Methods. Animals. All animals used in these studies were white Wistar male rats obtained from Charles River Breeding Labora-

Transcription

1 GASTROENTEROLOGY 1983 ;85 : An Evaluation of the Respective Roles of Portosystemic Shunting and Portal Hypertension in Rats Upon the Production of Gonadal Dysfunction in Cirrhosis D. H. VAN THIEL, J. S. GAVALER, c. F. COBB, and C. J. McCLAIN Departments of Medicine and Surgery. University of Pittsburgh. School of Medicine. Pittsburgh. Pennsylvania To evaluate the differential effects of portal hypertension and portosystemic shunting upon the endocrine changes that occur in men with advanced chronic liver disease, male rats underwent either partial portal vein ligation or direct portocaval anastomosis. Testicular mass was found to be reduced iii both models (p < 0.05). Similarly, estradiol levels were found to be increased (p < 0.05) in both models when compared with sham-operated controls. The increase in estradiol levels was greater in the animals with a complete shunt than in those aniriwls with incomplete shunts developed as a consequence of portal hypertension (p < 0.05). Luteinizing hormone levels were reduced (p < 0.01) in the animals with the greater estradiol levels. As expected, testosterone levels were reduced (p < 0.01) only in the animals with reduced luteinizing hormone levels. These data suggest that portosystemic shunting, and not portal hypertension per se, is responsible, at least in part, for the gonadal injury that occurs with advanced liver disease. The clinicai findings of spider angiomata, palmar erythema, gynecomastia, a female escutcheon, and a female body habitus are seen frequently in men with advanced chronic liver disease (cirrhosis) (1-3). Col- Received November 2, Accepted January 19, Address requests for reprints to: D. H. Van Thiel. M.D J Scaife Hall. University of Pittsburgh. School of Medicine, Pittsburgh, Pennsylvania This work was supported in part by grants from the National Institute for Alcohol Abuse and Alcoholism #AA04425, the Gastroenterological Medical Research Foundation of Southwestern Pennsylvania, and the Hunt Foundation. Dr. Van Thiel is the recipient of a Career Development Award #K01 AA00816 from the National Institute of Mental Health by the American Gastroenterological Association /83/$3.00 lectively these physical findings are termed the feminizing characteristics of cirrhosis (4,5). Such feminized cirrhotic men usually manifest overt evidence of portal hypertension such as esophageal varices, ascites, a caput medusa, and splenomegaly (6). The majority, but not all, have alcoholic liver disease (Laennec's cirrhosis) (1-6). Even in the absence of advanced liver disease, alcohol has been shown to produce hypogonadism manifested by testicular atrophy, loss of libido, reduced testosterone levels, and loss of male secondary sex characteristics (7-12). Collectively these findings have been termed the demasculinizing consequences of alcohol abuse (13). Demasculinization and feminization are not synonomous. Therefore, men can be demasculinized but not feminized as occurs with early alcoholic liver disease or, less often, they may be feminized but not demasculinized as can occur with digoxin or spironolactone therapy. Recently it has been suggested that portal hypertension may be an important factor that is responsible, at least in part, for the feminization of chronic alcoholic men (6,14). Whether portal hypertension per se, or the portal systemic shunting that occurs as a consequence of the intrinsic liver disease and the portal hypertension present in such men is the major determinant of the feminization in such men has not yet been addressed. To resolve this dilemma, the following study was performed. Methods Animals All animals used in these studies were white Wistar male rats obtained from Charles River Breeding Labora-

2 July 1983 SHUNTING AND ESTROGENS 155 Table 1. Body and Organ Weights of the Four Groups of Animals Studied Partial portal Sham portal Portal caval Sham shuntvein ligation vein ligation shunt ed animals Body weights (g) 348 ± ± ± ± 3 Liver weights (g) 14.1 ± ± ± ± 1.8 Testes weight (g) ± ± ± b ± Seminal vesicle/prostate ± 0.114b ± ± b ± weight o p < 0.01 compared with the respective control. b p < 0.01 compared with all other groups. tories, North Wilmington, Mass. They were housed in individual cages in a light-controlled environment with lights on at 7:00 AM and lights off at 7:00 PM daily. The animals were divided into four groups of 20 animals each as follows: Group 1 underwent partial portal vein ligation at age 35 days as described (14); group 2 underwent a sham partial portal vein ligation; group 3 underwent a complete end-to-side portal caval shunt at age 35 days (15); and group 4 underwent a sham portal caval shunt. They were fed a standard commercial rat laboratory diet (Wayne Lab Blox F4) obtained from Best Feeds (Oakdale, Pa.) from weaning until time of death at 90 days of age. Portal Venous Pressure Determinations Just before death the portal venous pressure was determined (14). Essentially, the superior mesenteric vein was identified, and a 25-gauge thin-wall needle attached to a manometer filled to a height of 25 cm with heparinized saline was inserted into it at the junction of the ileocolic branch and the most distal intestinal branch. The stopcock was opened, and at equilibrium the portal venous pressure was read from the manometer and recorded as reading 1. After closing the manometer to the portal venous system, the saline within the manometer was lowered by 4 cm, and after opening the manometer to the portal venous circulation, a second reading was obtained at equilibrium. The mean value obtained by adding the two readings and dividing by 2 was taken as the portal venous pressure. The volume of saline infused as a result of these two determinations was in all cases <1.0 ml. Gross and Microscopic Anatomy After weighing the animals, they were killed by exsanguination. The liver, spleen, testes, prostate, and seminal vesicles were removed and trimmed of all extraneous tissue, and weighed and representative sections were placed in Bouin's solution for histologic study. Histologic sections were prepared and examined with hematoxylin and eosin. The testes were examined for the size of seminiferous tubules, the number and types of germ cells contained within the seminiferous tubules, and the appearance and number of Leydig cells. The prostate and seminal vesicles were examined for androgen effect as determined by the appearance of the epithelial lining cells and evidence of secretory activity. Plasma Hormone Levels Plasma levels of estradiol and testosterone were determined in duplicate using radioimmunoassay methods established in our laboratory (16,17). All samples were measured in a single assay. The detection limit for the estradiol assay was 0.1 pg and for the testosterone assay, 10 pg. The intraassay variation was 8% and 5% respectively. Plasma rat luteinizing hormone (LH) was assayed using reagents supplied by the National Institute of Arthritis, Metabolism and Digestive Diseases (NIAMDD) Rat Pituitary Hormone Distribution Program. National Institute of Arthritis, Metabolism and Digestive Diseases-Rat LH RP-l was used as the reference preparation. The hormone was iodinated with using the choloramine-t method (18,19). The detection for the assay was 4 ng with an intraassay variation of <8%. Statistical Methods All data are reported as mean values ± SEM. Analysis of variance was used for all statistical analyses. Appropriate paired comparison procedures were used when necessary (20). A P value of <0.05 was considered as being significant. Results Both groups of animals with the surgically created portal systemic shunts had smaller liver weights (p < 0.01), smaller testicular (p < 0.05) weights, and smaller seminal vesicle prostate weights than did their respective controls (Table 1). In each case however, the animals with larger shunts Table 2. Spleen Weight and Portal Venous Pressure in the Four Groups of Animals Studied Partial Sham portal portal Portal Sham vein vein caval shunted ligation ligation shunt animals Spleen weight 738 ± ± ± ± 14 (mg) Portal venous 20.5 ± ± ± ± 0.6 pressure (em saline) o p < 0.01 compared with control.

3 Figure 1. Testicular histology of the four groups of animals studied: animals with po rtocaval shunts (A), animals with partial portal vein ligation (B), and controls for A and B (C and D, respectively). A is shown at X250. All other photomicrographs are shown at X100. (Hematoxylin and eosin.)... Z? U1 '" g. <: 0-<: Cl o or ::0 ttl...,z f5tri CIl..., ;> Cl r ;> tri..., ~ > Z <: '"...<11

4 July 1983 SHUNTING AND ESTROGENS 157 Figure 2. Prostate histology obtained from the animals with the portocaval shunts (A) and the animals with the partial portal vein ligation (B). The two control groups did not differ from B. Photomicrographs are shown at x 100. (Hematoxylin and eosin.) (the shunted animals) had smaller organ weights than did the animals with the lesser degree of shunting (partial portal vein ligation animals) (p < 0.01). In contrast, the animals with partial portal vein ligation had the largest spleens and the greatest portal venous pressure (Table 2). As expected, the animals with the end-to-side portal caval shunts had the smallest spleens and the lowest portal venous pressure. The histologic appearance of the liver of the four groups of animals did not differ (data not shown). In contrast, the testes of the shunted animals were obviously smaller than their respective controls (Table 1) and contained fewer germinal cells in the seminiferous tubules (Figure 1). Similarly, the epithelium of the prostate and seminal vesicles was less well developed in the animals with portal systemic shunts and the degree of shunting correlated semiquantitatively with the relative atrophy of these sex steroid-responsive tissues (on a histologic scale of 0 to 3 with 0 being normal and 3 being marked atrophy) (Figure 2). The testosterone levels from the two groups of sham-operated animals and the animals with partial portal vein ligation were similar (Table 3). In contrast, the testosterone levels in the animals with the portal caval shunts were reduced markedly (p < 0.01). Moreover, the LH levels of the two shamoperated control groups and the animals with partial portal vein ligation were similar (Table 3). In contrast, the LH levels in the animals with the portal caval shunts were markedly reduced (p < 0.01) compared with the other three groups (Table 3). The estradiol levels in both groups of animals with

5 158 VAN THIEL ET AL. GASTROENTEROLOGY Vol. 85, No.1 Table 3. Plasma Hormone Levels in the Four Groups of Animals Studied Partial portal Sham por- Portal Sham vein tal vein caval shunted ligation ligation shunt animals Testosterone 2.91 :t :t :t :t 0.34 (ng/mi) LH 50:t11 52:t10 12:t3 48:t9 (ng/ml) Estradiol 16.2 :t 0.7 b 11.9:t :t :t 0.4 (pg/mi) p < 0.01 compared with the control group. b p < 0.05 compared with the respective control group. surgically created portal systemic shunts were increased compared with the control animals (p < 0.05) and the increase was proportional to the degree of shunting produced. Discussion The present study examines the effect of two different degrees of portal systemic shunting [one, a high-grade complete shunt (portal caval shunt); the other, a low-grade shunt (partial portal vein ligation)] upon testicular size and plasma levels of the two major sex steroids and that of a single gonadotropin, LH. Both shunted models demonstrated a reduction in testicular size (p < 0.05) and the group with the larger shunts had lower plasma levels of LH as compared with the other three controls (p < 0.01). In addition, each was associated with an increase in plasma estradiol (p < 0.05) levels as compared with the two sham-operated controls. Moreover, the plasma hormone alterations produced were found to be proportional to the degree of portal systemic shunting allowed by the two models. A previous study that used one of these two models attempted to relate the change observed in plasma estrogen levels as occurring as a direct result of the portal hypertension produced by the model (14). In our study, one model (the shunted animals) has a marked reduction in portal venous pressure while the other model (partial portal vein ligated animals) had a marked increase in portal venous pressure when compared with the two control groups. Both experimental groups, however, had increased estradiol levels and reduced LH levels (Table 3). In contrast to the markedly different and divergent portal venous pressure levels observed between the two experimental models, both groups allowed portal systemic shunting to occur. The animals with the portal caval shunts had a 100% shunt, while the animals with the partial portal vein ligation had a somewhat lesser degree of shunting that was determined by the number of and the extent of the collateral vessels that developed as a consequence of the increased portal venous pressure produced (14). Plasma levels of estradiol increased in both groups, while plasma levels of LH decreased only in the animals with greater degree of shunting, which was associated with the greater estradiol levels. Testicular size and plasma testosterone levels were decreased markedly in the animals with highgrade shunts. The same animals also had the most marked alteration in estradiol and LH levels as noted above, suggesting that the high-grade portosystemic shunting produced might be responsible for the observed endocrine changes. In contrast, testicular size was only minimally reduced and testosterone and LH levels were not reduced in the animals with the lesser degrees of shunting (partial portal vein ligation animals). These data suggest, therefore, that portal venous shunting allows initially for an increase in estradiol levels and that with more advanced degrees of shunting, testicular atrophy, reduced LH, and reduced testosterone levels can be observed with yet a further increase in the plasma estradiol levels (Table 3). The increase in estradiol levels observed would appear to occur as a result of the shunting and to increase in direct proportion to the severity of the shunting and not, as has been suggested previously, as a direct result of an increased portal vein pressure. A pathophysiologic mechanism that may account for these findings is as follows. Presumably as a result of portosystemic shunting, weak adrenal androgens and estrogens in portal venous blood escape hepatic extraction and provide substrate for nonhepatic peripheral aromatization to estrogens. As estrogen levels increase, LH levels decrease, and ultimately testosterone levels decrease and testicular atrophy occurs. References 1. Lloyd CW, Williams RH. Endocrine changes associated with Laennec's cirrhosis of the liver. Am J Med 1948;4: Bennett HS, Baggenstoss AH, Butt HR. The testes, breast and prostate of men who die of cirrhosis of the liver. Am J Clin PathoI1950;20: Baker HWG, Burger HG, dekretser DM, et al. A study of the endocrine manifestations of hepatic cirrhosis. Q J Med 1976;45: Van Thiel DH, Lipsitz HD, Porter LE, et al. Gastrointestinal and hepatic manifestations of chronic alcoholism. Gastroenterology 1981;81: Van Thiel DH. Disorders of the hypothalamic-pituitary-gonadal axis in patients with liver disease. In: Zakim D, Boyer TD, eds. Hepatology, a textbook of liver diseases. Philadelphia:W. B. Saunders Co., 1982:

6 July 1983 SHUNTING AND ESTROGENS Van Thiel DH. Feminization of chronic alcoholic men: a formulation. Yale J Bioi Med 1979;52: Van Thiel DH, Lester R, Sherins RJ. Hypogonadism in alcoholic liver disease: evidence for a double defect. Gastroenterology 1974;67: Van Thiel DH, Gavaler JS, Lester R. Ethanol inhibition of vitamin A metabolism in the testes: possible mechanism for sterility in alcoholics. Science 1974;186: Ylikahri R, Huttunen M, Harkonen M, et al. Low plasma testosterone values in men during hangover. J Steroid Biochern 1974;5: Badr R, Bartke A. Effect of ethyl alcohol on plasma testosterone level in mice. Steroids 1974;23: Gordon GG, Altman K, Southren AL, et al. Effect of alcohol (ethanol) administration on sex hormone metabolism in normal men. N Engl J Med 1976;295: Van Thiel DH, Gavaler J. Lester R, et al. Alcohol induced testicular atrophy: an experimental model for hypogonadism occurring in chronic alcoholic men. Gastroenterology 1975;69: Van Thiel DH, Gavaler JS, Eagon PK, et al. Alcohol and sexual function. Pharmacol Biochem Behav 1980;13: Van Thiel DH, Gavaler JS, Sloane FL, et al. Is feminization in alcoholic men due in part to portal hypertension? A rat model. Gastroenterology 1980;78: Kennan FL. Modifications of the portocaval shunt in the rat. J Applied Physiol 1965;20: Hotchkiss JL, Atkinson LE, Knobil E. Time course of serum estrogen on luteinizing hormone concentrations during the menstrual cycle of the rheusus monkey. Endocrinology 1971;89: Nieschlag E, Loriaux D. Radioimmunoassay for plasma testosterone. Z Klin Cherne Klin Biochem 1972;10:i Greenwood RC, Hunter WM, Glover JS. The preparation of 131labeled human growth hormone of high specific radioactivity. Biochem J 1903;89: Sherins RJ, Viatukaitis J, Chramback A. Physical characteristics of hfsh and its desailyation products by isoelectric focusing and electrophoresis in polyacrylamide gel. Endocrinology 1972;92: Steel NGD, Torri JH. Principles and procedures of statistlcs: a biometrical approach. 2nd ed. New York: McGraw-Hill Book Co., 1980: