Cholecystokinin peptides in cerebrospinal fluid: a study in healthy male subjects

Regulatory Peptides 68 (1997) 57–61

Cholecystokinin peptides in cerebrospinal fluid: a study in healthy male subjects a, a b c b ¨ Tove Gunnarsson *, Thomas Eklundh , Mats Eriksson , G Ali Qureshi , Stefan Sjoberg , Conny Nordin d a

Department of Clinical Neuroscience and Family Medicine, Division of Psychiatry, Karolinska Institute, Huddinge Hospital, S-141 86 Huddinge, Sweden b Department of Internal Medicine, Karolinska Institute, Huddinge Hospital, S-141 86 Huddinge, Sweden c Clinical Research Center, Novum, Karolinska Institute, Huddinge Hospital, S-141 86 Huddinge, Sweden d ¨ Department of Psychiatry, University Hospital, S-581 85 Linkoping , Sweden Received 17 June 1996; revised 13 November 1996; accepted 13 November 1996

Abstract The clinical reliability of measuring cholecystokinin (CCK) peptides in the cerebrospinal fluid (CSF) has not been fully elucidated. Therefore, we have assayed CCK-8S and CCK-4 in CSF obtained from 14 healthy male subjects, lumbar-punctured at the L4–5 level following a strictly standardised procedure. CSF concentrations of free CCK-8S and free CCK-4 were used as dependent variables while age, height, body weight, atmospheric pressure and some other factors served as independent variables. It was shown that the CCK-8S ratio between the second (7–12 ml) and first (0–6 ml) CSF fractions, correlated significantly with the atmospheric pressure at the time of puncture. Neither CCK-8S nor CCK-4 displayed concentration gradients in CSF. The CCK-4 levels, expressed as pmol l 21 in the total amount of CSF were found to be positively correlated with the neuraxis distance in the lying position and negatively with the neuraxis distance in the sitting position. Furthermore, CCK-4, expressed as pmol l 21 per min of tapping-time (pmol l 21 min 21 ), showed a negative correlation with storage time, presumably mirroring a proteolytic process. CCK-8S and CCK-4 intercorrelated positively independently of whether expressed as pmol l 21 or pmol l 21 min 21 . In conclusion, the results of this exploratory study indicate that the neuraxis distance (in the sitting and lying positions) and storage-time have to be accounted for when interpreting data on CSF levels of CCK-4. Attention has to be paid to the potential influence of atmospheric pressure on the concentration ratio of CCK-8S.  1997 Elsevier Science B.V. Keywords: CCK-4; CCK-8S; Atmospheric pressure; Storage-time; Tapping-time

1. Introduction In recent years, the role of cholecystokinin (CCK) has been a focal point of increasing interest in experimental psychiatric research. In particular, CCK has been characterised as a neurotransmitter and a neuromodulator with specific action on noradrenergic, opioid and dopaminergic systems [1]. The CCK octapeptide (CCK-8) is the most prevalent form in the central nervous system (ClJS) [1]. A potential

*Corresponding author. Tel.: 1 46 8 7464605; Fax: 1 46 8 7795416. 0167-0115 / 97 / $17.00  1997 Elsevier Science B.V. All rights reserved PII S0167-0115( 96 )02104-0

role of CCK-8 has been assumed in the regulation of pain [2] and satiety [3]. Another CCK peptide worthy of attention is the tetrapeptide CCK-4, which is related experimentally to anxiety [4]. Intravenous administration of CCK-4 provokes panic attacks in patients with panic disorder (PD) [5] and in females suffering from premenstrual dysphoric disorder [6]. Furthermore, CCK-4 induces panic attacks in healthy subjects [7], but with a lower frequency than in patients predisposed to anxiety [5]. The anxiogenic properties of CCK-4 are blocked to some extent by oral CCK-B receptor antagonists [8,9]. The presence of CCK peptides in the cerebrospinal fluid

58

T. Gunnarsson et al. / Regulatory Peptides 68 (1997) 57 – 61

(CSF) was reported in 1978 by Rehfeld and Kruse-Larsen [10]. In subsequent clinical experiments, the CCK levels in CSF were found to be decreased in schizophrenic males, compared with females [11]. Lower concentrations of CCK-8 in the CSF have been reported in patients with PD [12] and in patients suffering from bulimia nervosa [13]. In 1993, Lydiard et al. [13] addressed the intriguing issue of factors influencing CCK concentrations in CSF. Their findings showed no influence of age and body weight on CCK-8 levels. Apart from that, the putative influence af other factors that are known to be of importance for monoamine metabolite levels, e.g. height, gender, length of the spine, tapping-time and atmospheric pressure (AP) [14–18] have not been evaluated as far as we know. The existence of CCK-8 and CCK-4 (and some other forms of CCK) in the human brain has recently been ¨ described by Lofberg et al. [19]. In line with that finding, we have now investigated the CSF concentrations of the sulphated CCK-8 (CCK-8S), which is the most abundant form of CCK-8 [20,21], and CCK-4 in healthy male subjects. The study design is exploratory and hypothesesgenerating.

7 and CCK-8 were used in the analyses. None of them interfered with the quantification of CCK-4 and CCK8S. Data on atmospheric pressure in the southern region of Stockholm were obtained from the meteorological station at the naval helicopter base at Berga, located 20 km southeast of the Huddinge Hospital. Monoamine transmitters and transmitter metabolites have also been analysed and the results have been reported separately [23]. The STATVIEW II (Abacus Concepts) and Statistica ( STATSOFT ) programmes were used. Parametric statistics were employed according to Kleinbaum et al. [24]. In repeated step-wise forward regression analyses (F to enter value set at 4.75), the CSF concentrations (expressed in pmol l 21 and pmol l 21 min 21 ) of CCK-8S and CCK-4 in the standardised amount of 12 ml [14] were used in turn as dependent variables, while age, body weight, height, tapping-time, AP, storage-time (2.46s.d 1.0 months), and the neuraxis distance (from the external occipital protuberance to the site of puncture) in the lying (ND L ) and sitting (ND S ) positions were used as regressors. When analysing data from the 6-ml fractions, non-parametric statistics [25] were used owing to deviations from normality. The study was approved by the Ethics Committee of the Huddinge University Hospital.

2. Subjects and methods The subjects comprised 14 males (aged 326S.D. 7, range 22–45 years) who were accepted after giving their informed consent. They were recruited among hospital staff members and their relatives. Exclusion criteria were evidence of alcohol or drug abuse and neurological, cardiovascular, hepatic, renal, haematopoietic, gastrointestinal or metabolic dysfunction. At the time of lumbar puncture, all were medication-free and healthy according to the history, physical examination and blood laboratory tests. All lumbar punctures were performed using a standardised technique [14] at the L 4–5 level at 8 a.m. after a minimum of 8 h of fasting along with strict bed rest. With the subject in a sitting position, CSF was drawn in two consecutive 6-ml fractions with a disposable needle (Becton-Dickinson 0.70 3 75 mm), from which the CSF was allowed to drip into a test-tube. Tapping-time was recorded using a stop-watch. The aliquots of 6-ml CSF were immediately protected from light, centrifuged and stored at 2 708C until analysis. For measurement of free CCK-4 and CCK-8S, the chemical identity of these peptides was ensured using a novel high-pressure liquid chromatography (HPLC) system for micropurification (SMART) and subsequent fast atom bombardment mass spectrometry [22]. The between-assay coefficient of variation (C.V.) was less than 3% for both peptides. The within-assay C.V. was 4% and 6% for CCK-4 and CCK-8S and the detection limit was 0.1 pmol l 21 and 0.5 pmol l 21 for CCK-4 and CCK8S, respectively (G.A. Qureshi, unpublished data). Commercially available CCK fragments such as CCK-5, CCK-

3. Results Clinical data are presented in Table 1, and concentrations of CCK-8S and CCK-4 in various fractions are shown in Table 2. In the first session, concentrations of CCK-8S and CCK-4 from the first (0–6 ml) and the second (7–12 ml) fractions were compared. Using the Wilcoxon signed rank test, we found no difference between fractions either when expressing the levels in pM (z CCK-8S 5 2 1.16; N.S; z CCK-4 5 2 0.20; N.S.) or when using pmol l 21 min 21 (z CCK-8S 5 2 0.41; N.S; z CCK-4 5 2 0.91; N.S.) In the second session, the CCK-8S and CCK-4 concentrations in the standardised amount of 12 ml (expressed Table 1 Clinical data on 14 male volunteers and experimental conditions Mean6S.D. Clinical data Age (years) Height (cm) Neuraxis distance sitting (cm) Neuraxis distance lying (cm) Body weight (kg) Experimental conditions Atmospheric pressure (hPa) Tapping-time (min) I II I 1 II

32.266.8 178.965.8 68.064.2 63.662.8 80.268.3 1013.668.7 4.561.8 4.561.7 9.063.1

T. Gunnarsson et al. / Regulatory Peptides 68 (1997) 57 – 61

59

Table 2 Concentrations of CCK-8S and CCK-4 in CSF Fraction

Concentration (mean6S.D.) CCK-8S

CCK-4

20.765.2 21.465.3 21.164.0

2.260.6 2.360.8 2.260.6

5.462.8 5.562.8 2.661.2

0.660.3 0.660.4 0.360.1

21

(pmol l ) 0–6 ml 7–12 ml 0–12 ml (pmol l 21 min 21 ) 0–6 ml 7–12 ml 0–12 ml

in pmol l 21 and pmol l 21 min 21 ) were used in repeated step-wise forward regression analyses. Expressed in pmol l 21 , the distances in the sitting and the lying positions made significant contributions to the variance in CCK-4 (Fig. 1). Using pmol l 21 min 21 , storage-time correlated negatively with CCK-4 (Fig. 2). No relationships were found whatsoever for CCK-8S. We also used the concentration ratio between fractions II and I. Then, CCK-8S correlated significantly with atmospheric pressure (Fig. 3). When investigating CCK-8S in the two 6-ml fractions, we found a correlation with AP in the first (r s 5 0.64; p 5 0.0205) but not the second (r s 5 0.0; N.S) fraction. In the third session, positive correlations between CCK4 and CCK-8S were found in the first (0–6 ml) CSF fraction, both when using pmol l 21 and pmol l 21 min 21 (Figs. 4 and 5). We also correlated concentrations of CCK-4 and CCK8S with monoamine precursors and transmitters and transmitter metabolites, as previously reported [23]. No correlations were found at a Bonferroni adjusted p , 0.05 significance level.

Fig. 2. Correlation between CCK-4 in 12 ml of CSF (pmol l 21 min 21 ) and storage-time (months) (r 5 0.57; p , 0.05; y 5 0.44 2 0.07x).

Fig. 3. Correlation between the concentration ratio of CCK-8S in CSF fractions II (7–12 ml) and I (0–6 ml) and atmospheric pressure (hPa) (r 5 0.59; p , 0.05; y 5 22.54 2 0.02x).

Fig. 4. Correlation between CCK-4 (pmol l 21 ) and CCK-8S (pmol l 21 ) in the first (0–6 ml) CSF fraction (r 5 0.56; p 5 0.05; y 5 0.74 1 0.07x).

Fig. 1. Plot of multiple regression including CCK-4 in 12 ml of CSF (pmol l 21 ) as the dependent variable ( y) and the neuraxis distance sitting (x 1 ) and distance lying (x 2 ) as regressors (R 5 0.75; F2:11 5 7.03; p , 0.05; y 5 4.98 2 0.13x 1 1 0.11x 2 )

4. Discussion Previous research on CCK in the CSF has mainly

60

T. Gunnarsson et al. / Regulatory Peptides 68 (1997) 57 – 61

Fig. 5. Correlation between CCK-4 (pmol l 21 min 21 ) and CCK-8S (pmol min 21 ) in the first (0–6 ml) CSF fraction (r 5 0.89; p 5 0.0001; y 5 0.05 1 0.09x).

focused on the octapeptide and larger forms of CCK. The results may be, at least in part, debatable since attention has not been paid to, e.g. height, gender, length of the spine, gradients, tapping-time and atmospheric pressure, factors that are known to be of importance for monoamine metabolite levels [14–18]. Setting out from animal experiments, there has been some discussion of the existence and role of CCK-4 in the CNS. In a study on pigs, this subject was addressed in 1986 by Rehfeld and Hansen [26] who concluded that it had to be demonstrated whether there are separate CCK-4 and CCK-5-synthesizing neurons, whether the small amounts of CCK-4 found represented a degradation of CCK-8 and / or CCK-5 or whether CCK-4 is a proper transmitter. Since then, the existence of CCK-4 in rat brain has been ¨ confirmed [22,27]. Interestingly, Lofberg et al. [19] recently reported that CCK-4 (as well as other forms of CCK) occurs in the human brain. Whether concentrations of CCK (and other compounds) in lumbar CSF really reflect the corresponding levels in the brain is an intriguing question. To our knowledge, our study is the first to systematically investigate the influence of various factors on the CSF levels of CCK-8S and CCK-4. We found that neither CCK-8S nor CCK-4 displayed clearcut concentration gradients on comparing the first and second fractions in the CSF. However, this statement is a qualified truth since the concentration ratio (fraction II / fraction I) of CCK-8S is correlated with AP (Fig. 3), implying that the presence of a gradient for CCK-8S might depend on the barometric pressure at the time of puncture. Just how AP exerts an effect is obscure but the finding that the impact on the concentration ratio is accounted for by a specific effect on the first (0–6 ml) fraction is notable. This might indicate that an AP above approximately 1017 hPa (Fig. 3) in some way reduces the elimination of CCK-8S from the lumbar compartment,

which is in line with previous suggestions of a lumbar cul-de-sac in which the CSF circulation is reduced or stagnated [28]. Concerning CCK-4, we found that ND L and ND S made significant contributions, but in opposite directions (Fig. 1). To explain this difference, we have to note that sitting up induces a movement of CSF from the brain to the lumbar region [29]. If we assume that this movement of CSF causes a dilution of CCK-4 in the lumbar compartment, the negative influence of ND S is then logical. The positive influence of ND L might then reflect a contribution of CCK-4 from the spinal cord and / or from the brain. This is in line with previous assumptions that CCK peptides are derived from the spinal cord [30] and with the fact that CCK-4 occurs in the human CNS [19]. Apart from the influence of ND L and ND S , we also note that CCK-4 (when expressed in nmol l 21 min 21 ) is influenced by storage-time. Our hypothesis is that, with increasing tapping-time, there is an increased activation of a proteolytic process that interacts with storage-time. Thus it might be necessary to pay attention to tapping-time as well as storage-time in future research on CCK-4 in CSF. Further evidence for the need to consider tapping-time is found in the fact that the significant correlation between CCK-4 and CCK-8S, expressed in pmol l 21 (Fig. 4) (confirming findings in human cortex [19]), is sharpened when expressing the concentrations in pmol l 21 min 21 (Fig. 5). Thus, our findings do not support previous suggestions by Lydiard et al. [12] that an increase in CCK-4 activity is compensated for by a decrease in CCK8S activity. In conclusion, the neuraxis distance (in the sitting and lying positions) and storage-time have to be accounted for when interpreting data on CCK-4 in the CSF of healthy male volunteers lumbar punctured in the sitting position after strict bedrest and fasting. For the concentration ratio of CCK-8S, a potential influence of atmospheric pressure has to be pointed out. Our unanimous results in a rather small number of subjects merit further study and elucidation.

Acknowledgments This study was supported by grants from the Ingrid and Fredrik Thuring Foundation, the Gadelius Foundation, the Trygg-Hansa Insurance Company and the Karolinska Institute. We thank Mr. Thomas Carlsson, assistant meteorologist, for providing data on atmospheric pressure ¨ and our research nurse, Mrs. Catharina Sjoberg, for her excellent assistance.

References [1] Albus M., Cholecystokinin, Prog. Neuropsychopharmacol. Biol. Psychiatry, 12 (1988) S5–S21.

T. Gunnarsson et al. / Regulatory Peptides 68 (1997) 57 – 61 [2] Pittaway K.M. and Hill R.G., Cholecystokinin and pain, Pain Headache, 9 (1987) 213–246. ´ A., Qureshi G.A. and Sodersten ¨ [3] Bednar I., Forsberg G., Linden P., Involvement of dopamine in inhibition of food intake by cholecystokinin octapeptide in male rats, J. Neuroendocrinol., 3 (1991) 491–496. [4] Bradwejn J. and Koszycki D., The cholecystokinin hypothesis of anxiety and panic disorder. In: J.R. Reeve Jr, V. Eysseleiti, T.E. Solomon and V.L.W. Go (Eds.), Cholecystokinin, The New York Academy of Sciences, New York, 1994, pp. 273–282. [5] Bradwejn J., Koszycki D. and Shriqui C., Enhanced sensitivity to cholecystokinin tetrapeptide in panic disorder, Arch. Gen. Psychiatry, 48 (1991) 603–610. [6] Le Melledo J.-M., Bradwejn J., Koscycki, D. and Bichet, D., Premenstrual dysphoric disorder and response to cholecystokinin tetrapeptide, Arch. Gen. Psychiatry, 52 (1995) 605–606. [7] de Montigny C., Cholecystokinin tetrapeptide induces panic-like attacks in healthy volunteers, Arch. Gen. Psychiatry, 46 (1989) 511–517. ¨ [8] Bradwejn J., Koszycki D., Couetoux du Tertre A., van Megen H., den Boer J., Westenberg H. and Annable L., The panicogenic effects of cholecystokinin–tetrapeptide are antagonized by L365 260, a central cholecystokinin receptor antagonist, in patients with panic disorder, Arch. Gen. Psychiatry, 51 (1994) 486–493. [9] Bradwejn J., Koszycki D., Paradis M., Reece P., Hinton J. and Sedman A., Effect of CI-988 on cholecystokinin tetrapeptide-induced panic symptoms in healthy volunteers, Biol. Psychiatry, 38 (1995) 742–746. [10] Rehfeld J.F. and Kruse-Larsen C., Gastrin and cholecystokinin in human cerebrospinal fluid. Immunochemical determination of concentrations and molecular heterogeneity, Brain Res., 155 (1978) 19–26. [11] Beinfeld M.C. and Garver D.L., Concentration of cholecystokinin in cerebrospinal fluid is decreased in psychosis: relationship to symptoms and drug response, Prog. Neuro-Psychpopharmacol. Biol. Psychiatry, 15 (1991) 601–609. [12] Lydiard R.B., Ballenger J.C., Laraia M.T., Fossey M.D. and Beinfeld M.C., CSF cholecystokinin concentrations in patients with panic disorder and in normal comparison subjects, Am. J. Psychiatry, 149 (1992) 691–693. [13] Lydiard R.B., Brewerton T.D., Fossey M.D., Laraia M.T., Stuart G., Beinfeld M.C. and Ballenger J.G., CSF cholecystokinin octapeptide in patients with bulimia nervosa and in normal comparison subjects, Am. J. Psychiatry, 150 (1993) 1099–1101. ˚ [14] Bertilsson L. and Asberg M., Amine metabolites in the cerebrospinal fluid as a measure of central neurotransmitter function: methodo˚ logical aspects. In: E. Usdin, M. Asberg, L. Bertilsson and F. ¨ Sjoqvist (Eds.), Frontiers in Biochemical and Pharmacological Research in Depression, Raven Press, New York, 1984, pp. 27–34. [15] Nordin C., Swedin A. and Zachau A., Tapping-time influences concentrations of 5-HIAA in the CSF, J. Psychiatr. Res., 27 (1993) 409–414.

61

¨ V. and Nordin C., Influence of tapping-time [16] Eklundh T., Fernstrom and atmospheric pressure on concentrations of monoamine metabolites in the cerebrospinal fluid — a prospective study in female volunteers, J. Psychiatr. Res., 28 (1994) 511–517. ¨ V., Swedin A. and Zachau A.C., [17] Nordin C., Eklundh T., Fernstrom Gradients of CSF monoamine metabolites: A comparison between male and female volunteers. J. Psychiatr. Res., 29 (1995) 133–140. ¨ L. and Wieselgren I.-M., Acid monoamine [18] Nordin C., Lindstrom metabolites in the CSF of healthy controls: A retrospective study, J. Psychiatr. Res., 30 (1996) 127–133. ¨ [19] Lofberg C., Harro J., Gottfries C-G. and Oreland L. Cholecystokinin peptides and receptor binding in Alzheimer’s disease, J. Neural Transm., 103 (1996) 851–860. [20] Sauter A. and Frick W., Determination of cholecystokinin tetrapeptide and cholecystokinin octapeptide sulfate in different rat brain regions by high-pressure liquid chromatography with electrochemical detection, Anal. Biochem., 133 (1983) 307–313. [21] Marley P.D., Rehfeld J.F. and Emson P.C., Distribution and chromatographic characterisation of gastrin and cholecystokinin in the rat central nervous system, J. Neurochem., 42 (1984) 1523–1535. ¨ [22] Qureshi G.A., Bednar I., Min Q., Sodersten P., Silberring J., Nyberg ¨ F. and Thornwall M., Quantitation and identification of two cholecystokinin peptides, CCK-4 and CCK-8S, in rat brain by HPLC and fast atom bombardment mass spectrometry, Biomed. Chromatogr., 7 (1993) 251–255. ¨ [23] Eklundh T., Eriksson M., Sjoberg S. and Nordin C., Monoamine precursors, transmitters and metabolites in cerebrospinal fluid. A prospective study in healthy male subjects, J. Psychiatr. Res., 30 (1996) 201–208. [24] Kleinbaum D.G., Kupper L.L. and Muller K.E., Applied Regression Analyses and Other Multivariable Methods, PWS-KENT Publishing Company, Boston, 1988. [25] Siegel S. and Castellan Jr. N.J., Nonparametric Statistics for the Behavioral Sciences, McGraw–Hill International editions, New York, 1988. [26] Rehfeld J.F. and Hansen H.F., Characterization of preprocholecystokinin products in the porcine cerebral cortex, J. Biol. Chem., 261 (1986) 5832–5840. ¨ [27] Pavlasevic S., Bednar I., Qureshi G.A. and Sodersten P., Brain cholecystokinin tetrapeptide levels are increased in a rat model of anxiety, NeuroReport, 5 (1993) 225–228. [28] Milhorat T. and Hammock M., Cerebrospinal fluid as reflection of internal milieu of brain. In: J. Wood (Ed.), Neurobiology of Cerebrospinal Fluid, Plenum Press, New York and London, 1983, pp. 1–24. [29] Magnaes B., Movement of cerebrospinal fluid within the craniospinal space when sitting up and lying down, Surg. Neurol., 10 (1978) 45–49. [30] Geracioti T.D., Nicholson W.E., Orth D.N., Ekhator N.N. and Loosen P.T. CSF cholecystokinin dynamics in the human. Abstract nr 505, Annual Meeting of the American Psychiatric Association, Miami, Florida, May 1995, p. 190.