28
1
13
9
17
31
16
20
15
25
8
⬍0.1
⬍0.1
⬍0.1
3
1
⬍0.1
⬍0.1
2
6
93
64
115
223
106
60
35
78
⬍0.1
51
43
74
124
71
45
40
82
⬍0.1
44
42
68
138
71
54
70
80
⬍0.1
112
102
198
390
178
174
93
214
2
a
Antiserum 92128 binds the different molecular forms of CCK with equimolar potency and displays no cross-reactivity with gastrin peptides
(28).
b
Antiserum 2604 binds the different molecular forms of gastrin with equimolar potency and displays no cross-reactivity with CCK peptides
(33). Gastrin was measured for control.
254
JCE & M • 2001
Vol. 86 • No. 1
REHFELD ET AL.
human CCK-22) to measure CCK precursors (34) and Ab 3208 to measure glycine-extended CCKs should be noted (35).
Gastrin measurements
Control measurements of related gastrin peptides were performed by
RIA using antiserum 2604, which binds all carboxyamidated gastrins
(gastrin-71, -34, -17, and -14) with equimolar potency regardless of their
degree of sulfation. Antiserum 2604 does not bind any CCK peptide (33).
Results
Reactivity of CCK-58 from the human intestinal mucosa
with antisera specific for the ␣-amidated C-terminus
CCK-58-like peptides in chromatographic fractions of
the neutral and acid extracts of the four biopsies (i.e. eight
individual extract pools) were assayed in five different
dilutions. As shown in Table 2, the assays using C-terminal
specific antisera measured the content of CCK-58-like pep-
tides differently. The assay using Ab 92128 have been
shown to measure CCK-58, CCK-33, CCK-22, and the standard peptide (CCK-8) with equimolar potency (28). Therefore, the results obtained using this antiserum are assumed
to be accurate. Table 2 then shows that the assays employing antisera 92127 and 92138 overestimate the amount
of CCK-58, whereas the assays employing antisera 8007
and 92132 measures CCK-58 with almost the same potency
as Ab 92128. Finally, the assays using antisera 2609, 2717
and 92136 underestimate the amount of CCK-58 in the
jejunal mucosa.
CCK in extracts of jejunal mucosa
When measured with the specific CCK assay that binds the
four main forms of CCK with equimolar potency (using Ab
92128), the human jejunal mucosa contained, as shown in
TABLE 3. Concentrations of pro-CCK and progastrin products in human jejunal mucosa (picomoles per g)
␣-Amidated
CCK
Gly-extended
CCK
Pro-CCK and larger
processing
intermediates
␣-Amidated
gastrin
Gly-extended
gastrin
Progastrin and larger
processing
intermediates
Neutral extract
Acidic extract
27.6 ⫾ 8.2
10.1 ⫾ 3.7
1.9 ⫾ 0.5
⬍0.1
25.8 ⫾ 8.3
30.3 ⫾ 10.0
0.8 ⫾ 0.5
⬍0.1
⬍0.1
⬍0.1
⬍0.1
⬍0.1
Total conc.
37.7 ⫾ 11.9
1.9 ⫾ 0.5
56.1 ⫾ 18.1
0.8 ⫾ 0.5
⬍0.1
⬍0.1
Values are the mean ⫾
SEM
(n ⫽ 5).
FIG. 2. Gel chromatography of acid (E) and neutral (F) extracts of human jejunal mucosa. The extracts were applied to Sephadex G-50 superfine
columns (1 ⫻ 100 cm) and eluted with 0.02 mol/L sodium veronal, pH 8.4, containing 1 g/l BSA. The chromatographic elutions were monitored
with the CCK-specific RIA using antibody 92128. The elution positions of known molecular forms of CCK are indicated. The elution of the large
nonamidated processing intermediates (right) was monitored by a RIA specific for glycine-extended CCKs (using Ab.3208) after each fraction
was incubated with trypsin and carboxypeptidase B. The specificity was subsequently corroborated by measurement of the enzyme-treated
fractions with the human CCK-specific RIA (using Ab89009) (34).
CCK-33 IS THE MAJOR HORMONAL CCK IN MAN
255
TABLE 4. Molecular pattern of CCK in human jejunal mucosa
(percentage)
CCK-58-like
CCK-33-like
CCK-22-like
CCK-8-like
18.4 ⫾ 11.6
35.0 ⫾ 13.2
15.2 ⫾ 7.7
31.4 ⫾ 15.9
Values are the mean ⫾
SEM
(n ⫽ 5).
Table 3, a total of 37.7 ⫾ 11.9 pmol bioactive CCK/g tissue
(mean ⫾ sem; n ⫽ 5); bioactive CCK were products of proCCK that are both carboxyamidated and O-sulfated. In addition, the mucosal extracts contained, on the average, 1.9
pmol/g tissue of the immediate precursor, the glycineextended CCK, and 56.1 pmol/g tissue of even less mature
processing-intermediates and pro-CCK. Tiny amounts of
amidated gastrin (⬍1 pmol/g) were also detected in the
control measurements (Table 3). Chromatography revealed
a molecular pattern as shown in Fig. 2. Of the amidated and
O-sulfated bioactive forms, the longer CCK-58- and CCK33-like peptides were extracted under acidic conditions,
whereas the neutral extracts contained mainly CCK-22- and
CCK-8-like forms (Fig. 2, left). CCK-33-like peptides were
predominant and more abundant than CCK-58- and CCK22-like peptides (Fig. 2 and Table 4). Nonamidated pro-CCK
and/or processing intermediates occurred mainly in the neutral intestinal extract (Fig. 2, right). As they were measured
using an assay against glycine-extended CCK after in vitro
cleavage of the chromatographic fractions with trypsin and
carboxypeptidase B, the chromatography suggests that proCCK is cleaved at four basic sites (or more) in its N-terminal
sequence.
Gel chromatography of the feline jejunal extracts revealed
a pattern in which CCK-58-like peptides predominated, but
where significant amounts of CCK-33-, CCK-22-, and CCK8-like peptides were also present (Fig. 3).
CCK in plasma extracts
The concentrations of carboxyamidated and O-sulfated
CCK in plasma rose 5-fold after a protein-rich meal, i.e. from
0.9 to 4.6 pmol/L (Table 5). As shown in Table 6 and Fig. 4
(left), bioactive CCK eluted by chromatography in peaks
corresponding to CCK-58, CCK-33, CCK-22, and CCK-8. Of
these, the CCK-33-like peptides predominated, and the CCK22-like form was the second most abundant. CCK-58 and
CCK-8 constituted minor fractions of the plasma CCK regardless of whether neutral or acid extractions were examined chromatographically (Table 6).
Also plasma contained substantial amounts of nonamidated CCK precursors extended beyond glycine following
the C-terminal bioactive site (Fig. 4, right). As in the intestinal
mucosa, the nonamidated processing intermediates of proCCK in plasma eluted in four major peaks. The exact molecular identities of these peaks were not examined in this
study.
Discussion
This study has shown that intestinal mucosa as well as
plasma contain a multitude of pro-CCK products. In a functional sense the most important products are the CCK-A
receptor ligands, i.e. the pro-CCK products that are both
FIG. 3. Gel chromatography of acid (E) and neutral (F) extracts of
feline jejunal mucosa. The extracts was applied to Sephadex G-50
superfine columns (1 ⫻ 100 cm) and eluted with 0.02 mol/L sodium
veronal, pH 8.4, containing 1 g/L BSA. The chromatographic elutions
were monitored with the CCK-specific RIA using antibody 92128. The
elution positions of known molecular forms of CCK are indicated.
TABLE 5. Concentrations of bioactive CCK in human plasma
before and at intervals after ingestion of a protein-rich meal
(picomoles per L)
0 min
30 min
90 min
150 min
0.9 ⫾ 0.4
2.6 ⫾ 0.9
4.6 ⫾ 1.3
4.3 ⫾ 1.5
Values are the mean ⫾ SEM (n ⫽ 5). The measurements were
performed by RIA using antiserum 92128 that binds only carboxyamidated and tyrosine 0-sulfated (i.e. bioactive) CCK peptides.
carboxyamidated and O-sulfated. Of these, CCK-33 appears
to be the most abundant CCK peptide in both intestinal
extracts (⬃35%; Table 4 and Fig. 2) and plasma (⬃53%; Table
6 and Fig. 4) in man. As mentioned previously, it is possible
that a small fraction of what here has been characterized as
CCK-33 is, in fact, CCK-39 (11).
The tissue extracts also contain substantial amounts of
CCK-8 (31%), which contrasts with the lower percentage of
CCK-8 in plasma (⬃10%). We attribute this discrepancy to a
difference in the clearance rates of circulating CCKs. Hence,
the short CCK-8 peptide is cleared considerably faster from
plasma than the longer CCK-33 (36). The low fraction of
CCK-58-like peptides in human plasma (⬃10%) cannot be
explained by a rapid metabolism of CCK-58. On the contrary.
If carboxyamidated CCK peptides are released from the I
cells to plasma in ratios similar to those found in the extracts
of intestinal tissue, a high fraction of CCK-58 should be
expected in peripheral plasma, because the long CCK-58
peptide, in analogy with other heterogeneous peptide systems (37, 38), survives longer in circulation than shorter
256
JCE & M • 2001
Vol. 86 • No. 1
REHFELD ET AL.
TABLE 6. Molecular pattern of CCK in normal human plasma in the fasting state and at intervals after a meal (percentage)
Sampling time
(min)
0
Neutral
Acida
30
Neutral
Acida
90
Neutral
Acida
150
Neutral
Acida
Mean
Neutral
Acida
CCK-58-like
CCK-33-like
CCK-22-like
CCK-8-like
7.2 ⫾ 3.3
6.4 ⫾ 3.4
50.6 ⫾ 20.7
36.5 ⫾ 16.8
33.6 ⫾ 17.5
47.6 ⫾ 24.9
7.6 ⫾ 2.9
9.5 ⫾ 4.3
10.7 ⫾ 4.0
7.7 ⫾ 3.6
53.3 ⫾ 19.0
35.0 ⫾ 16.9
26.0 ⫾ 8.7
46.7 ⫾ 16.6
10.0 ⫾ 3.7
10.5 ⫾ 4.3
12.3 ⫾ 4.2
9.2 ⫾ 3.9
57.3 ⫾ 24.3
35.5 ⫾ 18.4
23.2 ⫾ 8.0
46.3 ⫾ 18.6
7.3 ⫾ 3.2
9.0 ⫾ 3.4
12.3 ⫾ 3.8
14.0 ⫾ 5.1
48.3 ⫾ 14.7
42.7 ⫾ 15.5
29.7 ⫾ 12.4
27.3 ⫾ 14.8
9.7 ⫾ 4.8
16.0 ⫾ 7.3
10.6
9.3
52.8
37.4
28.2
42.0
8.6
11.2
Values are the mean ⫾ SEM (n ⫽ 5).
a
The blood was sampled into one tube containing only EDTA (neutral sample) and another also containing sodium acetate (acid sample).
For details, see text.
FIG. 4. Gel chromatography of acid (E) and neutral (F) extracts of basal and postprandial plasma from a healthy young human subject. The
concentrated plasma extracts were applied to Sephadex G-50 superfine columns (1 ⫻ 100 cm) and eluted with 0.02 mol/L sodium veronal (pH
8.4) containing 1 g/L BSA. The chromatographic elutions were monitored with the CCK-specific RIA using Ab.92128. The elution positions of
known molecular forms of CCK are indicated. The elution of the large nonamidated processing intermediates (right) were monitored by a RIA
specific for glycine-extended CCKs (using Ab.3208) after each fraction was incubated with trypsin and carboxypeptidase B. The specificity was
subsequently corroborated by measurement of the enzyme-treated fractions with the human CCK-specific RIA (using Ab.89009; see Ref. 34).
forms. Other mechanisms, therefore, have to be considered
to explain the discrepancy between the abundance of CCK-58
in tissue and that in plasma. Such explanation might be that
I cells contain a mixture of immature and mature secretory
granules, and that the immature granules contain a larger
fraction of the less processed CCK-58 than mature granules.
Accordingly, the prohormone convertases (PC 1/3, PC 2, and
other endoproteases) have had more time for truncation of
the peptides in mature granules. As a larger fraction of mature granules is secreted upon stimulation, the percentage of
CCK-33 IS THE MAJOR HORMONAL CCK IN MAN
CCK-58 released from I cells is presumably lower than the
percentage in tissue extracts. Notably, in vitro degradation at
neutral pH cannot explain the low fraction of CCK-58 in
plasma, as the fraction of CCK-58 in blood samples drawn in
acidified vials contained an even lower percentage of CCK58. Hence, our data cannot support the contention of rapid
in vitro cleavage at neutral pH (22).
One of the new results of this study is the demonstration
of large quantities of biologically inactive processing intermediates in both tissue and plasma. Due to the nature of the
analysis used for their demonstration, i.e. measurement of
glycine-extended CCK after trypsin and carboxypeptidase B
cleavage, these long nonamidated polypeptides are processing intermediates, which are extended beyond Gly84 at the
C-terminus of pro-CCK. In other words, they are not simple
N-terminal desocta or desnona fragments of longer bioactive
forms of CCK. Such inactive fragments are also present in
tissue and plasma (23, 39), but they have not been measured
in this study. The occurrence of substantial amounts of
nonamidated CCK processing intermediates in human
plasma has been corroborated by processing-independent
analysis (40).
The debate about the molecular nature of CCK in plasma,
in particular the nature of CCK in human plasma, during the
last decade has to a large extent been a debate about CCK-58.
In other words, is CCK-58 indeed the predominant hormonal
form of CCK in man (for review, see Ref. 24)? It is by now
generally agreed that the molecular pattern of CCK in plasma
varies among mammals. Hence, plasma from pigs, rats, and
rabbits contain mainly CCK-22 and CCK-8 (23, 24, 41),
whereas canine plasma has been claimed to harbor mainly
CCK-58 (30). In man, however, the reports have been particular controversial. Eysselein, Reeve and co-workers have
in analogy with their canine studies (30) maintained that
CCK-58 is the major form in human plasma (22), whereas
others have found human plasma to contain a heterogeneous
mixture, with CCK-33 and CCK-22 being the major forms (18,
21, 27, 28, 42).
Three explanations have been offered to explain the discrepancy concerning CCK-58 in plasma. First, it was suggested that the amount of CCK-58 in plasma has been underestimated, because of specific in vitro degradation at
neutral pH (22). Therefore, a procedure involving acidification during sampling and processing of blood was proposed
to prevent degradation (22). Even strict adherence to the
proposed procedure, however, did not increase the amount
of CCK-58 in human plasma. It only decreased the extraction
efficiency for the shorter forms, but still not to a degree
making CCK-58 the predominant form. Thus, acid extraction
results in false low and unequal extraction of different CCK
peptides in circulation. Specific in vitro degradation of
plasma CCK-58, therefore, cannot explain the discrepancy.
The second explanation suggests that CCK-58, in contrast
to other forms of CCK, is bound less well to CCK receptors
and antibodies specific for the common C-terminal epitope
(43, 44). Accordingly, Reeve et al. suggested that CCK-58 has
a structure in which the N-terminal part shields the Cterminal bioactive sequence from receptor or antibody binding (44, 45). Such shielding should reduce the detection by
bioassays and immunoassays unless the CCK-58-like mate-
257
rial is exposed by tryptic release of CCK-8 (45). It is difficult
to know whether the shielding idea is correct. It is, however,
an old empirical observation that some C-terminal-directed
antibodies may have a lower affinity for N-terminally extended peptides. This was described 2 decades ago for the
longer forms of CCK and its homolog gastrin (31, 46). At that
time, trypsination was also proposed to ensure accurate immunochemical quantitation of larger CCKs (31). Some antibodies do, however, bind the longer forms with the same
affinity as shorter forms (28, 33, 46). Thus, the new antiserum,
Ab 92128, binds CCK-58, CCK-33, and CCK-22 with the same
affinity as CCK-8 (28). Shielding, therefore, does not interfere
with measurements based on Ab 92128, which, as shown in
this study, discloses that CCK-58 is only a minor form of CCK
in human plasma and tissue, but is a major form in the feline
jejunum. In human plasma CCK-33 and -22 are the two most
abundant CCKs.
The third explanation addresses some technological problems in the studies, which have proposed CCK-58 to be the
major CCK component in human plasma (22, 30). As emphasized previously (24, 28, 47, 48) measurement of CCK in
plasma is unusually difficult for several reasons, including
extensive C-terminal homology between CCK and gastrin,
low plasma concentrations of CCK (⬎10-fold lower than
those of gastrin), and a high degree of molecular heterogeneity for both CCK and gastrin. For practical purposes,
plasma CCK assays therefore have to be sensitive RIAs specific for the C-terminus of CCK, but without cross-reactivity
with gastrin. The proponents of high levels of CCK-58 in
human plasma, however, used subtraction assay systems (22,
30). Subtraction based on combinations of a C-terminal assay,
which measures all the carboxyamidated CCKs and gastrins,
a gastrin-specific assay, and/or high pressure liquid chromatography separation of the molecular forms of gastrin and
CCK may in theory provide accurate measurements, but in
practice they never do (see also arguments in Ref. 47). First,
both CCK and gastrin circulate in forms of different lengths
and degrees of amino acid derivatization. The reactivity of all
circulating forms of CCK and gastrin therefore has to be
tested with both the cross-reacting and the gastrin-specific
assay. None of the laboratories employing subtraction assays
has to date reported such specificity evaluations. For instance, neither sulfated human gastrin-34, sulfated human
gastrin-14, nor sulfated gastrin-6 has been tested in RIAs or
high pressure liquid chromatography systems employed in
subtraction assays. Second, since gastrin circulate in concentrations 10 –50 times above those of CCK, even a minor gastrin component may easily appear as a major CCK component. Third, identification by reverse phase HPLC of the
various molecular forms of CCK and gastrin requires caution. Minute modifications by, for instance, methionyl oxidation or conservative amino acid substitutions among species (i.e. man vs. pig) may change the elution position of
sample or calibrator peptide.
Taken together, evaluation of available data indicates that
CCK-58 is not the major form of CCK in human plasma. The
predominant form is CCK-33, and CCK-22 is the second most
abundant. Moreover, the data suggest that subtraction assay
technology cannot be used to characterize CCK in plasma. A
sensitive and specific bioassay has been shown to provide
258
REHFELD ET AL.
accurate plasma CCK results (18), but it is too labor-intensive,
costly, and inconvenient. Therefore, only sensitive and specific CCK RIAs that recognize all the molecular forms of CCK
with similar potency can be used to characterize CCK in
human plasma.
Acknowledgments
The skillful technical and secretarial assistance of Alice von der Lieth,
Rikke Grønholt Pedersen, and Gitte Runge is gratefully acknowledged.
Also, the kind help of Anders H. Johnsen, D.Sc., in controlling the
amount and structure of synthetic peptides for standards and calibrations is gratefully acknowledged. This paper is dedicated to the memory
of Viktor Mutt, who died only 2 yr ago. Viktor Mutt pioneered the
identification of CCK in tissue, and he did it for good reasons as CCK-33
(7, 11).
References
1. Rehfeld JF. 1989 Cholecystokinin. In: Makhlouf GM, vol ed; Shultz SG, sect ed.
Handbook of physiology: the gastrointestinal system, neural and endocrine
biology. Baltimore: American Physiological Society; vol 11:337–358.
2. Liddle RA. 1994 Cholecystokinin. In: Walsh JH, Dockray GJ, vol eds. Gut
peptides. Comprehensive endocrinology. New York: Raven Press; 175–216.
3. Buffa R, Solcia E, Go VLW. 1976 Immunohistochemical identification of the
cholecystokinin cell in the intestinal mucosa. Gastroenterology. 70:528 –532.
4. Buchan AMJ, Polak JM, Solcia E. 1978 Electron immunohistochemical evidence for the human intestinal I-cells as source of CCK. Gut. 19:403– 407.
5. Rehfeld JF. 1978 Immunochemical studies on cholecystokinin. II. Distribution
and molecular heterogeneity in the central nervous system and small intestine
of man and hog. J Biol Chem. 253:4022– 4030.
6. Larsson L-I, Rehfeld JF. 1979 Localization and molecular heterogeneity of
cholecystokinin in the central and peripheral nervous system. Brain Res.
165:201–218.
7. Mutt V, Jorpes JE. 1971 Hormonal polypeptides of the upper intestine. Biochem J. 125:57–58.
8. Eysselein VE, Reeve JR, Shively JE, Hawke D, Walsh JH. 1982 Partial structure of a large canine cholecystokinin (CCK-58): amino acid sequence. Peptides. 3:687– 691.
9. Rehfeld JF, Hansen HF. 1986 Characterization of procholecystokinin products
in the porcine cerebral cortex: evidence of different processing pathways. J Biol
Chem. 261:5832–5840.
10. Eng J, Du B-H, Pan Y-C, Chang M, Hulmes JD, Yalow RS. 1985 Purification
and sequencing of a rat intestinal 22 amino acid C-terminal CCK fragment.
Peptides. 5:1203–1206.
11. Mutt V. 1976 Further investigations on intestinal hormonal polypeptides. Clin
Endocrinol (Oxf). 5:175s–183s.
12. Reeve JR, Eysselein VE, Walsh JH, Ben-Avram CM, Shively JE. 1986
New molecular forms of cholecystokinin: microsequence analysis of forms
previously characterized by chromatographic methods. J Biol Chem.
261:16392–16397.
13. Eberlein GE, Eysselein VE, Davis MT, et al. 1992 Patterns of prohormone
processing: order revealed by a new procholecystokinin-derived peptide. J Biol
Chem. 267:1517–1521.
14. Rehfeld JF. 1985 Neuronal cholecystokinin: one or multiple transmitters?
J Neurochem. 44:1–10.
15. Beinfeld MC. 1997 CCK biosynthesis and processing: recent progress and
future challenges. Life Sci. 61:2359 –2366.
16. Walsh JH, Lamers CBHW, Valenzuela JE. 1982 Cholecystokinin-octapeptidelike immunoreactivity in human plasma. Gastroenterology. 82:538 –544.
17. Calam J, Ellis A, Dockray GJ. 1982 Identification and measurement of molecular variants of cholecystokinin in duodenal mucosa and plasma. J Clin
Invest. 69:218 –225.
18. Liddle RA, Goldfine ID, Rosen MS, Taplitz RA, Williams JA. 1985 Cholecystokinin bioactivity in human plasma. J Clin Invest. 75:1144 –1152.
19. Maton PN, Selden AC, Chadwich VS. 1982 Large and small forms of cholecystokinin in human plasma: measurement using high pressure liquid chromatography and radioimmunoassay. Regul Pept. 4:251–260.
20. Jansen JBMJ, Lamers CBHW. 1983 Molecular forms of cholecystokinin in
human plasma during infusion of bombesin. Life Sci. 33:2197–2205.
JCE & M • 2001
Vol. 86 • No. 1
21. Cantor P, Rehfeld JF. 1987 The molecular nature of cholecystokinin in human
plasma. Clin Chim Acta. 168:153–158.
22. Eberlein GA, Eysselein VE, Hesse WH, Goebell H, Schaefer M, Reeve JR.
1987 Detection of cholecystokinin-58 in human blood by inhibition of degradation. Am J Physiol. 253:G477–G482.
23. Cantor P, Rehfeld JF. 1989 Cholecystokinin in pig plasma: release of components devoid of a bioactive COOH-terminus. Am J Physiol. 256:G53–G61.
24. Rehfeld JF. 1998 How to measure cholecystokinin in tissue, plasma and cerebrospinal fluid. Regul Pept. 78:31–39.
25. Eberlein GA, Eysselein VE, Goebell H. 1988 Cholecystokinin-58 is the major
molecular form in man, dog and cat but not in pig, beef and rat intestine.
Peptides. 9:993–998.
26. Eysselein VE, Eberlein GA, Schaeffer M, et al. 1990 Characterization of the
major form of cholecystokinin in human intestine: CCK-58. Am J Physiol.
258:G253–G260.
27. Ballinger A, McLoughlin L, Medbak S, Clark M. 1995 Cholecystokinin is a
satiety hormone in humans at physiological post-prandial plasma concentrations. Clin Sci. 89:375–381.
28. Rehfeld JF. 1998 Accurate measurement of cholecystokinin in plasma. Clin
Chem. 44:991–1001.
29. Wank SA. 1995 Cholecystokinin receptors. Am J Physiol. 269:G628 –G646.
30. Eysselein VE, Eberlein GA, Hesse WH, Singer MV, Goebell H, Reeve JR.
1987 Cholecystokinin-58 is the major circulating form of cholecystokinin in
canine blood. J Biol Chem. 262:214 –217.
31. Magistris L de, Rehfeld JF. 1980 A simple enzymatic procedure for radioimmunochemical quantitation of the large molecular forms of gastrin and cholecystokinin. Anal Biochem. 102:126 –133.
32. Rehfeld JF. 1978 Immunochemical studies on cholecystokinin. I. Development
of sequence-specific radioimmunoassays against porcine triacontatriapeptide
cholecystokinin. J Biol Chem. 253:4016 – 4021.
33. Rehfeld JF, Stadil F, Rubin B. 1972 Production and evaluation of antibodies
for the radioimmunoassay of gastrin. Scand J Clin Lab Invest. 30:221–232.
34. Paloheimo LI, Rehfeld JF. 1994 A processing-independent assay of human
procholecystokinin and its products. Clin Chim Acta. 229:49 – 65.
35. Hilsted L, Rehfeld JF. 1986 Measurement of precursors for ␣-amidated hormones by radioimmunoassay of glycine-extended peptides after trypsincarboxypeptidase B cleavage. Anal Biochem. 152:119 –126.
36. Gores GJ, LaRusso NF, Miller LJ. 1986 Hepatic processing of cholecystokinin
peptides. Am J Physiol. 250:G344 –G349.
37. Walsh JH, Isenberg JI, Ansfield J, Maxwell V. 1976 Clearance and acidstimulating action of human big and little gastrins in duodenal ulcer subjects.
J Clin Invest. 57:1125–1131.
38. Hansen CP, Stadil, Rehfeld JF. 1996 Hepatic and intestinal elimination of
endogenous gastrins in pigs. Digestion. 57:356 –361.
39. Blanke SE, Johnsen AH, Rehfeld JF. 1993 N-terminal fragments of intestinal
cholecystokinin: evidence for release of CCK-8 by cleavage on the carboxyl side
of Arg74 in proCCK. Regul Pept. 46:575–582.
40. Paloheimo LI, Rehfeld JF. 1995 Quantitation of procholecystokinin and its
products in plasma by processing-independent analysis. Clin Chim Acta.
238:21–33.
41. Fölsch UR, Cantor P, Wilms H, Schafmayer A, Becker HD, Creutzfeldt W.
1987 The role of cholecystokinin as a regulator of the feedback control of
pancreatic enzyme secretion in conscious rats. Gastroenterology. 92:449 – 458.
42. Kanayama S, Himeno S, Higashimoto Y, Yamasaki Y, Kitani T, Seiichiro T.
1987 Plasma cholecystokinin-octapeptide like immunoreactivity in patients
with hepatic cirrhosis. Life Sci. 41:1915–1920.
43. Turkelson CM, Reeve JR, Solomon TE. 1990 Low immunoreactivity of canine
cholecystokinin-58. Gastroenterology. 99:646 – 651.
44. Reeve JR, Eysselein VE, Ho FJ, Chew P, Vigna SR, Liddle RA, Evans C. 1994
Natural and synthetic CCK-58: novel reagents for studying cholecystokinin
physiology. Ann NY Acad Sci. 713:11–21.
45. Reeve JR, Eysselein VE, Rosenquist G, et al. 1996 Evidence that CCK-58 has
structure that influences its biological activity. Am J Physiol. 270:G860 –G868.
46. Rehfeld JF, de Magistris L, Andersen BN. 1981 Sulfation of gastrin: effect on
immunoreactivity. Regul Pept. 2:333–342.
47. Rehfeld JF. 1984 How to measure cholecystokinin in plasma? Gastroenterology. 87:434 – 438.
48. Rehfeld JF. 1988 Radioimmunoassay problems for gut hormones in the eighties. Adv Metab Disord. 11:45– 67.
49. Lacourse KA, Friis-Hansen L, Samuelson LC, Rehfeld JF. 1998 Altered
procholecystokinin processing in carboxypeptidase E-deficient mice: differential synthesis in neurons and endocrine cells. FEBS Lett. 436:61– 66.