The effect of liver denervation on meal patterns, body weight and body composition of rats

Physiology & Behavior, Vol. 33, pp. 661-667. Copyright o Pergamon Press Ltd., 1984.Printed in the U.S.A.

0031-9384/84$3.00 + .OO

The Effect of Liver Denervation on Meall Patterns, Body Weight and Body Composition of Rats’ LARRY

L. BELLINGER,** VERNE E. MENDEL,? FRED E. WILLIAMS* AND THOMAS W. CASTONGUAYS

*Department of Phys iology, Baylor College of Dentistry, Dallas, TX 75246 tDepartment of Animal Physiology and Animal Science and Food Intake Laboratory, University of California, Davis, CA 95616 *Department of Nutrition and Food Intake Laboratory, University of California Davis, CA 95616 Received BELLINGER,

16 January 1984

F. E. WILLIAMS AND T. W. CASTONGUAY. The effect of liver denervation on composition of rats. PHYSIOL BEHAV 33(4) 661-667, 1984.-Neural liver glucoreceptors have been proposed as a primary controller of food intake (FI). Male Sprague-Dawley rats were either sham operated or liver denervated (LD). LD rats had all tissue cut between the liver and the esophagus, stomach and upper 1 cm of the duodenum. The hepatic artery and surrounding tissue were also removed. Finally the hepatic portal vein and the bile duct were stripped clean and the former phenol treated. Three days after surgery animals were placed in modules for continuous computer monitoring of feeding behavior. At no time after surgery did the daily food intake or body weight of the groups differ signi&tnt~y. Meal size and frequency (light-dark distribution) were determined for 6 days and averaged. Neither parameter was altered by LD. During the next 6 months food intake and body weights bf the groups did not differ significantly. At sacrifice, body composition was directly determined with no significant differences observed between LD and sham operated rats. LD were confirmed histologically. Monoamine histofluorescence of the livers of rats subjected to liver denervation revealed an absence of the normal fluorescence seen on small blood vessels in liver parenchyma of sham operated rats. The data do not support the concept that liver glucoreceptors are a major controller of FL

meaf patterns,

L. L., V. E. MENDEL, body

Liver denervation

weight

and

body

Meal patterns

Body composition

THERE is considerable evidence that the liver contains neural glucoreceptors that send information to the central nervous system (CNS), including the hypothalamus [20, 22, 28, 301. A major issue in the literature is defining the physiological role of the neural liver glucoreceptors. For example it has been postulated that liver glucoreceptors play a primary regulatory role in feeding behavior [25,26]. The glucoreceptors have been proposed to affect both pre- and postabsorptive food intake [25,26]. F’reabsorptive satiety has been explained by ingested food acting on gut chemoreceptors which, via sympathetic nerves to the liver, hyperpolarize the liver glucoreceptors, while postabsorptive satiety is explained by absorbed glucose directly affecting the receptors [25,26]. This hypothesis has received support from some [ll, 13, 19, 23, 301 but not all [l-4, 17, 32,‘35] workers. A serious difficulty in accepting neural liver glucoreceptors as a primary controller of feeding behavior has been the lack of deficits in normal feeding behavior and weight gain in

animals subjected to total hepatic denervation [2-4, 201, hepatic vagal branch (where most glucoreceptor neural recordings are made [20,22]) transection [9, 13,331 or in animal and human total liver transplant patients [S, 6, 15, 18, 311. Nevertheless, following liver denervation subtle changes in daily meal parameters (meal size, length, frequency) and normal circadian distribution of these parameters may occur. One study [ 171 investigated the above parameters and found no effect of selective total liver denervation on any of them. However, this study has been criticized [26] on the grounds that liver glucoreceptor neurons may send fibers within the walls of the hepatic artery to reach the CNS. Also at least one laboratory (quoted by [ 111) has reported that partial liver denervation (i.e., hepatic vagal branch transection) disrupts the circadian feeding rhythm of rats. Thus, there is some controversy and cotiicting findings in the literature concerning liver denervation and feeding rhythms. The first part of the present study investigated daily, as well as, diurnal meal parameters in rats subjected to total

‘Supported in part by BCD research funds. ‘Requests for reprints should be addressed to Dr. Larry L. Bellinger, Department of Physiology, Baylor College of Dentistry, 3302 Gaston Ave, Dallas, TX 75246.

661

BELLINGER

662

liver denervation, which included severing and removing the hepatic artery. The long term effects of liver denervation on food intake, body weight gain and body composition are unknown, however the observation has been made that excessive weight gain may occur following liver transplantation in humans (see [3]). The second part of this study investigated the long term effects of liver denervation on these parameters. METHOD

Male Sprague-Dawley rats were purchased from Harlan Industries, Madison, WI, and upon arrival individually housed in a temperature controlled (25°C) room under a 1ight:dark ratio of 12:12. The rats were given powdered Purina Rat Chow and water ad lib. Also food intake was recorded daily for four days prior to operations. The rats were then randomly selected for denervation or sham operations. Following surgery (see below) the rats were given a three day recovery period and then placed, in groups (4-5 denervates and 3 controls), in individual self-contained feeding modules (LD schedules 12:12) for six-day recording sessions. Food intake was measured continuously from a nonspill cup sitting on a Mettler P-300 series balance. The cup was designed so any spillage fell on the balance. Output from each balance was continuously monitored by a PDP-I 1 computer (Digital Equipment Company; Maryland, MA). A meal was considered terminated if the rat did not eat for ten minutes [8]. After all rats had their feeding patterns recorded they were placed in individual stainless steel cages where they were maintained for six months with weekly body weight and total weekly food intake being recorded. After this period the animals were sacrificed and livers and surrounding tissues removed for histological verification of liver denervation (see below). The carcasses were then analyzed for fat and water using the procedures of Canolty and Koong [7], protein by Kjeldahl method and ash by pyrosis. Data were analyzed by ANOVA for repeated measurements and Student’s t-test. Surgery

The surgical procedures for liver denervations and sham denervations have been previously described in detail [3,4]. Briefly, the animals were fasted overnight then anesthetized with sodium pentobarbital(4 mg/lOO g body weight). A midline incision was made to expose the liver and stomach. The hepatic branch of the vagus nerve was isolated and severed by electro-cautery (to prevent reinnervation) as well all connections between the esophagus and liver. All connective tissue between the stomach and liver were severed. The hepatic artery and surrounding connective tissue were then isolated, tied (to prevent bleeding and reinnervation) and severed. Thus, any nerves that coursed with or within the walls of the hepatic artery were severed. All connective tissue attachments between the hepatic artery and hepatic portal vein were also broken. The portal vein and bile duct were

ET AL.

also stripped

of all loose connective tissue. Finally, the entire surface of the portal vein was treated for two minutes with a 9% phenol solution in 47% ethanol then rinsed with 5 ml of 0.9% saline. The bile duct was not phenol treated because its wall is too delicate to withstand the treatment. A total of 21 rats were denervated. Sham (n=lS) operated animals were opened and the stomach manipulated in a similar manner and time course to the denervated animals. In these animals, phenol-alcohol solution was applied to the left renal veins for two minutes and then rinsed with saline. Histology

Upon completion of testing, tissue samples were examined from both normal and denervated rats, including the portal vein with a portion of its primary tributaries up to a point where it begins to divide upon entering the liver. The hepatic artery and its branches, the hepatic artery proper, or severed remains, and the gastroduodenal artery were included in most samples. The tissues were fixed in 10% phosphate buffered formalin for 24 hours prior to dehydration and infiltration with paraffin. Cross sections of the portal vein were cut (6 pm) and stained with hematoxylin and eosin. Any denervated rats that demonstrated nerve bundles on the portal vein in the area stripped and treated were excluded from data analysis. To further validate the liver denervation technique, thirteen new rats were subjected to liver denervation (n=7) or sham operations (n=6). After ten days the animals were sacrificed and livers examined using the Glyoxylic Acid Histofluorescence Methods [IO] for tissue monoamines. Tissues were examined under epi-illumination using a Zeiss photoscope equipped with Zeiss 487703 filters. RESULTS

In all 21 liver denervated rats the portal vein area was histologically determined to be clear of nerve bundles (see Fig. la and lb). Within the liver parenchyma of sham operated rats fluorescent tracts were clearly ‘observed in association with small blood vessels (Fig. 2a); fluorescence was absent in the liver parenchyma of all denervated rats (Fig. 2b). Prior to surgery, the food intake of liver denervated (22.320.7 g) and sham operated rats (24.32 1.3 g) was similar. Following surgery, both group’s food consumption was slightly decreased, but had recovered to presurgery levels by the fifth postoperative day. At no time, during the recovery period or during the next six months, did the two group’s consumption differ significantly from one another. Similarly, body weights of the denervated and sham operated groups did not vary significantly from one another over the six month measurement period (Table 1). The grouped data (mean of six day measurement period) for 24 hr, F(1,30)=0.01, light phase, F(1,30)=0.28, and dark phase, F(1,30)=0.38, food consumption of the liver denervated and sham operated rats did not vary significantly between groups (Fig. 3). Twenty-four hr, F(1,30)=0.04, light

FACING PAGE FIG. la. The hepatic artery proper (A) is shown after meeting the hepatic portal vein (V). The bile duct (D) is to the left of the vein. Nerve bundles are indicated by arrows. Hematoxyline-eosin, 40x. FIG. lb. The hepatic portal vein (V) and the bile duct (D) after surgical denervation and phenol treatment. Please note the absence of nerve bundles. Hematoxylin-eosin, 40x.

LIVER DENERVATION AND MEAL PARAMETERS

663

BELLINGER

TABLE

N.S.

1 20 -

BODY WEIGHTS(g) OF LIVERDENERVATEDAND SHAM OPERATEDRATS FROMOPERATIONTO SACRIFICE Denervates Operation +4 days +I1 days +25 days +39 days +6 months

256.9 255.0 289.1 331.5 359.7 449.0

2 f f f f *

1.9 3.9 5.7 3.7 4.1 6.8

ET AL.

Shams 256.0 249.0 280.6 329.4 358.5 462.0

2 2 f * f f

2.1 4.2 5.8 5.0 6.1 6.4

P

NS NS NS NS NS NS

Mean f SEM. NS =nonsignificant . df varied: denervates n=21 to 17 and shams n=15 to 13 due to death of some rats as the result of respiratory infections.

lLD

S

Light Phase

LD

Dark Phase

S

24 Hours

FIG. 3. Food intake (24 hr. during light and dark phase) derived from grouped data of liver denervated (LD), n-21 and sham operated (S, n=15) rats. NS=nonsignificant.

F(1,30)=0.11, and dark phase, F(1,30)=0.61, meal size of the two groups also did not differ significantly (Fig. 4). Similarly, (Fig. 5) the two groups’ meal frequency did not deviate significantly over 24 hr, F( 1,30)=0.11, or during the light phase, F(1,30)=0.59, or dark phase, F(1,30)=0.05. All the above main effects x days interaction terms were nonsignificant. Body compositions (percentage H20, fat, protein and ash) were similar in both liver denervated and sham operated rats (Table 2). phase,

DISCUSSION

The data of this study are in agreement with the observation of Louis-Sylvestre and co-workers [17] who reported that liver denervation does not alter meal parameters. The present findings are also compatible with our earlier study [33 which showed no significant differences between liver denervated and sham operated rats in their initial daily meal size and length starting with the first day of postsurgery recovery. These results are also in agreement with previous investigations using the rat [3, 4, 171and the dog [2] which showed that selective total liver denervation does not alter daily food intake or body weight gain. Furthermore, the present data and earlier findings [9,17] do not agree with the reported observation that hepatic vagal branch transection alters the circadian feeding pattern of rats (quoted by [ 111). The histology study revealed that rats subjected to our liver denervation technique show a lack of monoamine fluorescence surrounding blood vessels in the liver parenchyma. This indicates that the denervation technique has removed

sympathetic innervation to the liver. Since meal parameters were not altered, using this denervation technique, these data fail to support the hypothesized importance of sympathetic innervation of liver parenchyma in promoting preabsorptive satiety [25,26]. The present data also indicates that liver denervation, in the rat, does not have a noticeable effect on long term food consumption, body weight gain or body composition. The excessive weight gain and apparent obesity reported in human liver transplant patients (see [3]) is a side effect of the immunosuppressants that are administered to these patients (personal communication with Dr. Starzl) and not the result of denervation. The present data would support such a conclusion. Interestingly, most neural recordings of hepatic glucoreceptors have been made using the hepatic branch of the vagus nerve [20,22], yet transection of this branch does not alter food consumption or body weight gain [9, 13, 331. Furthermore, feeding responses to various metabolic challenges such as insulin, 2-deoxy-d-glucose or epinephrine are not altered when the hepatic branch of the vagus is cut [33,34] or when total liver denervation is accomplished by selective denervations [4]. On the other hand, liver glucoreceptors have been reported to influence gastric acid secretion, gastric electromyogenic activity and the regulation of blood glucose through the modulation of adrenal medulla and pancreatic hormonal secretions 114, 21, 221. Notably these above effects are disrupted after hepatic vagal branch transection 114,221.Thus, the physiological role of liver glucoreceptors may be other than modulation of feeding.

FACING PAGE FIG. 2a. Section of liver parenchyma, from a sham operated rat, subjected to glyoxylic histofluorescence treatment. The arrow points to the fluorescence generated by monoamine nerve tracts associated with small blood vessel& The L denotes the lumen of one vessel. 60x. FIG. 2b. Section of liver parenchyma, from a liver denervated rat, subjected to glyocylic histofluorescence treatment. The arrow points to a small blood vessel whose lumen is denoted by a L. Note the absence of monoamine fluorescence after denervation. 38x.

LIVER DENERVATION

AND MEAL PARAMETERS

665

666

BELLINGER

ET AL.

N.S.

N.S.

l-

N.S.

N.S. -N.S.

LO

Light Phase LD

S

S

LD

Dark

Light Phase

LD

LD

S

S

Dark Phase

24 Hours

FIG. 5. Number of meals of LD and S rats. See legend Fig. 4.

S

24 Hours

Phase

FIG. 4. Meal size of LD and S rats. See legend Fig. 4.

TABLE 2 BODY COMPOSITION AT SACRIFICE OF LIVER DENERVATED AND SHAM OPERATED RATS

Liver Denervated (n= 17) Shams (n= 13) P<

% H,O

% Fat

% Protein

%Ash

59.4 * 0.9

9.2 ? 1.3

24.5 f 0.5

5.1 + 1.2

58.6 * 0.7

7.4 -c 1.6

26.5 ? 1.1

5.5 * 0.2

ns

ns

ns

ns

Mean f SEM. ns=nonsigniticant.

Nevertheless, it has also been reported [ 121 that transection of the hepatic branch of the vagus attenuates the anorexogenic effects of fructose, a sugar that can only slowly pass the blood-brain barrier [24]. While this could mean a disruption of hepatic glucoreceptor atferent information (neurophysiological studies of liver glucoreceptors using fructose do not support this view [20,221), the authors I121 note that their data also could support a possible loss of vagal efferent activity that may have affected feeding behavior by altering liver metabolism. The latter point became apparent after fructose infusions in denervated rats failed to restore liver glycogen or plasma glucose levels comparable to controls. This is not totally unexpected since it has been known for some time that vagal efferents to the liver can modulate the liver’s metabolism of carbohydrates [29]. In support of this type of modulation in the above study [12], there was a trend for systemic glucose concentrations to be lower in the denervated group. This would be in keeping with the hypothesis that liver glucoreceptors, via the hepatic branch of the vagus, influence blood glucose concentrations [2 1,221. Furthermore, normal meal onset has recently been correlated with slight decreases in systemic glucose concentration [ 161which is thought to be sensed by central receptors. Since the liver has the primary role in maintaining postabsorptive

plasma glucose homeostatis it is possible that liver denervation, by removing efferent as well as afferent nerves, may have altered hepatic metabolism so that under certain testing procedures an alteration in feeding behavior would be manifested. As we have mentioned previously [3,4] there is compelling evidence to suggest that in at least the rabbit, liverglucoreceptors may influence feeding behavior [23] when subjected to pharmacologically induced glucose excess, but even these findings are debated [27]. In summary the data reveal that selective total liver denervation in the rat produced no discernable effect on meal size or circadian patterns or on long term effects on food intake, body weight gain or body composition. These data lend no support to the hypothesis that liver glucoreceptors affect feeding behavior in the rat.

ACKNOWLEDGEMENTS The authors wish to thank Ms. Connie Williams for excellent technical assistance and Mary Lou Rodriguez for typing the manuscript.

LIVER DENERVATION

667

AND MEAL PARAMETERS RJ3lWWNCES

1. Bellinger, L. L., R. H. Birkhahn, G. J. Trietley and L. L. Bernardis. Failure of hepatic infusion of amino acids and/or glucose to inhibit onset of feeding in the deprived dog. J Neurosci Res 3: 162173, 1977. 2. BeUinger, L. L., G. J. Trietley and L. L. Bemardis. Failure of portal glucose and adrenaline infusions or liver denervation to affect food intake in dogs. Physiol Behav 16: 2w304, 1976. 3. Bellinger, L. L. and F. E. WiUiams. The effects of liver denervation on food and water intake in the rat. Physiol Eehav 26: 663-671, 1981. 4. Bellinger, L. L. and F. E. Williams. Liver denervation

5.

6.

7.

8.

9.

does not modify feeding responses to metabolic challenges or hypertonic NaCl induced water consumption. Physiol Behav 30: 463-470, 1983. Caine, R. Y., R. Williams, J. L. Dawson, I. D. AnseU, D. B. Evans, P. T. Flute, P. M. Herbertson, V. Joysey,. G. H. W. Keates, R. P. Knill-Jones, S. H. Mason, P. R. Millard, J. R. Pena, B. D. Pentlow, J. R. Salaman, R. A. Sells and P. A. CuUum. Liver transplantation in man. II. A report of two orthotopic liver transplants in adult recipients. Br Med J 4: 541546, 1968. Calne, R. Y., H. J. 0. White, D. E. Yoffa, R. R. Maginn, R. M. Binns, J. R. Samuel and V. P. Molina. Observations of orthotopic liver transplantation in the pig. Br Med J 2: 478-480, 1967. Canolty, N. L. and L. J. Koong. Evaluating protein quality by estimating energy deposition coefftcients. J Nutr 1OS: 15551565, 1978. Castonguay, T. W., D. E. Upton, P. M. B. Leung and J. S. Stem. Meal patterns in the genetically obese zucker rat: A reexamination. Physiol Behav 28: 911-916, 1982. Contreras, R. J. and T. Kosten. Changes in salt intake after abdominal vagotomy: evidence for hepatic sodium receptors. Physiol Behav 26: 575-582,

1981.

10. de la Torre, J. C. Standardization of the sucrose-potassium phosphateglyoxylic acid histofluorescence method for tissue monoamines. Neurosci Lett 17: 339-340, 1980. 11. Friedman, M. I. Hepatic-cerebral interactions in insulininduced eating and gastric acid secretion. Brain Res Bull 5: Suppl4, 63-68, 1980. 12. Friedman, M. I. and J. Granneman. Food intake and peripheral factors after recovery from insulin-induced hypoglycemia. Am J Physiol 244: R374-R382, 1983. 13. Geary, N. and G. P. Smith. Selective hepatic vagotomy blocks pancreatic glucagon’s satiety effect. PhysiolBehav 31: 391-394, 1983. 14. Granneman, J. and M. I. Friedman. Hepatic modulation of insulin-induced gastric acid secretion and EMG activity in rats. Am J Physiol 238: R346-R352, 1980. 15. Kamoda, N. and R. Y. Calne. Orthotopic liver transplantation in the rat. Transplantation 28: 47-50, 1979. 16. Louis-Sylvestre, .I. and J. LeMagnen. A fall in blood glucose levels precedes meal onset in free feeding rats. Neurosci Biobehav Rev 4: Suppl 1, 13-15, 1980. 17. Louis-Sylvestre, J., J.-M. Servant, R. Molimard and J. LeMagnen. Effect of liver denervation on the feeding pattern of rats. Am J Physiol 239: R66-R70, 1980. 18. MacdougaU, B. R. D., R. Y. Cahe, P. McMaster and R. Williams. Survival and rehabilitation after orthotopic liver transplantation. Lancet 1: 13261328, 1980.

19. Martin, J. R. and D. Novin. Decreased feeding in rats following hepatic-portal infusions of glucagon. Physiol Behav 19: 461-466, 1977. 20. Niijima, A. Afferent impulse discharge from glucoreceptors in the liver of the guinea pig. Ann NYAcad Sci 157: 690-700,1%9. 21. Niijima, A. Control of liver function and neuroendocrine regulation of blood glucose levels. In: Integrative Functions of the Autonomic Nervous System, edited by C. McC Brooks, K. Koizumi and A. Sato. Tokyo: University of Tokyo Press, 1979, pp. 68-83. 22. Niijima, A. Visceral tierents and metabolic function. Diabetologiu 20: 325-330, 1981. 23. Novin, D., D. A. VanderWeele and M. Rezek. Infusion of 2-deoxy-D-glucose into the hepatic-portal system causes eating: evidence for peripheral glucoreceptors. Science 181: 851860, 1973. 24. Park, C. R., L. H. Johnson, J. H. Wright, Jr. and H. Batsel. Effect of insulin on transport of several hexoses and pentoses into cells of muscle and brain. Am J Physiol 191: 13-18, 1957. 25. Russek, M. Hepatic receptors and the neurophysiological mechanisms controlling feeding behavior. In: Neurosciences Research, ~014, edited by S. Ehrenpreis and 0. C. Solnitzky. New York: Academic Press, 1971. DV. 213-282. 26. Russek, M. Current status of the-hepatostatic theory of food intake control. Appetite 2: 137-143, 1981. 27. Russell, P. J. D. and G. J. Mogenson. Drinking and feeding induced by jugular and portal infusion of 2-deoxy-D-glucose. Am J Physiol 229: 1014-1018, 1975. 28. Schmitt, M. Infhrences of hepatic portal receptors on hypothalamic feeding and satiety centers. Am J Physiol225: 1089-1095, 1973. 29. Shimazu, T. Central nervous system regulation of liver and adipose tissue metabolism. Diabetologia 20: 343-356, 1981. 30. Shimizu, N., Y. Oomura, D. Novin, C. V. GrijaIva and P. H. Cooper. Functional correlations between lateral hypothalamic glucose-sensitive neurons and hepatic portal glucose-sensitive units in the rat. Brain Res 265: 4%54, 1983. 31. Starzl, T. E., C. G. Groth, L. Brett-Schneider, I. Penn, V. A. Fulginiti, J. B. Moon, H. Blanchard, A. J. Martin, Jr. and K. A. Porter. Orthotopic homotransplantation of the human liver. Ann Surg 168: 392-415, 1968. 32. Strubbe, J. H. and A. B. Steffens. Blood glucose levels in portal and peripheral circulation and their relation to food intake in the rat. Physiol Behav 19: 303-307, 1977. 33. Tordoff, M. G., J. Hopfenbeck and D. Novin. Hepatic vagotomy (partial hepatic denervation) does not alter ingestive responses to metabolic challenges. Physiol Behav 28: 417-424, 1982. 34. Tordoff,

M. G., D. Novin and M. Russek. Effects of hepatic denervation on the anorexic response to epinephrine, amphetamine, and lithium chloride: A behavioral identification of glucostatic afferents. J Comp Physiol Psycho1 %: 361-375, 1982. 35. Williams, F. E. and L. L. BeUinger. The effect of portal and jugular glucose, mannitol and saline infusions on food intake, plasma glucose and insulin in dogs. The Physiologist 26: A106, 1983.