Clinical Chemistry 44:5
991–1001 (1998)
Endocrinology and
Metabolism
Accurate measurement of cholecystokinin
in plasma
Jens F. Rehfeld*
plasma measurements. Several laboratories have tried to
develop RIAs. The first difficulty in this endeavor was a
shortage of peptides for immunizations. Mutt and Jorpes
reported in 1971 the structure of porcine CCK-33 (7), but
synthesis of sulfated CCK-33 became possible only recently, and only limited amounts of natural CCK were
available. A batch of sulfated CCK-8 was synthesized (8),
but antibodies against it also reacted with gastrin. The
next difficulty was the isotopic labeling, because methionyl residues are oxidized easily, and CCK-8 contain two
such residues. The problem was solved by nonoxidative
labeling (9). The third difficulty is the low concentrations
of CCK in plasma, which requires antibodies of high
affinity and tracers of high specific activity. The largest
obstacle, however, is specificity. Hence, the antibodies
should bind the “active site” of CCK without binding the
homologous gastrin. Alignment of the active site sequences of CCK and gastrin (Fig. 1) illustrates the problem, which seems unsurmountable because antibody
binding sites are assumed to harbor epitopes of no more
than four to six amino acid residues (10). The specificity
problem is accentuated by the higher gastrin concentrations in plasma.
In spite of the somber prognosis, we have tried to raise
useful antibodies. Insufficient specificity of the first led to
design of haptens to yield antibodies that might bind both
the C-terminal phenylalanyl amide and the CCK-specific
tyrosyl sulfate in position 7 (as counted from the Cterminus). Using a directionally coupled CCK-12-analog,
we have raised such an antiserum.
Shortage of reliable plasma assays has hampered studies of cholecystokinin (CCK). The assay problems are
low plasma concentrations, extensive molecular heterogeneity, and close homology of CCK to gastrin, which
circulates in higher concentrations. To develop an accurate CCK RIA, antibodies were raised in rabbits, guinea
pigs, and mice in titers from 200 to 4 000 000. The
specificity of the antisera was tested with homologous
peptides, and tissue and plasma extracts. Rabbit 92128
produced antibodies in high titer (>500 000) with sufficient avidity (K°eff > 1012 mol21) and the desired specificity. The antiserum binds the bioactive forms of CCK
with equimolar potency and displays no reactivity with
gastrin. CCK concentrations in plasma from healthy
humans rose from 1.13 6 0.10 pmol/L (mean 6 SE, n 5
26) to 4.92 6 0.34 pmol/L after a mixed meal. Chromatography of human plasma revealed traces of CCK-58, a
predominance of CCK-33 and CCK-22, and moderate
amounts of CCK-8. The results show that it is possible
to produce specific CCK-antisera using a sulfated
CCK-12 analog.
The gut hormone cholecystokinin (CCK) regulates pancreatic growth, enzyme secretion, and contraction of the
gallbladder. Moreover, CCK influences intestinal motility
and satiety. CCK is also a widespread transmitter in the
nervous system. Finally, CCK is involved in pancreatic
carcinogenesis (for reviews, see references 1–3). The role
of circulating CCK, however, is in many instances still
obscure because quantitation of the hormone in plasma is
difficult (for review, see reference 4). Thus, earlier reports
on plasma CCK may be inaccurate because the assays
used did not possess the necessary sensitivity and specificity.
A sensitive and specific bioassay has been described
(5, 6), but it is labor-intensive and has seen limited use for
definition of CCK peptides
Human proCCK is a protein of 95 amino acid residues.
Sequence 84 – 86 (Gly-Arg-Arg) constitutes the crucial
amidation site, which requires processing by prohormone
convertases, carboxypeptidase E, and the amidation enzyme complex to release bioactive carboxyamidated CCK
peptides. The largest bioactive form is CCK-83, which
corresponds to the amidated sequence 1– 83 of proCCK.
This sequence is cleaved variably at four monobasic sites
to release CCK-58, CCK-33, CCK-22, and CCK-8, all of
which have the same C-terminal heptapeptide amide
Dept. of Clinical Biochemistry, Rigshospitalet, DK-2100 Copenhagen, Denmark.
*Author for correspondence. Fax 45 3545 4640; e-mail [email protected].
Received October 16, 1997; revision accepted December 22, 1997.
991
992
Rehfeld: Plasma cholecystokinin
Fig. 1. (Top) Primary structure of the O-sulfated heptapeptide amide sequence, which constitutes the specific “active site” of CCK; (Bottom)
corresponding gastrin sequence.
sequence (-Tyr(SO32 ) - Met - Gly - Trp - Met - Asp - PheNH2,
Fig. 1), which is necessary for receptor binding. Binding to
CCK-A receptors requires that the tyrosyl residue of the
heptapeptide amide is O-sulfated (Fig. 1), whereas CCK-B
receptors do not discriminate between sulfated and nonsulfated CCK-peptides or between CCK and gastrin peptides (for review, see references 1–3).
Materials and Methods
antigens
Natural porcine O-sulfated CCK-33 was generously donated by Viktor Mutt, Karolinska Institutet, Sweden.
Nonsulfated synthetic porcine CCK-29 and CCK-33 were
generous gifts from Ulf Ragnarsson, University of Uppsala, Sweden. The C-terminal amide, CCK-4, nonsulfated
porcine CCK-8, CCK-12, CCK-13, and an analog of human
O-sulfated CCK-12 (Gly-Gly-Asp-Arg-Asp-Tyr (SO4)Met-Gly-Trp-Met-Asp-Phe-NH2) were custom synthesized by ZENECA. The purity, structure, and amount of
the synthetic peptides were controlled in our laboratory
by reversed-phase HPLC, gas-phase microsequencing,
and mass spectrometry.
The first batch of natural porcine CCK-33 (30 mg, 20%
purity) was dissolved in 12 mL of 0.05 mol/L sodium
phosphate, pH 7.5, and divided into four portions, which
were stored at 220 °C. The next batches of natural porcine
CCK-33 (30 mg, 20% purity and 8 mg, 75% purity),
synthetic nonsulfated CCK-33 and CCK-29 (6 and 5 mg,
85% purity) were each dissolved in 1.0 mL of N,N-
dimethylformamide and conjugated to 25 or 30 mg of
bovine serum albumin (the Serum Institute) and dissolved in 2.5 mL of 0.05 mol/L sodium phosphate, pH 7.5,
by the addition of 135 mg of 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide HCl (Sigma Chemical) to give
molar ratios between the CCK-peptide, albumin, and
ethyl-carbodiimide of 1:0.1:40. The reagents were mixed
for 20 h at 20 °C, and each conjugate was then divided
into six portions and stored at 220 °C until immunization.
Sulfated and nonsulfated synthetic CCK-4, CCK-8,
CCK-12, and CCK-13 (2.0, 4.0, and 6.0 mg of each) were
dissolved in 10 mL of 0.05 mol/L sodium phosphate, pH
7.5, and conjugated to 15 mg of bovine serum albumin by
dropwise addition of 100 mL of 500 g/L glutaraldehyde.
The solution was mixed for 4 h at 20 °C, applied to a
calibrated Sephadex G-10 column (10 3 60 mm), and
eluted at 20 °C with 0.05 mol/L sodium phosphate, pH
7.5, in fractions of 1.0 mL. The void volume fractions
containing the conjugate were pooled, divided into six
portions, and stored at 220 °C until immunization.
tracers
Two types of tracers were used: Sera from rabbits immunized with CCK-33 and CCK-29 were examined with
CCK-33 labeled by nonoxidative conjugation of [125I]hydroxyphenylpropionic-succinimide ester to either the
a- or e- NH2 groups of the N-terminal lysyl residue
as detailed elsewhere (9, 11, 12). Sera from rabbits immunized with CCK-13, CCK-12, CCK-8, CCK-4, and
993
Clinical Chemistry 44, No. 5, 1998
the corresponding analogs were examined with BoltonHunter-labeled CCK-8 (Amersham). The tracers displayed specific radioactivities of 1500 –1700 kCi/mol.
Thus, the 1000-cpm tracer used for the RIA incubations
corresponds to ;0.5 fmol of peptide.
immunizations
Randomly bred white Danish rabbits (n 5 78) were used
for immunizations in nine series of 6 –16 rabbits each. In
addition, one series of 29 guinea pigs and one of 8 mice
were immunized (Table 1). On the basis of our experience
with the production of antibodies towards other peptides
(13–21), we chose a specific immunization procedure for
all of the animals: The first portion of the antigen was
suspended in 8.5 g/L saline to a volume of 5 mL and
emulsified with an equal volume of Freund’s complete
adjuvant (the Serum Institute). Two subcutaneous injections of the mixture were given over the hips in amounts
corresponding to 100 mg of CCK-33 (;40 mg of CCK-12 or
25 mg of CCK-8) per animal. Five or more booster injections using Freund’s incomplete adjuvant were administered simultaneously at 8-week intervals, using one-half
of the initial dose of antigen per immunization. The
rabbits were bled from an ear vein 10 days after immunization. Guinea pigs and mice were bled by cardiac
puncture 10 days after immunization. Sera from the
bleedings were separated and stored at 220 °C.
antiserum evaluation
Four characteristics of sera from the immunized animals
were examined: (a) Titer was defined as the serum dilution that binds 33% of the 0.5-fmol tracer at equilibrium.
(b) Affinity (expressed by the “effective” equilibrium
constant (K°eff) was determined as the slope of the curve at
zero peptide concentration in a Scatchard plot (22, 23).
(c) Specificity was determined in percentage as the molar
ratio of the concentrations of the CCK standard and the
related peptide that produced a 50% inhibition of the
binding of the tracer. (d) Homogeneity of the antibodies
with respect to binding kinetics as expressed by the Sips
index (24). An index of 1.0 indicates homogeneity of both
the tracer and the antiserum in the binding, otherwise
seen only for monoclonal antibodies (25). The ability of
the peptides to displace tracers from the antisera was
tested in peptide concentrations of 0, 3, 10, 30, 100, 1000,
10 000, and 100 000 pmol/L.
ria procedure
The RIA was carried out as an equilibrium system at pH
8.4 in disposable plastic tubes, using 0.02 mol/L barbital
buffer, pH 8.4, containing 1 g/L bovine serum albumin
(Ortho). Incubation mixtures of 2.4 mL contained 2.0 mL
of antiserum dilution, 250 mL of tracer solution (giving
1000 cpm, corresponding to 0.5 fmol of freshly prepared
125
I-labeled CCK-peptide), and 150 mL of standard solution or sample. The assays were set up at room temperature and incubated at 4 °C for 48 h to reach equilibrium.
Antibody-bound (B) and free (F) tracers were separated
by the addition of 0.5 mL of a suspension of 20 mg of
activated charcoal (Merck) and blood plasma (equivolume mixture of buffer and outdated human plasma from
the Blood Bank of Rigshospitalet, Denmark) in 0.02 mol/L
sodium phosphate, pH 7.4, to each tube. The tubes were
centrifuged for 10 min at 2000 rpm, and the supernatant
(B) and sedimented charcoal (F) were counted in automatic g-scintillation counters for 5 min. The binding
percentage was calculated as B 2 [(B 1 F) 3 D]/(B
1 F) 2 [(B 1 F) 3 D] 3 100, where the “damage”
(D) is defined as B 3 100/(B 1 F) in the absence of
antiserum. The damage was usually 2–3%, and the
antisera were used at dilutions giving a binding percentage of ;33%. All samples were assayed in duplicate.
Table 1. Antigens and antibody characteristics in the production of CCK antisera.
Immunization seriesa
Antigenb
Antibody titer,c range
Antibody affinity,d
range
Antibody homogeneity,e
range
I (4464–4492)
II (A1–A7)
III (4696–4704)
IV (R1–R8)
V (R9–R16)
VI (R17–R24)
VII (M1–M8)
VIII (1557–1564)
IX (4815–4822)
X (8007–8012)
XI (92 125–92 140)
pCCK-33 (s)
pCCK-33 (s)
pCCK-33 (s)
pCCK-29 (ns)
pCCK-33 (ns)
CCK-4
CCK-8 (ns)
pCCK-33 (s)
pCCK-12 (ns)
pCCK-13 (ns)
hCCK-12 (s)
12 000–50 000
,1000–2000
,1000–30 000
,1000–2000
,1000–1000
,1000
,1000
4000–540 000
,1000
1000–2 000 000
2000–4 000 000
0.20–211
,0.01
,0.01–12
,0.01
,0.01
,0.01
,0.01
0.66–44
,0.01
0.04–154
,0.01–326
0.40–1.02
a
Rabbits were used for immunization except in series I (guinea pigs) and VII (mice).
CCK peptide covalently coupled to bovine serum albumin.
c
Final antiserum dilution in RIA.
d
Effective equilibrium constant (K ° eff(1010mol21)) according to Ekins and Newman (24).
e
Sips index of homogeneity [a, (25)].
b
0.74–1.04
0.92–1.01
0.65–1.02
0.62–1.03
994
Rehfeld: Plasma cholecystokinin
assay reliability
The assays were evaluated with respect to detection limit,
specificity, between- and within-assay reproducibility,
and accuracy.
tissue extracts
Biopsies of human jejunal mucosa were obtained from the
Department of Surgical Gastroenterology and porcine
jejunal mucosa from anesthetized pigs at the Department
of Experimental Pathology, Rigshospitalet. The tissue
samples were immediately frozen in liquid nitrogen. The
frozen tissue was cut in pieces of a few milligrams, boiled
in water (10 mL/g tissue) for 20 min, homogenized, and
centrifuged at 10 000g for 30 min at 4 °C. The supernatant
was decanted (neutral extract), and the pellet was redissolved in ice-cold acetic acid (10 mL/g), left at room
temperature for 20 min, and centrifuged as described
above (acid extract).
plasma extracts
Blood samples were collected into chilled tubes containing 3.9 mmol of EDTA per mL of blood. Within 30 min, the
samples were centrifuged at 3000g at 4 °C for 10 min. The
plasma was stored at 220 °C until extraction, which was
performed as follows. One volume of plasma (usually 1.0
mL) was mixed with two volumes of 960 mL/L ethanol
on a whirlmixer for 10 s. The mixture was then centrifuged in 30 min at 1200g; the supernatant was decanted
and evaporated at 37 °C in a speed-vac concentrator (SVC
200 H, Savant). 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 15 healthy females and 11 healthy males
(mean age, 36 years) after an overnight fast. The meal
consisted of an omelet (two eggs mixed with 10 g of flour,
25 mL of cream, salt, and pepper) with two slices of bacon,
250 mL of orange juice, 250 mL of milk, 250 mL of yogurt,
and two slices of toasted bread with butter and cheese,
i.e., 1470 calories of which 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.
slowly into 150 mL of 20 g/L trifluoroacetic acid (TFA)
under constant stirring (26). This mixture was extracted
on Sep-Pak C-18 cartridges (Waters Associates) prewashed with 10 mL of 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
with a flow rate of 1 mL/min. After the cartridge was
washed with 10 mL of 13 mmol TFA/L, the CCK-peptides
were eluted by 2 mL of 800 mL/L ethanol containing 13
mmol TFA/L. Evaporation of the eluates was performed
as described. All steps were performed consecutively
without freezing the plasma or extracts.
chromatography
One milliliter of tissue extract or plasma concentrate was
applied to a Sephadex G-50 superfine column (10 3 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 bovine serum albumin at
4 °C with a flow rate of 4 mL/h. Fractions of 1.0 mL were
collected. The columns were calibrated with human CCK33, CCK-22, and CCK-8, as well as with 125I-albumin and
22
NaCl to indicate void (Vo) and total (Vt) volumes.
enzyme analysis
To measure the immunoreactivity of the larger endogenous forms of CCK using antiserum 92128, chromatographic fractions of the jejunal tissue extracts were also
measured after tryptic cleavage. Each fraction was incubated with trypsin (100 mg/L Trypsin-TPCK ,Worthington) for 30 min at 20 °C. Tryptic cleavage was terminated
by boiling the fraction for 10 min. Similarly, tissue extracts
were also cleaved with trypsin before chromatography to
ensure their CCK nature. Principles and details of the
tryptic analysis have been described elsewhere (27).
gastrin measurements
Control measurements of gastrin in plasma were performed with a RIA using antiserum 2604, which binds all
bioactive, carboxyamidated gastrins (gastrin-71, -34, -17,
and -14) with equimolar potency irrespective of their
degree of sulfation (14, 15). The assay does not measure
CCK-peptides.
plasma for chromatography
Four healthy persons (two of each sex, 23- to 40-years-old)
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 (26) 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) per 5 mL of blood (acid sample (26)).
Immediately after centrifugation, 50 mL of the neutral
plasma samples was extracted directly on Sep-Pak cartridges, and 50 mL of the acidified plasma was poured
gastrin infusions
After an overnight fast, six healthy male volunteers (24- to
29-years-old) were intubated with a nasogastric tube (AN
10, Anderson Amplers). The position of the tube was
controlled by fluoroscopy. The gastric contents were
aspirated by intermittent pump suction. The infusions
were preceded by an equilibrium period of at least 30 min,
during which saline was infused. Subsequently, increasing doses of synthetic human nonsulfated gastrin-17
(0 – 60 pmol z kg z h; Sigma Chemical) were administered
continuously into a cubital vein. Each dose (0, 10, 30, and
60 pmol z kg z h) was administered for 45 min. Venous
blood samples were drawn from the opposite cubital vein
995
Clinical Chemistry 44, No. 5, 1998
every 15 min into EDTA tubes as described above. The
tubes were immediately placed in crushed ice and centrifuged at 4 °C within 30 min. The plasma samples were
stored at 220 °C until the measurement of gastrin and
CCK.
ethics
The studies reported here were approved by the local
committee for ethics in studies of human subjects.
Results
antibody production
The response to the immunizations varied between and
within the series (Table 1). The modest or low antibody
titers in the early series (I-VII) may be due to inadequate
methods for measuring hapten in the antigen-conjugate, a
low degree of hapten purity, degraded peptides, and less
efficient coupling procedures. Nevertheless, a few antisera in series I and III had titers and binding affinities of
some usefulness (Table 1). Two of these (4478 and 4698)
were also specific for CCK and have proved useful for the
characterization of CCK-peptides in tissue (9, 28 –30).
However, they were not useful for plasma measurements
because there was too little antiserum (small volume of
heartblood from guinea pigs) in titers that were too
modest (,50 000) and because the animals died before the
quality of their antibodies was known. In addition, only
antiserum 4698 had the desired specificity, but it lacked
sufficient affinity to measure plasma concentrations in
fasting subjects.
High titers and affinities were obtained in series VIII
and X (Table 1), but none of the antisera had the specificity necessary for plasma measurement. Although several
high-titer and high-affinity antisera were C-terminal directed (1559, 1560, 1564, 8007, and 8011), they all crossreacted with gastrin peptides to a degree making them
useless for plasma measurement. One high-titer anti-
serum in series VIII (1561) was monospecific for sequence
16 –20 of porcine CCK-33. This antiserum has proved
valuable for plasma measurement and affinity purification of porcine N-terminal desocta- and desnona-CCK-58,
CCK-39 and CCK-33 fragments (31, 32). Unfortunately,
the epitope was specific for a porcine CCK sequence.
Hence, species variations of the sequence N-terminal for
the bioactive C-terminal heptapeptide amide make antisera like 1561 useless for plasma measurements in man
and other nonporcine species.
The last series (XI) was composed of 16 rabbits that all
responded; some with high titers and affinities (Table 1).
Among the five antisera containing C-terminal-directed
antibodies with binding affinities sufficient for measurement of basal CCK concentrations in plasma and with
titers sufficient for long-term measurements (Table 2),
only one antiserum (92128) displayed a specificity acceptable for the plasma measurement of CCK. In addition, it
also binds the nonmammalian CCK homologs, cionin and
caerulein, with full potency (Table 2).
specificity characteristics of antiserum 92128
The epitope for antiserum 92128 was delineated on the
basis of the reactivity with related peptides (Table 2). The
specificity was then corroborated by measurements of
tissue and plasma extracts (Figs. 2 and 3). The results
confirmed that antiserum 92128 binds O-sulfated CCK-8,
CCK-22, CCK-33, and CCK-58 (as shown by tryptic cleavage) with equimolar potency (Table 2, Figs. 2 and 3).
The single most crucial specificity problem for plasma
CCK assays is interference from circulating gastrins. This
point was examined in different ways: (a) Chromatography of plasma extracts revealed that antiserum 92128
bound no gastrin (data not shown); (b) no correlation
existed between CCK and gastrin concentrations in human plasma samples covering a wide range of concentrations (Fig. 4); and (c) infusion of gastrin-17 into human
Table 2. Specificitya of high-affinity (K°, > 1012 mol21) and high-titer (final dilution, >0.5 3 106) CCK antisera in
immunization series XI.
Peptides
Ab.b 92127
Ab. 92128
Ab. 92132
Ab. 92136
Ab. 92137
CCK-8 (sulfated)
CCK-8 (nonsulfated)
CCK-22 (sulfated)
CCK-33 (sulfated)
CCK-5
CCK-8-gly (sulfated)
CCK-8-gly (nonsulfated)
Gastrin-17 (sulfated)
Gastrin-17 (nonsulfated)
Cionin (disulfated)
Caerulein (sulfated)
100.0
31.0
95.0
92.0
,0.01
,0.01
,0.01
88.0
82.0
10.0
39.0
100.0
,0.01
96.0
103.2
,0.01
,0.01
,0.01
0.4
,0.01
119.0
160.0
100.0
30.0
92.3
84.6
20.0
,0.01
4.0
836.0
368.0
112.0
108.0
100.0
16.0
96.0
100.0
1.7
,0.01
2.5
52.0
16.0
97.0
98.0
100.0
32.0
88.0
91.5
6.5
,0.01
3.0
550.0
248.0
104.0
110.0
a
The specificity was measured by the potency of CCK-related peptides to displace
relative to that of sulfated CCK-8 (100%).
b
Ab, antibody.
125
I-CCK-8(s) from the antisera. The potencies are expressed in percentages
996
Rehfeld: Plasma cholecystokinin
Fig. 2. Gel chromatography of acid and neutral extracts of porcine jejunal mucosa.
The extracts were applied to Sephadex G-50 superfine columns (1 3 100 cm) and eluted with 0.02 mol/L sodium veronal, pH 8.4, containing 1 g/L bovine serum
albumin. The chromatographic runs were monitored with the CCK-specific radioimmunoassay using antibody 92128 (Table 2) before (E) and after (F) incubation of each
fraction with trypsin. The elution positions of known molecular forms of CCK are indicated by arrows. Similar chromatographic patterns were obtained with extracts of
human jejunal mucosa (data not shown).
subjects did not increase CCK concentrations in plasma.
On the contrary, gastrin induced a decrease in plasma
CCK (Fig. 5). Stripping of the antiserum showed that CCK
circulates mainly as CCK-22 and CCK-8 in rabbit 92128
(data not shown).
Plasma CCK concentrations in healthy young human
subjects in the basal state were 1.13 6 0.10 pmol/L
(mean 6 SE, n 5 26), whereas the peak concentration at 60
min (or later) after onset of the meal was 4.92 6 0.34
pmol/L (mean 6 SE, n 5 26; Fig. 6).
ria reliability
The CCK concentration two SD below the binding percentage at zero binding, (i.e., mean 6 SD, 33.5 6 0.4%, n 5 10),
was 0.1 pmol/L, which was defined as the detection limit of
the assay. According to Ekins and Newman (23) this detection limit corresponds to 50 amol CCK in the assay tube.
The intra- and interassay variation at different concentrations within the working range of the assay ranged
between 5% and 15% (Table 3).
Accuracy evaluations by the addition of CCK-peptides
to plasma, the dilution of plasma samples with high
concentrations of endogenous CCK, and the mixing
plasma samples with high and low concentrations of CCK
showed a high degree of correlation with deviations
,15% from the expected concentrations.
Discussion
This study has shown that it is possible to develop an
assay for the accurate quantitation of CCK in plasma. The
decisive step is the production of a high-affinity and
high-titer antiserum with negligible or no binding of the
closely related gastrins. Two features seemed critical for
such production. First, the size and structure of the
CCK-hapten is crucial. An octapeptide (like CCK-8) is too
small for the production of high-titer and high-affinity
antibodies with the desired specificity. All antibodies
against CCK-8 have thus far bound gastrin, because the
CCK-specific portion of CCK-8 is too small for the induction of specific antibodies (Fig. 1). Therefore, the poor
results with immunization against small CCK-peptides
Clinical Chemistry 44, No. 5, 1998
997
Fig. 3. Gel chromatography of acid (E) and neutral (F) extracts of basal and postprandial plasma from two healthy young human subjects (left and
right panels, respectively).
The concentrated extracts were applied to Sephadex G-50 superfine columns (1 3 100 cm) and eluted with 0.02 mol/L sodium veronal, pH 8.4, containing 1
g/L bovine serum albumin. The chromatographic runs were monitored with the CCK-specific radioimmunoassay using antibody 92128 (Ab. 92128) (Table 2). The
elution positions of known molecular forms of CCK are indicated by arrows. Note that CCK-58 like peptides were present in only modest amounts in both acid
and neutral plasma extracts.
led to the design of a CCK-12 analog corresponding to
O-sulfated CCK-10 extended at the N-terminus with a
diglycine bridge for carrier coupling–the idea being
that coupling at the N-terminus would inevitably produce antibodies directed towards the amidated Cterminus (common also for gastrin) but at the same
time expose the CCK-specific sequence 3–5 of CCK-10
(Asp-Tyr (SO32 )-Met-; see Fig. 1). In other words, this
hapten should offer a chance for the production of an
antibody specific for the entire bioactive CCK-site–the
sulfated heptapeptide amide sequence. For theoretical
reasons (10), however, such a chance should be small
because antibody combining sites usually have a size
corresponding to the epitopes of five or six small amino
acid residues. Therefore, the second precaution to be
taken was to immunize a larger number of animals to
increase the chance of obtaining an antiserum with the
desired specificity. This explains why series XI was
composed of 16 rabbits.
It may seem odd to describe the development of a RIA
for a gut hormone in the 1990s. Such studies were
generally reported in the 1970s and early 1980s (for a
998
Rehfeld: Plasma cholecystokinin
Fig. 4. The relationship between CCK and gastrin concentrations in plasma from 49 healthy human subjects.
The samples were obtained from randomized population study. The CCK concentrations were measured with the radioimmunoassay using antibody 92128 (ordinate);
the gastrin concentrations were measured with a RIA using antibody 2604 (abscissa).
review, see reference 33). But the problems in developing
a reliable CCK-RIA for plasma measurements have by far
exceeded those of other gut hormones and of peptide
hormones in general (4).
Since 1969, various reports have claimed to measure
CCK in plasma (6, 34 – 63). Meaningful measurements,
however, were not reported until the mid-1980s
(6, 47, 52, 53, 57– 62); most of the antisera used in these
assays suffer shortcomings. Some antisera bind gastrins
to such a degree (51, 54, 59), that apparent hypercholecystokininemia has been assumed to occur during
clinical hypergastrinemias and during infusion of gastrin. Thus, the phenomenon of gastrin-suppressed CCK
secretion (Fig. 5) has thus far escaped detection. Moreover, such antisera cannot be used in the search for
CCK-producing tumors. Another shortcoming has been
that too few bleedings, low titer, and/or suboptimal
assay technology have limited the amounts of antiserum available for distribution so that only a few
laboratories have successfully measured CCK in
plasma in a reliable manner.
Another shortcoming has been the variable extent to
which the molecular forms of CCK in plasma have been
measured. This again has contributed to the confusion
about the molecular nature of CCK in plasma
(5, 6, 26, 31, 42, 43, 48, 61, 63–71). The molecular pattern
reported thus far has varied considerably both in man
(6, 42, 43, 48, 61, 63– 67) and between mammalian species.
Hence, the evidence in favor of predominance of either
CCK-8 (42, 43) or CCK-58 (26, 64) is declining, and a
pattern of mixed occurrences of CCK-33, CCK-22, and
CCK-8 is emerging (Fig. 3). Even in acidified human
plasma, CCK-58 predominated neither in the basal state
nor after a meal (Fig. 3). In addition, the use of different
analytical principles in the earlier studies, such as RIAs
with different antisera, subtraction assays, combined RIA
and enzyme assays, and bioassays have contributed to the
confusion.
Clinical Chemistry 44, No. 5, 1998
999
As mentioned in the introduction, a single bioassay has
met the reliability demands for accurate plasma CCK
measurements (5, 6). However, the complexity, laborintensiveness, and costs for this bioassay measurements
have precluded broader use. Therefore, the present RIA
using antiserum 92128 should open avenues for new
studies of the old hormone, CCK.
Fig. 5. Effect of gastrin-17 infusion in three different doses on CCK
secretion in six healthy young human subjects.
The plasma CCK concentrations (mean 6 SE) were measured with the CCKspecific RIA using antibody 92128 (Fig. 2 and Table 2).
We gratefully acknowledge the skillful technical assistance of Alice von der Lieth, Rikke Grønholt Pedersen,
and Gitte Runge Hansen. We also gratefully acknowledge the kind help of Anders H. Johnsen in controlling
the amount and structure of synthetic peptides. We
appreciate the generous support with natural CCK
peptides from Viktor Mutt (Stockholm, Sweden) and
with synthetic CCK peptides from Ulf Ragnarsson
(Uppsala, Sweden) and Miguel Ondetti (Princeton, NJ)
in the early phases of the study. We also appreciate the
kind permission of Kurt Borch (Linköping, Sweden)
and Morten Wøjdemann (Copenhagen, Denmark) to
use some of their plasma samples for CCK measurements. Finally, I thank Cathrine Ørskov and Lea Palo-
Fig. 6. Concentrations of CCK in
plasma from healthy young human
subjects (n 5 26) 60 min before
and 125 min after ingestion of a
meal.
The meal was consumed in 10 min.
1000
Rehfeld: Plasma cholecystokinin
Table 3. Reproducibility of plasma CCK measurements
(pmol/L) using Aba 92128 (n 5 10).
Mean
Within-assay reproducibility
3.76
7.56
15.41
Between-assay reproducibility
3.79
7.45
15.32
a
SD
SEM
95% confidence
Intervals
0.32
0.56
2.03
0.10
0.18
0.64
3.52–3.99
7.16–7.96
13.96–16.86
0.43
0.71
1.40
0.14
0.24
0.47
3.42–4.16
6.84–8.07
14.27–16.38
16.
17.
18.
19.
Ab, antibody.
heimo for help with the application to the local ethical
committee for studies in human subjects. The study was
supported by grants from the Danish Medical Research
Council, the Danish Cancer Union, the Danish Biotechnology Program for Peptide Research, the Gangsted
Foundation, and the Vissing Foundation.
20.
21.
22.
23.
References
1. Rehfeld JF. Cholecystokinin. In: Makhlouf GM, ed. Handbook of
physiology: the gastrointestinal system. Vol. II. Neural and endocrine biology. Bethesda, MD: American Physiological Society
1989:337–58.
2. Liddle RA. Cholecystokinin. In: Walsh JH, Dockray GJ, eds. Gut
peptides. Comprehensive endocrinology, revised series. New
York: Raven Press 1994:175–216.
3. Wank SA. Cholecystokinin receptors. Am J Physiol 1995;269:
G628 – 46.
4. Rehfeld JF. How to measure cholecystokinin in plasma? Gastroenterology 1984;87:434 – 8.
5. Liddle RA, Goldfine ID, Williams JA. Bioassay of cholecystokinin in
rats: effects of food, trypsin inhibitor and alcohol. Gastroenterology 1984;87:542–9.
6. Liddle RA, Goldfine ID, Rosen MS, Taplitz RA, Williams JA.
Cholecystokinin bioactivity in human plasma. J Clin Invest 1985;
75:1144 –52.
7. Mutt V, Jorpes JE. Hormonal polypeptides of the upper intestine.
Biochem J 1971;125:57p– 8p.
8. Ondetti MA, Rubin B, Engel SL, Pluscec J, Sheehan JT. Cholecystokinin-pancreozymin: recent developments. Am J Dig Dis 1970;
15:149 –56.
9. Rehfeld JF. Immunochemical studies on cholecystokinin. I. Development of sequence-specific radioimmunoassays against porcine triacontratriapeptide cholecystokinin. J Biol Chem 1978;253:4016–21.
10. Schechter I. Mapping of the combining sites of antibodies specific
for poly-L-alanine determinants. Nature 1970;228:639 – 41.
11. Sankaran H, Deveney CW, Goldfine ID, Williams JA. Preparation of
biologically active radioiodinated cholecystokinin for radioreceptor
assay and radioimmunoassay. J Biol Chem 1979;254:9349 –51.
12. Jensen RT, Lemp GF, Gardner JD. Interaction of cholecystokinin
with specific membrane receptors on pancreatic acinar cells. Proc
Natl Acad Sci U S A 1980;77:2079 – 83.
13. Paloheimo LI, Rehfeld JF. A processing-independent assay of
human procholecystokinin and its products. Clin Chim Acta 1994;
229:49 – 65.
14. Rehfeld JF. Disturbed islet-cell function related to endogenous
gastrin release. J Clin Invest 1976;58:41–9.
15. Rehfeld JF, Stadil R, Rubin B. Production and evaluation of
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
antibodies for the radioimmunoassay of gastrin. Scand J Clin Lab
Invest 1972;30:221–32.
Hilsted L, Rehfeld JF. Measurement of precursors for a-amidated
hormones by radioimmunoassay of glycine-extended peptides
after trypsin-carboxypeptidase-B cleavage. Anal Biochem 1986;
152:119 –26.
Bardram L, Rehfeld JF. Processing-independent radioimmunoanalysis: a general analytical principle applied to progastrin and
its products. Anal Biochem 1988;175:537– 43.
Fahrenkrug J, Schaffalitzky de Muckadell OB, Rehfeld JF. Production and evaluation of antibodies for the radioimmunoassay of
secretin. Scand J Clin Lab Invest 1976;36:281–7.
Larsson LI, Fahrenkrug J, Schaffalitzky de Muckadell OB, Sundler
F, Håkanson R, Rehfeld JF. Localization of vasoactive intestinal
polypeptide (VIP) to central and peripheral neurons. Proc Natl
Acad Sci U S A 1976;73:3197–200.
Rehfeld JF. Sequence-specific radioimmunoassays for gastrin,
cholecystokinin and somatostatin. Biomed Res 1980;1:73– 8.
Duve H, Rehfeld JF, East P, Thorpe A. Localisation of sulfakinin
neuronal pathways in the blowfly Calliphora vomitoria. Cell Tissue
Res 1994;275:177– 86.
Scatchard G. The attraction of proteins for small molecules and
ions. Ann N Y Acad Sci 1949;51:660 –72.
Ekins R, Newman B. Theoretical aspects of saturation analysis.
Acta Endocrinol 1970;147:11–36.
Sips R. On the structure of a catalyst surface. J Chem Phys
1948;16:490 –5.
Rehfeld JF, Bardram L, Cantor P, Hilsted L, Johnsen AH. The
unique specificity of antibodies in modern radioimmunochemistry.
Scand J Clin Lab Invest 1989;49:41– 4.
Eysselein VE, Eberlein GA, Hesse WH, Singer MV, Goebell H,
Reeve JR Jr. Cholecystokinin-58 is the major circulating form of
cholecystokinin in canine blood. J Biol Chem 1987;262:214 –7.
Magistris Lde, Rehfeld JF. A simple enzymatic procedure for
radioimmunochemical quantitation of the large molecular forms of
gastrin and cholecystokinin. Anal Biochem 1980;102:126 –33.
Rehfeld JF. Immunochemical studies on cholecystokinin. II. Distribution and molecular heterogeneity in the central nervous
system and small intestine of man and hog. J Biol Chem 1978;
253:4022–30.
Larsson LI, Rehfeld JF. Localisation and molecular heterogeneity
of cholecystokinin in the central and peripheral nervous system.
Brain Res 1979;165:201–18.
Hökfelt T, Rehfeld JF, Skirboll L, Ivemark B, Goldstein M, Markey
K. Evidence for coexistence of dopamine and cholecystokinin in
mesolimbic neurons. Nature 1980;285:476 – 8.
Cantor P, Rehfeld JF. Cholecystokinin in pig plasma: release of
components devoid of a bioactive COOH-terminus. Am J Physiol
1989;256:G53– 61.
Blanke SE, Johnsen AH, Rehfeld JF. N-terminal fragments of
intestinal cholecystokinin: evidence for release of CCK-8 by cleavage on the carboxyl side of Arg-74 in proCCK. Regul Pept
1993;46:575– 82.
Rehfeld JF. Radioimmunoassay problems for gut hormones in the
eighties. Adv Metab Dis 1988;11:45– 67.
Young JD, Lazarus L, Chisholm DJ, Atkinson FFV. Radioimmunoassay of pancreozymin-cholecystokinin in human serum. Nucl Med
1969;10:743–5.
Harvey RF, Dowsett L, Hartog M, Read AE. A radioimmunoassay
for cholecystokinin-pancreozymin. Lancet 1973;II:826 – 8.
Reeder DD, Becker HD, Smith NJ, Rayford PL, Thompson JC.
Measurement of endogenous release of cholecystokinin by radioimmunoassay. Ann Surg 1973;1978:304 –10.
Thompson JC, Fender HR, Ramus NI, Villar HV, Rayford PL.
Clinical Chemistry 44, No. 5, 1998
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
Cholecystokinin metabolism in man and dogs. Ann Surg 1975;
182:496 –503.
Schlegel W, Raptis S, Grube D, Pfeiffer EF. Estimation of cholecystokinin-pancreozymin (CCK) in human plasma and tissue by a
specific radioimmunoassay and the immunohistochemical identification of pancreozymin-producing cells in the duodenum of
humans. Clin Chim Acta 1977;80:305–16.
Marshall CE, Egberts EH, Johnson AG. An improved method for
estimating cholecystokinin in human serum. J Endocrinol 1978;
79:17–27.
Byrnes DJ, Henderson L, Borody T, Rehfeld JF. Radioimmunoassay of cholecystokinin in human plasma. Clin Chim Acta 1981;
111:81–9.
Thompson JC, Fried GM, Ogden WD, Fagan CJ, Inoue K, Werner I,
Watson LC. Correlation between release of cholecystokinin and
contraction of the gallbladder in patients with gallstones. Ann Surg
1982;195:670 –5.
Calam J, Ellis A, Dockray GJ. Identification and measurement of
molecular variants of cholecystokinin in duodenal mucosa and
plasma. J Clin Invest 1982;69:218 –25.
Walsh JH, Lamers CB, Valenzuela JE. Cholecystokinin-octapeptide-like immunoreactivity in human plasma. Gastroenterology
1982;82:538 – 44.
Burhol PG, Jenssen TG, Lygren I, Schultz TB, Jorde R, Waldum H.
Iodination with iodo-gen and radioimmunoassay of cholecystokinin (CCK) in acidified plasma, CCK release and molecular CCK
components in man. Digestion 1982;23:156 – 68.
Maton PN, Selden AC, Chadwich VS. Large and small forms of
cholecystokinin in human plasma: measurement using high pressure liquid chromatography and radioimmunoassay. Regul Pept
1982;4:251– 60.
Chang TM, Chey WY. Radioimmunoassay of cholecystokinin. Dig
Dis Sci 1983;28:456 – 68.
Jansen JBMJ, Lamers CBHW. Radioimmunoassay of cholecystokinin in human tissue and plasma. Clin Chim Acta 1983;131:
305–16.
Jansen JBMJ, Lamers CBHW. Molecular forms of cholecystokinin
in human plasma during infusion of bombesin. Life Sci 1983;33:
2197–205.
Himeno S, Tarui S, Kanayama S, Kuroshima, T, Shinomura Y,
Hayashi C, et al. Plasma cholecystokinin responses after ingestion of liquid meal and intraduodenal fat, amino acids or hydrochloric acid in man: analysis with region-specific radioimmunoassay. Am J Gastroenterol 1983;78:703–7.
Izzo RS, Brugge WR, Praismann M. Immunoreactive cholecystokinin in human and rat plasma: correlation of pancreatic secretion
in response to CCK. Regul Pept 1984;9:21–34.
Becker HD, Werner M, Schafmayer A. Release of radioimmunologic cholecystokinin in human subjects. Am J Surg 1984;147:
124 – 8.
Miazza B, Palma R, Lachance JR, Chayvialle JA, Jonard PP,
Modigliani R. Jejunal secretory effects of intraduodenal food in
humans. Gastroenterology 1985;88:1215–22.
Slaff J, Wolfe MM, Toskes PP. Elevated fasting cholecystokinin
levels in pancreatic exocrine impairment: evidence to support
feedback regulation. J Lab Clin Med 1985;105:282–5.
Cantor P. Evaluation of a radioimmunoassay for cholecystokinin in
human plasma. Scand J Clin Lab Invest 1986;46:213–21.
1001
55. Isaacs PET, Ladas S, Forgacs IC, Dowling RH, Ellam SV, Adrian TE,
Bloom SR. Comparison of effects of ingested medium- and
long-chain triglycerides on gallbladder volume and release of
cholecystokinin and other gut peptides. Dig Dis Sci 1987;32:
481– 6.
56. Ohgo S, Takemura J, Oki Y, Nishizono F, Ishikawa E, Yoshimi T, et
al. Radioimmunoassay of cholecystokinin in plasma. Clin Chem
1988;34:1579 – 84.
57. Philipp E, Pirke KM, Kellner MB, Krieg JC. Disturbed cholecystokinin secretion in patients with eating disorders. Life Sci 1991;
48:2443–50.
58. Beardshall K, Deprez P, Playford RJ, Alexander M, Calam J. Effect
of chymotrypsin on human cholecystokinin release: use of clostripain in the validation of a new radioimmunoassay. Regul Pept
1992;40:1–12.
59. Höcker M, Schmidt WE, Creutzfeldt W, Choudhury AR, Nustede R,
Schafmayer A, Fölsch UR. Determination of plasma cholecystokinin (CCK) concentrations by bioassay and radioimmunoassay in
man. A critical evaluation. Regul Pept 1992;37:255– 69.
60. French SJ, Murray B, Rumsey RDE, Sepple CP, Read NW. Preliminary studies on the gastrointestinal responses to fatty meals in
obese people. Int J Obesity 1993;17:295–300.
61. Ballinger A, McLoughlin L, Medbak S, Clark M. Cholecystokinin is
a satiety hormone in humans at physiological post-prandial
plasma concentrations. Clin Sci 1995;89:375– 81.
62. Paloheimo LI, Rehfeld JF. Quantitation of procholecystokinin and
its products in plasma by processing-independent analysis. Clin
Chim Acta 1995;238:21–33.
63. Gaisano HY, Reilly W, Go VLW, Olivero D, Miller L J. Large
molecular forms of cholecystokinin circulating in humans. Pancreas 1986;1:148 –53.
64. Cantor P, Rehfeld JF. The molecular nature of cholecystokinin in
human plasma. Clin Chim Acta 1987;168:153– 8.
65. Eberlein GA, Eysselein VE, Hesse WH, Goebell H, Schaeffer M,
Reeve JR Jr. Detection of cholecystokinin-58 in human blood by
inhibition of degradation. Am J Physiol 1987;253:G477– 82.
66. Jansen JBMJ, Lamers CBHW. Molecular forms of cholecystokinin
in plasma from normal and gastrectomized human subjects
following a fat meal. Peptides 1987;8:801–5.
67. Kanayama S, Himeno S, Higashimoto Y, Yamasaki Y, Kitani T,
Seiichiro T. Plasma cholecystokinin-octapeptide like immunoreactivity in patients with hepatic cirrhosis. Life Sci 1987;41:1915–
20.
68. Rehfeld JF, Holst JJ, Jensen SL. The molecular nature of vascularly
released cholecystokinin from the isolated, perfused porcine
duodenum. Regul Pept 1982;3:15–28.
69. Rehfeld JF. The molecular nature of cholecystokinin in plasma: an
in vivo immunosorption study in rabbits. Scand J Gastroenterol
1994;29:110 –21.
70. Fölsch UR, Cantor P, Wilms H, Schafmayer A, Becker HD,
Creutzfeldt W. The role of cholecystokinin as a regulator of the
feedback control of pancreatic enzyme secretion in conscious
rats. Gastroenterology 1987;92:449 –58.
71. Lindén A, Carlquist M, Hansen S, Uvnäs-Moberg K. Plasma
concentrations of cholecystokinin, CCK-8, and CCK-33/39 in rats
determined by a method based on enzyme digestion of gastrin
before HPLC and RIA detection of CCK. Gut 1989;30:213–22.