- Email: [email protected]
Basal and hyperinsulinemia-induced immunoreactive hypothalamic insulin changes in lean and genetically obese Zucker rats revealed by microdialysis
258
Brain Research, 611 (1993) 258-263 ~'~ 1993 Elsevier Science Publishers B.V. All rights reserved 0006-8993/93/$06.00
BRES 18811
Basal and hyperinsulinemia-induced immunoreactive hypothalamic insulin changes in lean and genetically obese Zucker rats revealed by microdialysis K. Gerozissis, M. Orosco, C. Rouch and S. Nicolaidis Neurobiologie des Rdgulations, Collbge de France, Paris (France) (Accepted 8 December 1992)
Key words: lmmunoreactive insulin; Hypothalamus; Ventromedial nucleus; Paraventricular nucleus; Microdialysis; Genetically obese Zucker rat
Lean and genetically obese Zucker rats were implanted with permanent intravenous catheters and a guide canula was aimed at the region of the ventromedial (VMH) and paraventricular (PVN) nuclei to measure immunoreactive insulin collected by means of microdialysis. Preliminary experiments assessed the validity of a novel assay of insulin in microdialysates by a sensitized radioimmunoassay technique. This method was then used to measure basal levels of insulin and those induced by i.v. infusion of 0.5 U of insulin over 30 min in both lean and obese rats. Basal hypothalamic immunoreactive insulin levels were lower in the obese rats than in the lean Zucker rats. When insulin was infused i.v. for 30 min, hypothalamic immunoreactive insulin showed an increase in the 30-60 min sample, which was twice as great in the obese rats. Two facts suggest that the insulin found in the microdialysates was of cerebral, not vascular origin: the short latency in the response and the finding that the response was greater in obese rats.
INTRODUCTION
The role of insulin in the control of food intake and body weight attracts considerable interest. The first investigation showing the direct impact of insulin on central structures used slow and prolonged insulin infusion in the medial hypothalamus of the rat ~3. Such infusions result in hypophagia and, more importantly, in a permanent reduction of body weight t3. Subsequent studies using, hypothalamic and intracerebral infusions of insulin verified this phenomenon in both the rat and the baboon 12'29. The involvement of insulin in the regulation of body weight was also assessed by numerous investigations in the genetically obese homozygous Zucker rat (fa/fa) which shows a peripheral resistance to insulin z3 and whose food intake is suppressed less by centrally administered insulin 1°. Most of these approaches were mainly pharmacological and consisted in the observation of a behavioural effect after the administration of insulin. Physiological studies however, showed several abnormalities in the
obese Zucker rat, including hyperinsulinemia 32, a high insulin concentration in the cerebrospinal fluid but a low ratio of CSF to plasma insulin 2~. Furthermore, low hypothalamic insulin levels have been reported in both the obese and the lean heterozygous (Fa/fa) Zucker rat 2, which suggests the importance of cerebral insulin. The question of the origin of cerebral insulin has been difficult to answer so far. Some authors believe that peripheral insulin reaches the cerebrospinal fluid and possibly the hypothalamic structures 25'3°. Others believe that the blood-brain barrier prevents the entry of insulin into the brain ~'9. Some evidence suggests that mRNA for insulin is present in hypothalamic cells of rats and humans ~'~'3~ Discrepancies in the literature made it difficult to ascertain the physiological involvement of central insulin in the control of body weight and food intake. The pharmacological microinjection approach could not provide an answer. Neither did the reciprocal approach, measuring brain tissue insulin by ex vivo methods, because these assays do not allow dynamic correlations of CSF and peripheral insulin. The introduction of the in vivo technique of brain
Correspondence: M. Orosco, Neurobiologie des R6gulations, Coll~ge de France, 11, place Marcelin Berthelot,75231 Paris Cedex 05, France.
259 microdialysis solves some of these problems because this method allows the measurement of neurochemicals released in the brain extracellular space and therefore allows one to determine the time course of possible variations of cerebral insulin during peripheral insulin infusion. Because insulin has not previously been measured by means of brain microdialysis, it was necessary to measure the relation between peripherally infused insulin and hypothalamic tissue insulin. This correlation is important to know because in most studies, changes in CSF insulin were found to follow changes in plasma insulin 3°. In this investigation, we assayed variations in hypothalamic immunoreactive insulin in response to peripheral insulin infusion in obese and lean Zucker rats. Previous studies have found that the hypothalamus of the obese rat contains low concentrations of insulin-like material. The assay was performed in the feeding-related structures of the median hypothalamus including the ventromedial (VMH) and paraventricular (PVN) nuclei. This area is the site of the inhibitory action of insulin microinjections on feeding 13'14. MATERIALS
AND METHODS
Animals Female Zucker obese (fa-fa) or lean (Fa-Fa) rats, 16 weeks old, were used in this study. They were bred in the Laboratoire d'Endocrinologie, Universit6 Paris-Sud, Orsay, France. Age-paired Wistar rats were also used for comparison of basal levels and validation tests. U p o n arrival in the laboratory, they were housed individually in cylindrical Plexiglas cages designed to allow stressless chronic infusions in the unrestrained animals. T h e temperature of the room was maintained at 2 4 + I°C and the lights were turned on from 06.00 to 18.00 h. Food and water were available ad libitum.
Surgery for chronic hypothalamic microdialysis and simultaneous venous catheterization Rats were anesthetized with ketamine (Imalgbne, M6rieux, 150 m g / k g ) after pretreatment by a muscle relaxant, xylasine (Rompun, Bayer). T h e indwelling venous catheter was implanted according to a previously described technique 15. Briefly, a silicone rubber (Silastic) catheter, filled with viscous polyvinylpyrrolidone solution (20% w / v ) , was inserted in the right external jugular vein approximately 5 m m before it dives u n d e r the clavicle. W h e n pushed in as far as a collar fitted in a position defined by the rat's body weight, the tip arrives at the auricular cavity. T h e vein was tied around the catheter and held in place by further sutures. T h e other end of the catheter was pulled subcutaneously through a slit in the skin at the top of the skull. Once the venous catheter was secured, the animal was placed in a stereotaxic frame (Kopf Instruments). A guide cannula (Carnegie) for the microdialysis probe was aimed at the space between the PVN and the VMH. The coordinates of the guide tip were (according to the atlas of Paxinos and Watson 17) - 1.9 m m anterior, 0.5 m m lateral to and 7 m m ventral to bregma. The dialysis probe protruded 2 m m beyond the guide-tube reaching a point 9 m m ventral to bregma. T h e guide cannula, the venous catheter and a securing device (screw), were fixed to the skull with stainless-steel screws and dental cement. At least one week was allowed for postoperative recovery before the experiments began. During this period, the animals were accus-
tomed to the experimental conditions in their home cages. In order to accustom the rats to the conditions of the experiments, they were permanently connected to the infusion and dialysis system of catheters which were protected by a metal sheath and kept out of the animal's reach by m e a n s of a counterbalanced beam. This arrangement allowed normal movements. The animals showed normal sleep, feeding and body weight gain. All experiments were performed in the animal's own home cages.
Microdialysis procedure The microdialysis m e m b r a n e s (Carnegie) were 2 mm long, with a diameter of 0.5 m m and a 20,000 kDa molecular weight cut-off. According to our in vitro calibration test, the relative recovery of immunoreactive insulin was 3.7%. The perfusion was performed with a Ringer type solution containing 147 m M Na ÷, 2.3 m M Ca 2÷, 4 m M K ÷ and 155.6 m M C I with 0.1% bovine serum albumin added. A system of catheters connected to a two-way swivel 15 placed on the beam, allowed independent perfusion of fluid into the probe and sampling. The flow rate (2 ~ l / m i n ) allowed the collection of 60 /xl fractions every 30 min.
Experimental protocol Because the experiment was conducted in the home cage where the tubes were permanently connected to the rat, the only change on the day of experiment was just the insertion of the microdialysis probe through the guide and the i.v. infusion. The first samples were taken 5 h after insertion of the probe, the venous perfusion was set up by connecting through one channel of the two-way swivel a plastic tube to the head piece of the venous catheter and by switching on the pump. After collecting 4 baseline dialysis samples with aqueous vehicle (control), 0.5 U of insulin (Actrapid, Novo) in a volume of 0.5 ml was then infused intravenously over a 30 min period. Food was removed during the experiment in order not to interfere with the effect of exogenous insulin.
Analysis of immunoreactive insulin in the dialysates Immunoreactive insulin in the dialysates was measured by m e a n s of a radioimmunoassay (RIA), using rat insulin as reference (Novo Labs) and a commercial antiserum raised in guinea pigs, immunized with h u m a n insulin, cross-reacting at 100% with rat insulin. [IZsl]Porcine insulin was used as a tracer (CIS, France). Free insulin was separated from bound insulin by immunoprecipitation, using a second antibody and polyethylene glycol. The R I A was performed in two steps (24 h preincubation in the absence of tracer followed by 90 min incubation). After the last incubation, the tubes were centrifuged, the supernatant was discarded and the radioactivity of the precipitate (bound fraction) was measured in a gamma-counter. T h e tests for validation are exposed in the section results.
Histology At the end of each experiment, the animals were given an overdose of pentobarbital and cardiac perfusion of saline and then of 10% formol was performed. The brain was removed, further hardened in 10% formol, and then serial coronal sections were cut (60 /zm). Following glycerin staining, they were observed at low magnification by an observer unaware of the data obtained during the microdialysis experiments to check for the position of the dialysis probe (Fig. 1). The animals whose brain did not show a correct placement of the guide and the probe were discarded by the independent observer.
Statistics The mean basal levels (before insulin infusion) were calculated for each animal. As in all the experiments of this kind, the percentage of variation at the peak effect of insulin infusion, relative to the m e a n of the four baseline samples was calculated for each animal. T h e significance of insulin effect in each group was assessed by Student's paired t-test.
260 The results were then expressed as mean basal levels+S.E.M. (pg/ml) and as means of percentage variations+S.E.M, for each group of rats, obese or lean. The statistical significance between the two groups was calculated using Student's t-test for basal levels and for percentage of variations.
IMMUNOREACTIVE I N S U L I N (pglml)
2OO-
El
RESULTS
Improvement in RIA sensitivity After a preincubation of 24 h, followed by a 30 min incubation in a reaction volume of 200/~1 containing 50 /xl of Ringer solution, the sensitivity of the R I A was improved tenfold (8.25 p g / t u b e and 100 p g / t u b e for 50% displacement, in sensitized and normal RIA, respectively), with a detection limit of 0.75pg/tube. The intra-assay coefficient was 4.5%. Diluted rat plasma insulin was quantified simultaneously with the one-step and the two-step method. Similar values were obtained (60, 31, 40 and 60, 30, 35 p g / m l , respectively). Several dilutions of rat plasma were tested and the values obtained were similar (288 + 16 p g / m l , mean + S.E.M.). We have followed the kinetics of immunoreactive insulin in the perfusates. As previous studies have shown for most substances 24, insulin concentration, high during the first period, probably because initial lesion of the tissue, then fell and remained relatively stable after a 4-h lag (Fig. 2). In a parallel study, a K + concentration (56 mM) known to induce a peptide release provoked a fivefold increase in insulin level (Fig. 2)
Basal levels The mean basal level of hypothalamic immunoreactive insulin was 63.9 + 7.8 p g / m l in lean rats (n = 8 determinations from 4 animals) and 37.8 ___4.7 p g / m l
S0
0
[ 90
I
150
I
[
I
I
210
270
330
390
TIME ( m i n )
Fig. 2. Immunoreactive insulin (pg/ml) in median hypothalamic microdialysates during a 7-h period, Normal Ringer-BSA solution was replaced by 56 mM K + Ringer-BSA solution 340 min after insertion of the probe.
in obese rats (n = 10 determinations from 4 animals). This difference was statistically significant (t = 2.86, P < 0.05). Wistar rats (n = 28 determinations from 6 animals) gave values similar to lean rats (57.5 _+ 0.4 p g / m l )
Effect of insulin infusion Intravenous insulin increased hypothalamic dialysate immunoreactive insulin by 60.6 + 14.8% in the 30 to 60 min period following the onset of the infusion in lean rats. However, this increase did not reach statistical significance. ~Hypothalamic immunoreactive insulin increased by 246.2 + 63% in the 30 to 60 min period following the onset of the infusion in obese rats. This increase was statistically significant (t = 3.54, P < 0.05). This i.v. insulin-induced enhancement of hypothalamic immunoreactive insulin was thus much larger in obese than in lean rats. The difference in percent increases between the two groups was statistically significant (t = 2.84, P < 0.05) (Fig. 3). DISCUSSION
Fig. 1. Coronal section at the level of anterior hypothalamus depicting the location of the microdialysis probe aimed at the space beween the paraventricular nucleus (PVN) and the ventromedian hypothalamus (VMH).
This study is the first in our knowledge to measure changes in concentrations of immunoreactive insulin in brain, using microdialysis. This measurement was accomplished by means of an improved method of RIA. The tenfold increase in R I A sensitivity was obtained using a classic modification of sequential incubation of ligand and tracer 7'28. The improved R I A method was validated by means of several tests, including stimulation tests, dilution studies and comparison with values obtained by a classic one-step R I A procedure. Some of the above tests were performed with plasma instead of
261 % CHANGES IN IMMUNOREACTIVE INSULIN / BASAL LEVELS
I
I 300 ~
200 -
lO0 -
o 30 min samples infusion LEAN
infusion OBESE
Fig. 3. Changes in immunoreactive insulin ( % / b a s a l levels) in median hypothalamic microdialysates induced by peripheral insulin infusion in lean and obese Zucker rats. Basal levels are given in the text. * P < 0.05 difference between insulin effects in lean and obese rats
microdialysates. Obviously, unlike plasma, microdialysates contain little, if any, nonspecific antigen-like or antibody-like material that could interfere with the specific ligand (insulin)-antibody binding. In fact, one of the major advantages of the microdialysis technique is that the membrane prevents large molecules from diffusing into the perfusate, keeps the samples purified, and protects molecules from being broken down by enzymes once they are recovered from the tissue. Consequently, the chemical analysis can usually be performed without clean-up steps 27 and avoids problems related to the loss of material entailed by purification procedures. However, as the very low insulin concentrations does not permit a complete series of validation and identification tests, we are using the term 'immunoreactive insulin' in the present study. Beyond the validation of RIA, care was taken also in this study to control the experimental conditions of the microdialysis technique. As previous studies have found for other cerebral substances 24, a progressive decline followed by stable basal levels was also observed in the case of insulin. Because 4 h were necessary for the plateau to occur, the determination of the release of immunoreactive insulin was started 5 h after the insertion of the microdialysis probe. The administration of high concentrations of K + via counter-dialysis several hours after the begining of microdialysis triggered a clear-cut peak of insulin response, which further suggests that the system was effective. The presence of bovine serum albumin in the perfusate overcomes the problems of adsorption of insulin by the probes and the tubes. The in vitro relative recovery was 3.7%, which corresponds to that obtained with similar high molecular weight substances H. It is
known that an estimate of the actual concentration of the substance of interest (insulin) in the extracellular fluid can be performed by calculating the relative recovery. Because we are aware that in vitro recoveries may not precisely reflect in vivo conditions e7, the resuits are presented only as the amounts of substance that are recovered in the fluid inside the dialysis probe. Three main results were obtained in this study. First, in lean Zucker rats as well as in genetically obese Zucker rats, i.v. insulin infusion produced a rapid elevation of the basal level of immunoreactive hypothalamic insulin. This elevation appeared in the second fraction (30 to 60 min) collected after the onset of the infusion. Dynamic assays of cerebrospinal fluid insulin have previously suggested that peripheral insulin could reach the brain tissue via its penetration in the cerebroventricular space following a dumped and delayed profile 3°. It was also suggested that insulin delivery to the neuropil may be due to a specialized brain transport system (transcytosis)~6'19" Alternatively, brain insulin may be produced locally as suggested by findings on expression of insulin mRNA in fetal but also adult brain cells localized precisely in the periventricular hypothalamuss'~s'31. However, other investigators, using the polymerase chain reaction method did not verify the notion of endogenous production of insulin in the hypothalamic tissuesS.The present data suggest (but do not prove) that peripheral insulin passes rapidly into the hypothalamus, or perhaps the peptide is even produced there, as some authors suggest sAs'31. The second result deals with the basal level of hypothalamic immunoreactive insulin, which is significantly higher in the lean than in the genetically obese rat. This observation confirms and extends a previous report of such a difference in basal insulin content of tissue homogenates of the entire hypothalamus2. Although we lack comparative measurements between various structures, our data suggest that the lower levels of insulin found in the VMH and PVN are responsible for the impaired regulation of food intake and body weight of the obese rat. The third result shows that, despite lower basal hypothalamic levels in the obese rat, the increase in local insulin in response to a peripheral load is significantly more pronounced. Again this differential increase is observed in the hypothalamic region that has been proposed to be the central site of action at which insulin acts in its role in body weight regulation and adjusment of feeding. Since insulin in the VMH and PVN region seems to inhibit feeding, its large changes in the obese rat may account for the lengthening of the intermeal interval in these hyperphagic animals 3. Taking these results into account, it is possible to
262
speculate on the important question of whether the insulin found in the hypothalamus was synthesized locally or whether it was derived from the insulin that was injected peripherally. In the latter case, the classically proposed mechanism would be an uptake from the plasma to the CSF and then a diffusion to various brain structures, including the hypothalamus 3°. Another transport mechanism, from the plasma to the brain parenchyma, either actively by means of transcytosis 16'19 or via fenestrated capillaries t'4'26 and subsequently to the CSF, has also been proposed. Our results in the lean rats are consistent with the hypothesis that the hypothalamic insulin is of peripheral origin. Although other investigators have suggested that the uptake of insulin from plasma to the CSF is slower in the obese rat 22, we observed a larger increase in the obese rat. If the hypothalamic insulin of obese rats is of peripheral origin, then the diffusion from the CSF to the brain parenchyma must be accelerated in these animals. According to the present knowledge, less insulin should have passed through the insulin resistant blood-CSF parenchyma barriers of the obese rat than through the normal barriers of the lean rat 2°. In conclusion, the use of the microdialysis technique for collecting samples of extracellular fluid from the brain for insulin assays in the genetically obese Zucker rat provides a new approach to the investigation of the origin of cerebral insulin and its possible role in obesity. Our data favour the possibility of a local cerebral insulin production that can respond to a peripheral signal such as an intravenous insulin infusion. More experimental manipulations and shorter sample collection intervals are needed before we can understand the role of insulin in the hypothalamus. However, these procedural changes will require more sensitive assays of insulin. But the use of microdialysis in freely moving rats bearing intravenous catheters is already capable of contributing in our understanding of the central mechanisms of peripheral peptides that act in parallel on distinct brain structures and induce regulatory or pathological processes. Acknowledgements. The authors would like to thank Marie-Jo Meile for histological preparation, Simon Thornton for the English, Neil Carlson for editorial improvement and Daniel Gripois and MarieFrance Blouquit for the gift of Zucker rats. This work was supported by grants MRT (Minist~re de la Recherche et de la Technologie) and BSN with the contribution of INRA (Institut National de Recheche Agronomique).
REFERENCES 1 Albrecht, J., Wroblewska, B. and Mossakovski, M.J., The binding of insulin to cerebral capillaries and astrocytes of the rat, Neurochem. Res., 7 (1982) 489-494.
2 Baskin, D.G., Stein, L.J., lkeda, H., Woods, S.C., Figlewicz, D.P., Porte, D.Jr., Greenwood, M.R.C. and Dorsa, D.M., Genetically obese Zucker rats have abnormally low brain insulin content, Lift, Sci.,36 (1985) 627-633. 3 Becket, E.E. and Grinker, J.A., Meal patterns in the genetically obese Zucker rat. Physiol. Beha~,., 18 (1977) 685-692. 4 Catalan, R.E., Martinez, A.M., Aragones, M.D., Miguel, B.G. and Robles, A., Insulin action on brain microvessels: Effect on alkaline phosphatase, Biochem. Biophys. Res. Commun., 150 (1988) 583-590. 5 Coker, G.T.I., Studelska, D., Harmon, S., Burke, W. and O'Malley, K.L., Analysis of tyrosine hydroxylase and insulin transcripts in human neuroendocrine tissues. MoL Brain Res., 8 (1990) 93-98. 6 Eng, J. and Yalow, R., Evidence against extrapancreatic insulin synthesis, Proc.Natl. Acad. Sci. USA, 78 (1981) 4576-4578. 7 Ezan, E., Mamas, S., Rougeot, C. and Dray, F., Strategies for developing specific and sensitive hapten radioimmunoassay. In C. Patrono and B.A. Peskar (Eds.), Handbook of Experimental Pharmacology, Springer, Berlin, 1987, pp. 143-179. 8 Giddins, S., Chirgwin, J. and Permutt, M., Evaluation of rat insulin messenger RNA in pancreatic and extrapancreatic tissues, Diabetologia, 27 (1985) 343-347. 9 Havrankova, J.M., Brownstein, M. and Roth, J., Insulin and insulin receptors in rodent brains, Diabetologia Suppl., 20 (1981) 268-273. 1(1 Ikeda, H., West,D.B., Pustek, J.J., Figlewicz, D.P., Greenwood, M.R.C., Porte, D.Jr. and Woods, S.C., Intraventricular insulin reduces food intake and body weight of lean but not obese Zucker rats, Appetite, 7 (19861 381-386. 11 Kendrick, K.M., Use of microdialysis in neuroendocrinology, Methods Enzymol., 168 (1989) 182-205 12 Mc Gowan, M.K., Andrews, K.M., Kelly, J. and Grossman, S.P., Effects of chronic intrahypothalamic infusion of insulin on food intake and diurnal meal patterning in the rat, Behat~. Neurosci., 104 (1990) 371-383. 13 Nicolffidis, S., M6canisme nerveux de l'~quilibre t~nerg6tique, Journ~es Annuelles de Diabetologie de l'Hotel Dieu de Paris, 1 (1978) 152-156. 14 Nicolffidis, S., Lateral hypothalamic control of metabolic factors related to feeding, Diabetologia, 20 (1981) 426-434. 15 Nicolffidis, S., Rowland, N., Meile, M.J., Marfaing-Jallaz, P. and Pesez, A., A flexible technique for long-term infusions in unrestrained rats, Pharmacol. Biochem. Behaz~., 2 (1974) 131-136. 16 Pardridge, W.M., Receptor-mediated peptide transport through the blood-brain barrier, Endocrinol. Reu., 7 (1986) 314-330. 17 Paxinos, G. and Watson, C., The rat brain in stereotaxic coordinates, Academic Press, New York, 1986. 18 Schechter, R., Sadiq, H.F. and Devaskar, S.U., Insulin and insulin mRNA are detected in neuronal cell cultures maintained in an insulin-free/serum-free medium. J. Histochem. Cytochem., 38 (1990) 829-836. 19 Schwartz, M.W., Bergman, R.N., Kahn, S.E., Taborsky, G.J., Fisher, L.D., Sipols, A.J., Woods, S.C., Steil, G.M. and Porte, D.J., Evidence for entry of plasma insulin into cerebrospinal fluid through an intermediate compartment in dogs. Quantitative aspects and implications for transport, Z Clin. Ira,est., 88 (1991) 1272-1281. 20 Schwartz, M.W., Figlewicz, D.P., Kahn, S.E., Baskin, D.G., Greenwood, M.R.C. and Porte, D.Jr., Insulin binding to brain capillaries is reduced in genetically obese, hyperinsulinemic Zucker rats, Peptides, 11 (1990) 467-472. 21 Stein, L.J., Dorsa, D.M., Baskin, D.G., Figlewicz, D.P., Ikeda, H., Frankman, S.P., Greenwood, M.R.C., Porte, D.Jr. and Woods, S.C., Immunoreactive insulin levels are elevated in the cerebrospinal fluid of genetically obese Zucker rats, Endocrinology, 113 (1983) 2299-2301. 22 Stein, L.J., Dorsa, D.M., Baskin, D.G., Figlewicz, D.P., Porte, D.Jr. and Woods, S.C., Reduced effect of experimental peripheral hyperinsulinemia to elevate cerebrospinal fluid insulin concentrations of obese Zucker rats, Endocrinology, 121 (1987) 1611-1615. 23 Stern, J.S., Johnson, P.R., Batchelor, B., Zucker, L. and Hirsch,
263
24
25
26 27
J., Pancreatic insulin release and peripheral tissue resistance in Zucker obese rats fed high and low carbohydrate diets, Am. J. PhysioL, 228 (1975) 543-548. Ungerstedt, U., Measurement of neurotransmitter release by intracranial dialysis. In C.A. Marsden (Ed.), Measurement of Neurotransmitter Release In l/'ivo, Wiley, 1984, pp. 81-105 Vanderweele, D.A., Pi-Sunyer, F.X., Novin, D. and Bush, M.J., Chronic insulin infusion suppresses food ingestion and body weight gain in rats, Brain Res. Bull., 5 (1980) 7-11. Van Houten, M. and Posner, B.I., Insulin binds to brain blood vessels in vivo, Nature, 282 (1979) 623-625. Westerink, B.H.C., Damsma, G., Rollema, H., De Vries, J.B. and Horn, A.S., Scope and limitations of in vivo brain dialysis: a comparison of its application to various neurotransmitter systems, Life Sci., 41 (1987) 1763-1776.
28 Wiley, K.P., Simple and complex antibody reactions in radioimmunoassay and the prediction of assay characteristics, J. Immunol. Methods, 84 (1985) 343-358 29 Woods, S.C., Lotter, E.C., Mc Kay, D. and Porte, D., Chronic intracerebroventricular infusion of insulin reduces food intake and body weight in baboons, Nature, 282 (1979) 503-505. 30 Woods, S.C. and Porte, D. Jr., Relationship between plasma and cerebrospinal insulin levels of dogs, Am. J. Physiol., 233 (1977). E331-E334. 31 Young, W.S., Periventricular hypothalamic cells in the rat brain contain insulin mRNA, Neuropeptides, 8 (1986) 93-97. 32 Zucker, L.M. and Antoniades, H.N., Insulin and obesity in the Zucker genetically obese rat "fatty", Endocrinology, 90 (1972) 1320-1330.
Brain Research, 611 (1993) 258-263 ~'~ 1993 Elsevier Science Publishers B.V. All rights reserved 0006-8993/93/$06.00
BRES 18811
Basal and hyperinsulinemia-induced immunoreactive hypothalamic insulin changes in lean and genetically obese Zucker rats revealed by microdialysis K. Gerozissis, M. Orosco, C. Rouch and S. Nicolaidis Neurobiologie des Rdgulations, Collbge de France, Paris (France) (Accepted 8 December 1992)
Key words: lmmunoreactive insulin; Hypothalamus; Ventromedial nucleus; Paraventricular nucleus; Microdialysis; Genetically obese Zucker rat
Lean and genetically obese Zucker rats were implanted with permanent intravenous catheters and a guide canula was aimed at the region of the ventromedial (VMH) and paraventricular (PVN) nuclei to measure immunoreactive insulin collected by means of microdialysis. Preliminary experiments assessed the validity of a novel assay of insulin in microdialysates by a sensitized radioimmunoassay technique. This method was then used to measure basal levels of insulin and those induced by i.v. infusion of 0.5 U of insulin over 30 min in both lean and obese rats. Basal hypothalamic immunoreactive insulin levels were lower in the obese rats than in the lean Zucker rats. When insulin was infused i.v. for 30 min, hypothalamic immunoreactive insulin showed an increase in the 30-60 min sample, which was twice as great in the obese rats. Two facts suggest that the insulin found in the microdialysates was of cerebral, not vascular origin: the short latency in the response and the finding that the response was greater in obese rats.
INTRODUCTION
The role of insulin in the control of food intake and body weight attracts considerable interest. The first investigation showing the direct impact of insulin on central structures used slow and prolonged insulin infusion in the medial hypothalamus of the rat ~3. Such infusions result in hypophagia and, more importantly, in a permanent reduction of body weight t3. Subsequent studies using, hypothalamic and intracerebral infusions of insulin verified this phenomenon in both the rat and the baboon 12'29. The involvement of insulin in the regulation of body weight was also assessed by numerous investigations in the genetically obese homozygous Zucker rat (fa/fa) which shows a peripheral resistance to insulin z3 and whose food intake is suppressed less by centrally administered insulin 1°. Most of these approaches were mainly pharmacological and consisted in the observation of a behavioural effect after the administration of insulin. Physiological studies however, showed several abnormalities in the
obese Zucker rat, including hyperinsulinemia 32, a high insulin concentration in the cerebrospinal fluid but a low ratio of CSF to plasma insulin 2~. Furthermore, low hypothalamic insulin levels have been reported in both the obese and the lean heterozygous (Fa/fa) Zucker rat 2, which suggests the importance of cerebral insulin. The question of the origin of cerebral insulin has been difficult to answer so far. Some authors believe that peripheral insulin reaches the cerebrospinal fluid and possibly the hypothalamic structures 25'3°. Others believe that the blood-brain barrier prevents the entry of insulin into the brain ~'9. Some evidence suggests that mRNA for insulin is present in hypothalamic cells of rats and humans ~'~'3~ Discrepancies in the literature made it difficult to ascertain the physiological involvement of central insulin in the control of body weight and food intake. The pharmacological microinjection approach could not provide an answer. Neither did the reciprocal approach, measuring brain tissue insulin by ex vivo methods, because these assays do not allow dynamic correlations of CSF and peripheral insulin. The introduction of the in vivo technique of brain
Correspondence: M. Orosco, Neurobiologie des R6gulations, Coll~ge de France, 11, place Marcelin Berthelot,75231 Paris Cedex 05, France.
259 microdialysis solves some of these problems because this method allows the measurement of neurochemicals released in the brain extracellular space and therefore allows one to determine the time course of possible variations of cerebral insulin during peripheral insulin infusion. Because insulin has not previously been measured by means of brain microdialysis, it was necessary to measure the relation between peripherally infused insulin and hypothalamic tissue insulin. This correlation is important to know because in most studies, changes in CSF insulin were found to follow changes in plasma insulin 3°. In this investigation, we assayed variations in hypothalamic immunoreactive insulin in response to peripheral insulin infusion in obese and lean Zucker rats. Previous studies have found that the hypothalamus of the obese rat contains low concentrations of insulin-like material. The assay was performed in the feeding-related structures of the median hypothalamus including the ventromedial (VMH) and paraventricular (PVN) nuclei. This area is the site of the inhibitory action of insulin microinjections on feeding 13'14. MATERIALS
AND METHODS
Animals Female Zucker obese (fa-fa) or lean (Fa-Fa) rats, 16 weeks old, were used in this study. They were bred in the Laboratoire d'Endocrinologie, Universit6 Paris-Sud, Orsay, France. Age-paired Wistar rats were also used for comparison of basal levels and validation tests. U p o n arrival in the laboratory, they were housed individually in cylindrical Plexiglas cages designed to allow stressless chronic infusions in the unrestrained animals. T h e temperature of the room was maintained at 2 4 + I°C and the lights were turned on from 06.00 to 18.00 h. Food and water were available ad libitum.
Surgery for chronic hypothalamic microdialysis and simultaneous venous catheterization Rats were anesthetized with ketamine (Imalgbne, M6rieux, 150 m g / k g ) after pretreatment by a muscle relaxant, xylasine (Rompun, Bayer). T h e indwelling venous catheter was implanted according to a previously described technique 15. Briefly, a silicone rubber (Silastic) catheter, filled with viscous polyvinylpyrrolidone solution (20% w / v ) , was inserted in the right external jugular vein approximately 5 m m before it dives u n d e r the clavicle. W h e n pushed in as far as a collar fitted in a position defined by the rat's body weight, the tip arrives at the auricular cavity. T h e vein was tied around the catheter and held in place by further sutures. T h e other end of the catheter was pulled subcutaneously through a slit in the skin at the top of the skull. Once the venous catheter was secured, the animal was placed in a stereotaxic frame (Kopf Instruments). A guide cannula (Carnegie) for the microdialysis probe was aimed at the space between the PVN and the VMH. The coordinates of the guide tip were (according to the atlas of Paxinos and Watson 17) - 1.9 m m anterior, 0.5 m m lateral to and 7 m m ventral to bregma. The dialysis probe protruded 2 m m beyond the guide-tube reaching a point 9 m m ventral to bregma. T h e guide cannula, the venous catheter and a securing device (screw), were fixed to the skull with stainless-steel screws and dental cement. At least one week was allowed for postoperative recovery before the experiments began. During this period, the animals were accus-
tomed to the experimental conditions in their home cages. In order to accustom the rats to the conditions of the experiments, they were permanently connected to the infusion and dialysis system of catheters which were protected by a metal sheath and kept out of the animal's reach by m e a n s of a counterbalanced beam. This arrangement allowed normal movements. The animals showed normal sleep, feeding and body weight gain. All experiments were performed in the animal's own home cages.
Microdialysis procedure The microdialysis m e m b r a n e s (Carnegie) were 2 mm long, with a diameter of 0.5 m m and a 20,000 kDa molecular weight cut-off. According to our in vitro calibration test, the relative recovery of immunoreactive insulin was 3.7%. The perfusion was performed with a Ringer type solution containing 147 m M Na ÷, 2.3 m M Ca 2÷, 4 m M K ÷ and 155.6 m M C I with 0.1% bovine serum albumin added. A system of catheters connected to a two-way swivel 15 placed on the beam, allowed independent perfusion of fluid into the probe and sampling. The flow rate (2 ~ l / m i n ) allowed the collection of 60 /xl fractions every 30 min.
Experimental protocol Because the experiment was conducted in the home cage where the tubes were permanently connected to the rat, the only change on the day of experiment was just the insertion of the microdialysis probe through the guide and the i.v. infusion. The first samples were taken 5 h after insertion of the probe, the venous perfusion was set up by connecting through one channel of the two-way swivel a plastic tube to the head piece of the venous catheter and by switching on the pump. After collecting 4 baseline dialysis samples with aqueous vehicle (control), 0.5 U of insulin (Actrapid, Novo) in a volume of 0.5 ml was then infused intravenously over a 30 min period. Food was removed during the experiment in order not to interfere with the effect of exogenous insulin.
Analysis of immunoreactive insulin in the dialysates Immunoreactive insulin in the dialysates was measured by m e a n s of a radioimmunoassay (RIA), using rat insulin as reference (Novo Labs) and a commercial antiserum raised in guinea pigs, immunized with h u m a n insulin, cross-reacting at 100% with rat insulin. [IZsl]Porcine insulin was used as a tracer (CIS, France). Free insulin was separated from bound insulin by immunoprecipitation, using a second antibody and polyethylene glycol. The R I A was performed in two steps (24 h preincubation in the absence of tracer followed by 90 min incubation). After the last incubation, the tubes were centrifuged, the supernatant was discarded and the radioactivity of the precipitate (bound fraction) was measured in a gamma-counter. T h e tests for validation are exposed in the section results.
Histology At the end of each experiment, the animals were given an overdose of pentobarbital and cardiac perfusion of saline and then of 10% formol was performed. The brain was removed, further hardened in 10% formol, and then serial coronal sections were cut (60 /zm). Following glycerin staining, they were observed at low magnification by an observer unaware of the data obtained during the microdialysis experiments to check for the position of the dialysis probe (Fig. 1). The animals whose brain did not show a correct placement of the guide and the probe were discarded by the independent observer.
Statistics The mean basal levels (before insulin infusion) were calculated for each animal. As in all the experiments of this kind, the percentage of variation at the peak effect of insulin infusion, relative to the m e a n of the four baseline samples was calculated for each animal. T h e significance of insulin effect in each group was assessed by Student's paired t-test.
260 The results were then expressed as mean basal levels+S.E.M. (pg/ml) and as means of percentage variations+S.E.M, for each group of rats, obese or lean. The statistical significance between the two groups was calculated using Student's t-test for basal levels and for percentage of variations.
IMMUNOREACTIVE I N S U L I N (pglml)
2OO-
El
RESULTS
Improvement in RIA sensitivity After a preincubation of 24 h, followed by a 30 min incubation in a reaction volume of 200/~1 containing 50 /xl of Ringer solution, the sensitivity of the R I A was improved tenfold (8.25 p g / t u b e and 100 p g / t u b e for 50% displacement, in sensitized and normal RIA, respectively), with a detection limit of 0.75pg/tube. The intra-assay coefficient was 4.5%. Diluted rat plasma insulin was quantified simultaneously with the one-step and the two-step method. Similar values were obtained (60, 31, 40 and 60, 30, 35 p g / m l , respectively). Several dilutions of rat plasma were tested and the values obtained were similar (288 + 16 p g / m l , mean + S.E.M.). We have followed the kinetics of immunoreactive insulin in the perfusates. As previous studies have shown for most substances 24, insulin concentration, high during the first period, probably because initial lesion of the tissue, then fell and remained relatively stable after a 4-h lag (Fig. 2). In a parallel study, a K + concentration (56 mM) known to induce a peptide release provoked a fivefold increase in insulin level (Fig. 2)
Basal levels The mean basal level of hypothalamic immunoreactive insulin was 63.9 + 7.8 p g / m l in lean rats (n = 8 determinations from 4 animals) and 37.8 ___4.7 p g / m l
S0
0
[ 90
I
150
I
[
I
I
210
270
330
390
TIME ( m i n )
Fig. 2. Immunoreactive insulin (pg/ml) in median hypothalamic microdialysates during a 7-h period, Normal Ringer-BSA solution was replaced by 56 mM K + Ringer-BSA solution 340 min after insertion of the probe.
in obese rats (n = 10 determinations from 4 animals). This difference was statistically significant (t = 2.86, P < 0.05). Wistar rats (n = 28 determinations from 6 animals) gave values similar to lean rats (57.5 _+ 0.4 p g / m l )
Effect of insulin infusion Intravenous insulin increased hypothalamic dialysate immunoreactive insulin by 60.6 + 14.8% in the 30 to 60 min period following the onset of the infusion in lean rats. However, this increase did not reach statistical significance. ~Hypothalamic immunoreactive insulin increased by 246.2 + 63% in the 30 to 60 min period following the onset of the infusion in obese rats. This increase was statistically significant (t = 3.54, P < 0.05). This i.v. insulin-induced enhancement of hypothalamic immunoreactive insulin was thus much larger in obese than in lean rats. The difference in percent increases between the two groups was statistically significant (t = 2.84, P < 0.05) (Fig. 3). DISCUSSION
Fig. 1. Coronal section at the level of anterior hypothalamus depicting the location of the microdialysis probe aimed at the space beween the paraventricular nucleus (PVN) and the ventromedian hypothalamus (VMH).
This study is the first in our knowledge to measure changes in concentrations of immunoreactive insulin in brain, using microdialysis. This measurement was accomplished by means of an improved method of RIA. The tenfold increase in R I A sensitivity was obtained using a classic modification of sequential incubation of ligand and tracer 7'28. The improved R I A method was validated by means of several tests, including stimulation tests, dilution studies and comparison with values obtained by a classic one-step R I A procedure. Some of the above tests were performed with plasma instead of
261 % CHANGES IN IMMUNOREACTIVE INSULIN / BASAL LEVELS
I
I 300 ~
200 -
lO0 -
o 30 min samples infusion LEAN
infusion OBESE
Fig. 3. Changes in immunoreactive insulin ( % / b a s a l levels) in median hypothalamic microdialysates induced by peripheral insulin infusion in lean and obese Zucker rats. Basal levels are given in the text. * P < 0.05 difference between insulin effects in lean and obese rats
microdialysates. Obviously, unlike plasma, microdialysates contain little, if any, nonspecific antigen-like or antibody-like material that could interfere with the specific ligand (insulin)-antibody binding. In fact, one of the major advantages of the microdialysis technique is that the membrane prevents large molecules from diffusing into the perfusate, keeps the samples purified, and protects molecules from being broken down by enzymes once they are recovered from the tissue. Consequently, the chemical analysis can usually be performed without clean-up steps 27 and avoids problems related to the loss of material entailed by purification procedures. However, as the very low insulin concentrations does not permit a complete series of validation and identification tests, we are using the term 'immunoreactive insulin' in the present study. Beyond the validation of RIA, care was taken also in this study to control the experimental conditions of the microdialysis technique. As previous studies have found for other cerebral substances 24, a progressive decline followed by stable basal levels was also observed in the case of insulin. Because 4 h were necessary for the plateau to occur, the determination of the release of immunoreactive insulin was started 5 h after the insertion of the microdialysis probe. The administration of high concentrations of K + via counter-dialysis several hours after the begining of microdialysis triggered a clear-cut peak of insulin response, which further suggests that the system was effective. The presence of bovine serum albumin in the perfusate overcomes the problems of adsorption of insulin by the probes and the tubes. The in vitro relative recovery was 3.7%, which corresponds to that obtained with similar high molecular weight substances H. It is
known that an estimate of the actual concentration of the substance of interest (insulin) in the extracellular fluid can be performed by calculating the relative recovery. Because we are aware that in vitro recoveries may not precisely reflect in vivo conditions e7, the resuits are presented only as the amounts of substance that are recovered in the fluid inside the dialysis probe. Three main results were obtained in this study. First, in lean Zucker rats as well as in genetically obese Zucker rats, i.v. insulin infusion produced a rapid elevation of the basal level of immunoreactive hypothalamic insulin. This elevation appeared in the second fraction (30 to 60 min) collected after the onset of the infusion. Dynamic assays of cerebrospinal fluid insulin have previously suggested that peripheral insulin could reach the brain tissue via its penetration in the cerebroventricular space following a dumped and delayed profile 3°. It was also suggested that insulin delivery to the neuropil may be due to a specialized brain transport system (transcytosis)~6'19" Alternatively, brain insulin may be produced locally as suggested by findings on expression of insulin mRNA in fetal but also adult brain cells localized precisely in the periventricular hypothalamuss'~s'31. However, other investigators, using the polymerase chain reaction method did not verify the notion of endogenous production of insulin in the hypothalamic tissuesS.The present data suggest (but do not prove) that peripheral insulin passes rapidly into the hypothalamus, or perhaps the peptide is even produced there, as some authors suggest sAs'31. The second result deals with the basal level of hypothalamic immunoreactive insulin, which is significantly higher in the lean than in the genetically obese rat. This observation confirms and extends a previous report of such a difference in basal insulin content of tissue homogenates of the entire hypothalamus2. Although we lack comparative measurements between various structures, our data suggest that the lower levels of insulin found in the VMH and PVN are responsible for the impaired regulation of food intake and body weight of the obese rat. The third result shows that, despite lower basal hypothalamic levels in the obese rat, the increase in local insulin in response to a peripheral load is significantly more pronounced. Again this differential increase is observed in the hypothalamic region that has been proposed to be the central site of action at which insulin acts in its role in body weight regulation and adjusment of feeding. Since insulin in the VMH and PVN region seems to inhibit feeding, its large changes in the obese rat may account for the lengthening of the intermeal interval in these hyperphagic animals 3. Taking these results into account, it is possible to
262
speculate on the important question of whether the insulin found in the hypothalamus was synthesized locally or whether it was derived from the insulin that was injected peripherally. In the latter case, the classically proposed mechanism would be an uptake from the plasma to the CSF and then a diffusion to various brain structures, including the hypothalamus 3°. Another transport mechanism, from the plasma to the brain parenchyma, either actively by means of transcytosis 16'19 or via fenestrated capillaries t'4'26 and subsequently to the CSF, has also been proposed. Our results in the lean rats are consistent with the hypothesis that the hypothalamic insulin is of peripheral origin. Although other investigators have suggested that the uptake of insulin from plasma to the CSF is slower in the obese rat 22, we observed a larger increase in the obese rat. If the hypothalamic insulin of obese rats is of peripheral origin, then the diffusion from the CSF to the brain parenchyma must be accelerated in these animals. According to the present knowledge, less insulin should have passed through the insulin resistant blood-CSF parenchyma barriers of the obese rat than through the normal barriers of the lean rat 2°. In conclusion, the use of the microdialysis technique for collecting samples of extracellular fluid from the brain for insulin assays in the genetically obese Zucker rat provides a new approach to the investigation of the origin of cerebral insulin and its possible role in obesity. Our data favour the possibility of a local cerebral insulin production that can respond to a peripheral signal such as an intravenous insulin infusion. More experimental manipulations and shorter sample collection intervals are needed before we can understand the role of insulin in the hypothalamus. However, these procedural changes will require more sensitive assays of insulin. But the use of microdialysis in freely moving rats bearing intravenous catheters is already capable of contributing in our understanding of the central mechanisms of peripheral peptides that act in parallel on distinct brain structures and induce regulatory or pathological processes. Acknowledgements. The authors would like to thank Marie-Jo Meile for histological preparation, Simon Thornton for the English, Neil Carlson for editorial improvement and Daniel Gripois and MarieFrance Blouquit for the gift of Zucker rats. This work was supported by grants MRT (Minist~re de la Recherche et de la Technologie) and BSN with the contribution of INRA (Institut National de Recheche Agronomique).
REFERENCES 1 Albrecht, J., Wroblewska, B. and Mossakovski, M.J., The binding of insulin to cerebral capillaries and astrocytes of the rat, Neurochem. Res., 7 (1982) 489-494.
2 Baskin, D.G., Stein, L.J., lkeda, H., Woods, S.C., Figlewicz, D.P., Porte, D.Jr., Greenwood, M.R.C. and Dorsa, D.M., Genetically obese Zucker rats have abnormally low brain insulin content, Lift, Sci.,36 (1985) 627-633. 3 Becket, E.E. and Grinker, J.A., Meal patterns in the genetically obese Zucker rat. Physiol. Beha~,., 18 (1977) 685-692. 4 Catalan, R.E., Martinez, A.M., Aragones, M.D., Miguel, B.G. and Robles, A., Insulin action on brain microvessels: Effect on alkaline phosphatase, Biochem. Biophys. Res. Commun., 150 (1988) 583-590. 5 Coker, G.T.I., Studelska, D., Harmon, S., Burke, W. and O'Malley, K.L., Analysis of tyrosine hydroxylase and insulin transcripts in human neuroendocrine tissues. MoL Brain Res., 8 (1990) 93-98. 6 Eng, J. and Yalow, R., Evidence against extrapancreatic insulin synthesis, Proc.Natl. Acad. Sci. USA, 78 (1981) 4576-4578. 7 Ezan, E., Mamas, S., Rougeot, C. and Dray, F., Strategies for developing specific and sensitive hapten radioimmunoassay. In C. Patrono and B.A. Peskar (Eds.), Handbook of Experimental Pharmacology, Springer, Berlin, 1987, pp. 143-179. 8 Giddins, S., Chirgwin, J. and Permutt, M., Evaluation of rat insulin messenger RNA in pancreatic and extrapancreatic tissues, Diabetologia, 27 (1985) 343-347. 9 Havrankova, J.M., Brownstein, M. and Roth, J., Insulin and insulin receptors in rodent brains, Diabetologia Suppl., 20 (1981) 268-273. 1(1 Ikeda, H., West,D.B., Pustek, J.J., Figlewicz, D.P., Greenwood, M.R.C., Porte, D.Jr. and Woods, S.C., Intraventricular insulin reduces food intake and body weight of lean but not obese Zucker rats, Appetite, 7 (19861 381-386. 11 Kendrick, K.M., Use of microdialysis in neuroendocrinology, Methods Enzymol., 168 (1989) 182-205 12 Mc Gowan, M.K., Andrews, K.M., Kelly, J. and Grossman, S.P., Effects of chronic intrahypothalamic infusion of insulin on food intake and diurnal meal patterning in the rat, Behat~. Neurosci., 104 (1990) 371-383. 13 Nicolffidis, S., M6canisme nerveux de l'~quilibre t~nerg6tique, Journ~es Annuelles de Diabetologie de l'Hotel Dieu de Paris, 1 (1978) 152-156. 14 Nicolffidis, S., Lateral hypothalamic control of metabolic factors related to feeding, Diabetologia, 20 (1981) 426-434. 15 Nicolffidis, S., Rowland, N., Meile, M.J., Marfaing-Jallaz, P. and Pesez, A., A flexible technique for long-term infusions in unrestrained rats, Pharmacol. Biochem. Behaz~., 2 (1974) 131-136. 16 Pardridge, W.M., Receptor-mediated peptide transport through the blood-brain barrier, Endocrinol. Reu., 7 (1986) 314-330. 17 Paxinos, G. and Watson, C., The rat brain in stereotaxic coordinates, Academic Press, New York, 1986. 18 Schechter, R., Sadiq, H.F. and Devaskar, S.U., Insulin and insulin mRNA are detected in neuronal cell cultures maintained in an insulin-free/serum-free medium. J. Histochem. Cytochem., 38 (1990) 829-836. 19 Schwartz, M.W., Bergman, R.N., Kahn, S.E., Taborsky, G.J., Fisher, L.D., Sipols, A.J., Woods, S.C., Steil, G.M. and Porte, D.J., Evidence for entry of plasma insulin into cerebrospinal fluid through an intermediate compartment in dogs. Quantitative aspects and implications for transport, Z Clin. Ira,est., 88 (1991) 1272-1281. 20 Schwartz, M.W., Figlewicz, D.P., Kahn, S.E., Baskin, D.G., Greenwood, M.R.C. and Porte, D.Jr., Insulin binding to brain capillaries is reduced in genetically obese, hyperinsulinemic Zucker rats, Peptides, 11 (1990) 467-472. 21 Stein, L.J., Dorsa, D.M., Baskin, D.G., Figlewicz, D.P., Ikeda, H., Frankman, S.P., Greenwood, M.R.C., Porte, D.Jr. and Woods, S.C., Immunoreactive insulin levels are elevated in the cerebrospinal fluid of genetically obese Zucker rats, Endocrinology, 113 (1983) 2299-2301. 22 Stein, L.J., Dorsa, D.M., Baskin, D.G., Figlewicz, D.P., Porte, D.Jr. and Woods, S.C., Reduced effect of experimental peripheral hyperinsulinemia to elevate cerebrospinal fluid insulin concentrations of obese Zucker rats, Endocrinology, 121 (1987) 1611-1615. 23 Stern, J.S., Johnson, P.R., Batchelor, B., Zucker, L. and Hirsch,
263
24
25
26 27
J., Pancreatic insulin release and peripheral tissue resistance in Zucker obese rats fed high and low carbohydrate diets, Am. J. PhysioL, 228 (1975) 543-548. Ungerstedt, U., Measurement of neurotransmitter release by intracranial dialysis. In C.A. Marsden (Ed.), Measurement of Neurotransmitter Release In l/'ivo, Wiley, 1984, pp. 81-105 Vanderweele, D.A., Pi-Sunyer, F.X., Novin, D. and Bush, M.J., Chronic insulin infusion suppresses food ingestion and body weight gain in rats, Brain Res. Bull., 5 (1980) 7-11. Van Houten, M. and Posner, B.I., Insulin binds to brain blood vessels in vivo, Nature, 282 (1979) 623-625. Westerink, B.H.C., Damsma, G., Rollema, H., De Vries, J.B. and Horn, A.S., Scope and limitations of in vivo brain dialysis: a comparison of its application to various neurotransmitter systems, Life Sci., 41 (1987) 1763-1776.
28 Wiley, K.P., Simple and complex antibody reactions in radioimmunoassay and the prediction of assay characteristics, J. Immunol. Methods, 84 (1985) 343-358 29 Woods, S.C., Lotter, E.C., Mc Kay, D. and Porte, D., Chronic intracerebroventricular infusion of insulin reduces food intake and body weight in baboons, Nature, 282 (1979) 503-505. 30 Woods, S.C. and Porte, D. Jr., Relationship between plasma and cerebrospinal insulin levels of dogs, Am. J. Physiol., 233 (1977). E331-E334. 31 Young, W.S., Periventricular hypothalamic cells in the rat brain contain insulin mRNA, Neuropeptides, 8 (1986) 93-97. 32 Zucker, L.M. and Antoniades, H.N., Insulin and obesity in the Zucker genetically obese rat "fatty", Endocrinology, 90 (1972) 1320-1330.