Vitamin A modulation of xenobiotic-induced hepatotoxicity in rodents.

Vitamin A (VA, retinol) has been shown to modulate cells of the immune system. When rats are pretreated with VA (75 mg/kg/day) for 7 days, there is greatly potentiated liver damage upon subsequent exposure to hepatotoxicants such as CCl4. This potentiated damage can be blocked by superoxide dismutase or catalase, suggesting that reactive oxygen species are playing a major role in the increased liver injury. The studies reported here examined VA-induced modulation of CCl4 hepatotoxicity in different strains of male rats, female rats, and different strains of male mice. Also, the role of VA-induced weight loss on potentiation of CCl4 injury was investigated. Rats or mice were dosed with VA (retinol) at 75 mg/kg/day, po, for 7 days. In an additional VA dose-response study, mice were given VA at 18.8, 37.5, or 75 mg/kg/day, po, for 7 days. On day 8 they were given a dose of CCl4 which elicited mild hepatic damage. On day 9 they were necropsied. Male and female Sprague-Dawley rats, and male Fischer-344 and athymic nude rats pretreated with VA had an approximately 10-fold increase in liver damage as compared to vehicle controls. Pretreatment of male Balb/C, C3H/HeJ, Swiss-Webster, or athymic nude mice resulted in a marked reduction of CCl4-induced hepatic damage. In the dose-response study in mice, increasing doses of VA elicited increasing amounts of protection from CCl4-induced liver injury. Paired feeding studies revealed that VA-induced weight loss (or decreased weight gain) had no effect on subsequent VA-induced potentiation (rats) or protection (mice) from hepatic damage caused by CCl4.(ABSTRACT TRUNCATED AT 250 WORDS)


Introduction
Many chemicals, such as carbon tetrachloride, allyl alcohol, and acetaminophen are known to cause liver damage upon exposure. Likewise, numerous compounds, such as vitamin A (retinol), endotoxin (lipopolysaccharide), and methyl palmitate have been demonstrated to modulate the function of the immune system after their administration (1,2). Little is known, however, about the role of the immune system in the progression of liver damage and even less well characterized are the effects of immune system modulators on xenobiotic-induced hepatotoxicity.
Kupffer cells [(KC), resident macrophages of the liver] and neutrophils polymorphonuclear (PMN) are phagocytic white blood cells that have recently been shown to play a role in liver injury (3)(4)(5). Studies suggest that the contribution of KC or PMN to liver injury varies depending on the hepatotoxic agent. Regardless of which type of white blood cell is involved, release of cytotoxic factors are responsible for increased hepatocellular damage. Examples of cytotoxic factors are reactive species of oxygen (ROS) such as superoxide anion and nitric oxide, and proteases (6,7). Presumably, when hepatocytes or hepatic nonparenchymal cells (NPC) are damaged, chemical mediators are released which stimulate KC and PMN to release ROS and proteases resulting in increased injury to surrounding hepatocytes and NPC. Therefore, immunomodulators could affect KC and PMN in one of two ways; either to suppress their function resulting in decreased hepatic damage or to increase their function, causing increased or potentiated damage. Experiments in this laboratory have been principally focused on vitamin A-induced modulation of hepatotoxicity caused by xenobiotic compounds.
Vitamin A (VA, retinol) is found in many foods and is available as an over-thecounter food supplement. Besides its requirement for normal cellular function, it has been shown to be beneficial as an adjunct to cancer chemotherapy and in the prevention of some types of skin cancer. Among its physiological effects, VA (as retinol or retinoic acid) can activate macrophages in vivo and in vitro (2,8).
Original investigations in this laboratory sought to determine the consequences of high doses of VA on the subsequent exposure to a hepatotoxicant such as carbon tetrachloride (CC14). It was discovered that the administration of VA (retinol) by oral gavage to male Sprague-Dawley rats at a dose of 75 mg/kg/day for 7 days, greatly potentiated the hepatotoxicity of a subsequent ip dose of CC14 as determined by increased plasma ALT activity and increased histologic hepatic damage (9). In a VA dose-response study in rats, increasing doses of VA from 30 to 75 mg/kg/day, po, for 7 days, elicited increasing potentiation of CC)4 -induced liver injury (10).
Administration of superoxide dismutase (SOD) or catalase eliminated the potentiated damage, implicating release of ROS by KC as the cause of the increased damage (9). In addition, when KC isolated from VA-pretreated rats were stimulated in vitro, they produced significantly greater amounts of superoxide anion as compared to KC from controls (11). However, it was noted that VA-treated rats had reduced weight gain or lost weight during the treatment period, while the corn oil controls had a gain in body weight of approximately 20% during the treatment period. Since the preceeding experiments were only performed in male Sprague-Dawley rats and because of concern about weight loss in VA-treated animals, the following experiments were designed to answer three questions. First, do other strains of rats, and do female rats exhibit the same VA-induced Environmental Health Perspectives Aquasol A Drops (retinol, Astra Pharmaceutical Products, Westborough, MA) or its vehicle (10% propylene glycol, 7% Tween 20 in distilled water) or retinol (Fluka Chemical Co., Ronkonkoma, NY) dissolved in corn oil (rats in paired-feeding study) were administered to each animal by oral gavage at a dosage of 75 mg/kg/day. In a VA dose-response study, male Swiss-Webster mice were dosed with vehicle or Aquasol A at 18.8, 37.5, or 75 mg/kg/day, po. In mice, doses of VA greater than 75 mg/kg/day for 7 days were judged to be too toxic for evaluation. Rats and mice were dosed for 7 consecutive days. On day 8 rats were dosed with corn oil (vehicle) or CCl4 (Aldrich Chemical Co., Milwaukee, WI) at 0.2 ml/kg (nude and Sprague-Dawley) or 0.1 ml/kg (F-344), ip. Because of the sensitivity of mice to CC14-induced liver damage, its very steep-dose response in this species, and due to strain differences, three different doses of CCI4, administered by the ip route, were necessary to obtain comparable degrees of liver injury in these mouse strains. Therefore, on day 8 Balb/C mice were dosed with CC14 at 12.5 pl/kg, Swiss-Webster mice at 20 pIl/kg and C3H/HeJ and athymic nude mice at 15 pl/kg. Twenty-four hours after the administration of CC14, the animals were killed by CO2 inhalation and exsanguinated through the posterior vena cava. Heparinized blood samples were immediately collected and plasma separated for determination of ALT activities using a Beckman DU-7 spectrophotometer (Sigma kit No. 59-20, Sigma, St. Louis, MO). A section of liver was fixed in 10% neutral buffered formalin, embedded in paraffin, sectioned at 5 to 6 pm, stained with hematoxylin and eosin and examined microscopically.
In paired-feeding studies, male Sprague-Dawley rats (250-300 g) or male Swiss-Webster mice (25 g) were individually housed. To ensure that each VA-vehicle animal ate the same amount as a VAtreated, the amount of rodent-chow eaten by each VA-treated animal was weighed daily, and the same percentage of food per gram body weight was given to its pair-fed VA-vehicle control animal the next day. With regard to VA and CCd4 dosing, and necropsy procedures, these animals were treated identically to those above.

Results
Pretreatment of male F-344, athymic nude, and male and female Sprague-Dawley rats for 7 days with VA (75 mg/kg/day, po) resulted in potentiation of hepatic damage as determined by increased plasma ALT activities and histological examination of their livers (Table 1). Although CC14 at this dose (0.2 ml/kg ip) caused greater baseline toxicity in male athymic nude and female Sprague-Dawley rats as compared to male Sprague-Dawley and F-344 rats, VA-pretreatment resulted in an approximately 10-fold increase in CCl4-induced hepatotoxicity. Histologically, rats treated with vehicle/CCl4 had mild centrilobufardamage in the two to four rows of hepatocytes surrounding the central vein. These cells were gready enlarged and contained several large, clear intracytoplasmic vacuoles. Hepatocytes immediately adjacent to the central vein were necrotic with shrunken, fragmented eosinophillic cytoplasm and pyknotic nuclei.
With VA pretreatment, this damage was more extensive with severe degeneration and necrosis that included both centrilobular and midzonal hepatocytes (Figure 1).
Pretreatment of male Swiss-Webster, Balb/C, C3H/HeJ (macrophage impaired), and athymic nude mice for 7 days with VA (75 mg/kg/day, po) resulted in protection from CCl4-induced hepatic damage as determined by plasma ALT activities and microscopic examination of liver sections. Balb/C and Swiss-Webster mice had an approximately 4-fold protection, while C3H/HeJ had approximately 10-fold protection, and athymic nude mice, a 40-fold protection ( Table 2). The VA doseresponse study in male Swiss-Webster mice revealed that as the daily dose of VA increased, from 18.8 to 75 mg/kg/day, the degree of protection against CCl4-induced liver injury also increased ( Figure 2). Histologically, mice treated with vehicle/ CCl4 had moderate degeneration and necrosis of centrilobular hepatocytes extending to midzonal regions, similar to that seen in rats. With VA pretreatment, the centrilobular hepatic damage was much reduced with minimal hepatocyte damage around the central vein in Swiss-Webster and Balb/C mice and virtually no hepatocyte damage seen in C3H/HeJ or athymic nude mice (Figure 3).
In the paired feeding study, rats fed adlibitum and dosed with VA-vehicle gained A (A) Swiss-Webster, Veh/CCI4 0.02 ml/kg. There is (C) C3H/HeJ, Veh/CCI4 0.015 ml/kg. There is severe severe degeneration and necrosis of centrilobular degeneration and necrosis of centrilobular hepatocytes hepatocytes.
with infiltration of numerous leucocytes.      VA (75 mg/kg/day, po) for 7 days or its vehicle followed by CC14 (0.2 ml/kg, ip). Mean ± SEM. *, significantly different from Veh/CCI4 and Veh/CCI4 (pair-fed) groups at p<0.05. 40 ± 4 g during the course of the study. Animals given daily doses of VA gained only 6 ± 2 g during the study, while rats pair-fed to the VA-treated rats only gained 17 ± 4 g during treatment. Following dosing with CC14 (0.2 ml/kg ip), VA-vehicle and pairfed rats had minimal elevations in plasma ALT activities (approximately 325-350 U/L) while VA-pretreated rats had an approximate 1 0-fold increase of plasma ALT activities (Table 3). Male Swiss-Webster mice and their pair-fed vehicle controls both lost approximately 25% of their body weight over the 7 day VA dosing period ( Figure 4). However, when dosed with CC14, VA-pretreated mice had an approximately 10-fold reduction in plasma ALT activities as compared to both the pair-fed and non-pair-fed CCI4 groups (Table 4).

Discussion
Vitamin A (retinol) is a widely used, overthe-counter food supplement. Since it is fat soluble and stored in the Ito cells of the liver, many adults have increasing hepatic stores of it through their lifetimes (12). Studies over the last decade have shown that VA can modulate elements of the immune system, including macrophages. Vitamin A and other retinoids (retinal and retinoic acid) stimulate peritoneal macrophages to increased phagocytosis and interleukin production (2) and alveolar macrophages to increased phagocytosis and tumoricidal activity (8). Retinol, retinoic acid, and retinyl palmitate have been shown to increase the release of superoxide anion by phagocytic cells (13,14). In turn, the secretory products of stimulated peripheral blood monocytes and macrophages, including KC, have been shown to be involved in hepatic injury following exposure to hepatotoxicants (6,7,15). Superoxide dismutase (SOD) and catalase (CAT) degrade superoxide anion and hydrogen peroxide, respectively. Methyl palmitate (MP) administration inhibits the function of macrophages, including Kupffer cells. In this laboratory, it has been demonstrated that the administration of SOD, CAT, or MP to rats alleviates the VA-potentiated increase in CC)4-induced hepatic damage (9). This suggests that the potentiated liver damage is due, at least in part, to reactive oxygen species released by these phagocytic cells.
In the studies reported here, VA potentiated CC14-induced hepatotoxicity to a similar degree in male and female Sprague-Dawley rats and in F-344 and athymic nude rats. Vitamin A potentiation of CC14 in athymic nude rats suggests that potentia-tion can occur in the absence of T-lymphocytes in this strain. Fischer-344 rats have been shown to be more susceptible to pulmonary oxidant-induced injury (16). Such a strain difference may also be expressed in the liver of F-344 rats since Smith et al. (17) have shown them to be more susceptible to the hepatotoxic effects of the redox cycling compound, diquat. In our VA model, they responded in a similar manner as both the Sprague-Dawley and athymic nude rats, except that they required a lower dose of CC14 to produce a comparable degree of liver injury.
Mice responded differently than rats in this particular model of chemical induced hepatotoxicity. As seen in the control animals in these studies, mice are more susceptible to CCI4-induced hepatotoxicity than rats. Doses required to produce comparable elevations in serum ALT activities were approximately one-tenth of those required for rats. However, the most surprising result was that pretreatment of mice with VA protected mice from CC14induced liver injury.
Mice exhibited considerable strain variability with respect to VA protection. Balb/C and Swiss-Webster mice had an approximately 4-fold protection; C3H/HeJ (macrophage impaired), 10-fold protection; and athymic nude mice (T-lymphocyte deficient), 40-fold protection. In the dose-response study in Swiss-Webster mice, increasing doses of VA up to the highest possible dose elicited increasing protection from CCl4-induced liver injury. This is in contrast to rats, which show increasing potentiation of CC14 with increasing doses of VA (10). The mechanism(s) responsible for VA-associated protection of CC14-induced hepatotoxicity in mice remain(s) to be elucidated. It is possible that differences in retinol metabolism and tissue distribution between rats and mice may explain, in part, the species differences reported here. In rats, hepatic retinol and retinyl palmitate concentrations increased following VA treatment (10). We are currently determining hepatic concentrations of retinol and retinyl palmitate in mice following VA treatment at 75 mg/kg for 7 days. In addition, studies of the effects of VA on CC14 metabolism by murine hepatocytes and on phagocytic cell function are currently in progress.
It was observed that oral gavage with VA resulted in weight loss in mice and a reduction of weight gain in rats. Since VA can cause anorexia and weight loss in humans, this finding was not unexpected (18) were conducted in both species to determine the role of food intake in VA modulation of CCI4-induced hepatotoxicity. In both rats and mice, vehicle control animals, pair-fed to VA animals, had reductions in weight gain (rats) or decreases in body weight (mice) similar to the VA animals. However, when dosed with CC14, there was no alteration of liver injury. Therefore, in these studies, VA-associated weight loss or decrease of weight gain was not associated with VA modulation of CC14-induced liver injury. Recent studies in our laboratory indicate that VA administration to rats may induce increased production of tumor necrosis factor-ax(TNF-x, data not shown). Since TNF-a (19) can result in weight loss, this is a possible cause of the weight loss or decreased weight gain (rats) that was observed.
These studies demonstrate that large doses of VA can modulate the hepatotoxic effects of subsequent administration of CC14. The wide variability of responses between species and among strains of the same species (mice) make it difficult to predict the responses that will be seen in other species, including humans. However, retinoids have been reported to increase macrophage and monocyte activity in peo-ple. After consumption of excessive amounts of VA some humans could be exposed to other immunomodulators and also to potentially toxic chemicals. Therefore, it will be important to better characterize how VA and other immunomodulators interact with the wide variety of chemicals (drugs, environmental pollutants, etc.) to which people are exposed and to determine the mechanisms that explain the differences observed in mice and rats with respect to the effects of VA on chemical induced hepatotoxicity. Such information will assist in the extrapolation of these results obtained in animals to humans.