Inhibition of hepatocellular carcinoma by glycyrrhizin in diethylnitrosamine-treated mice

2To whom correspondence should be addressed Glycyrrhizin (GL) is widely used in Japan as a therapeutic agent for chronic active liver diseases. However, its action on hepatocarcinogenesis remains to be elucidated. To clarify its effect, mice treated with diethylnitrosamine (NDEA) with or without GL were analyzed. Five-week-old male BALB/c mice were divided into two groups, GL (n J 50) and C (n J 47). Mice in the GL group intramuscularly received 2 mg of GL 3 days a week, and mice in the C group received the same volume of saline in the same way. After 2 weeks, the mice were treated with an i.p. injection of 75 mg/kg body wt of NDEA weekly for 3 weeks and 100 mg/kg body wt of NDEA weekly for the following 3 weeks. Thirty additional mice that did not receive NDEA treatment were divided into two groups, GC (n J 15) and SC (n J 15). They received GL or saline, respectively. Mice in the 4 groups were killed every 5 weeks after the last injection of NDEA from 7 weeks to 32 weeks. Liver function tests such as AST and albumin were significantly improved in the GL group compared with the C group (P < 0.05, each). Although liver nodules appeared in the C group at 22 weeks, they were not observed until 32 weeks in the GL group. At 32 weeks, the mean number of liver tumors, composed of adenoma and hepatocellular carcinoma (HCC), in the GL group was 0.71, which was significantly decreased compared with 1.64 of the C group (P < 0.05). The mean number of HCC in the GL group was 0.29/liver, which was lower than 0.82/liver in the C group (P < 0.05). The incidence rate of HCC at 32 weeks was 64% in the C group and 21% in the GL group (P < 0.05, C versus GL group). Our results suggest that GL treatment inhibits the occurrence of HCC.

© Oxford University Press 59 indicate that the highest prevalence of anti-HCV positivity is found in Japan (66-94%), that a relatively high prevalence (57-79%) is found in Africa, Italy and France, and that a lower prevalence (Ͻ40%) is typical in the USA, Switzerland, Austria, China and Hong Kong (2).HCC can develop late in the course of viral infection, especially of chronic HCV infection, after the development of cirrhosis (4).HCV carriers demonstrate higher rates of developing HCC than HBV carriers (5); the incidence of HCC in HCV carriers is 4.2-fold higher than in HBV carriers, and it is 572-fold higher than in normal individuals (6).Therefore, prophylaxis of HCC occurrence from chronic viral hepatitis is likely as important as the progress of HCC therapy.
Glycyrrhizin (GL), a molecule of glycyrrhetic acid (GA) and two molecules of glucuronic acid, is an aqueous extract from licorice root (7).This compound is well known for its anti-inflammatory properties in Chinese medicine (8).GL and the GA derivatives exhibit anti-ulcerous (9), anti-viral (10)(11)(12) and interferon (IFN)-inducing activity (13).GL was proven to be useful for therapy of chronic hepatitis (14).Indeed, double blind controlled trials showed that GL improved hepatic function tests such as GOT, GPT and γ-GTP in patients with chronic active hepatitis (15,16).Thereafter, GL has often been used in treatment for chronic active liver diseases in Japan.We examined the effect of GL on hepatocarcinogenesis in mice treated with diethlynitrosamine (NDEA).An NDEA model was used because nitrate and nitrosamine synthesis is increased in viral hepatitis (17,18).Therefore, a NDEA model may be useful as a model of hepatocarcinogenesis due to viral hepatitis.

Chemicals
NDEA was purchased from Wako Pure Chemical Industries (Osaka, Japan).GL was kindly provided from Minophagen Pharmaceutical (Tokyo, Japan).

Animals
Five-week-old male BALB/c mice, weighing 22-23g, were purchased from Charles River Japan (Yokohama, Japan).The mice were fed a CE-2 diet (CLEA Japan, Tokyo, Japan) and were given water ad libitum.Light and dark alternated at 12 h intervals.All animals received humane care in compliance with the guidelines of our university.

Experimental procedure
Ninety-seven BALB/c mice were divided into two groups, GL (n ϭ 50) and C (n ϭ 47).The experimental schedule is shown in Figure 1.In brief, mice in the GL group intramuscularly received 2 mg/mouse of GL 3 days a week, and mice in the C group received the same volume (0.1 ml) of saline in the same way (Figure 1A).Two weeks after the start of administration of GL or saline, an i.p. of 75 mg/kg body wt of NDEA was given weekly for 3 weeks (19).Then, 100 mg/kg body wt of NDEA was weeky given for the following 3 weeks.Thirty additional mice without NDEA treatment were divided into two groups, GC (n ϭ 15) and SC (n ϭ 15).Mice in the GC group intramuscularly reveived 2 mg/mouse of GL 3 days a week, and mice in the SC group received 0.1 ml of saline in the same way (Figure 1B).Mice in the four groups were killed every 5 weeks from 7 to 25 weeks.The numbers of mice killed in the GL group at 12, 17, 22, 27 and 32 weeks were 9, 9, 9, 9 and 14, respectively, and those in the C group at 12, 17, 22, 27 and 32 weeks were 9, 9, 9, 9 and 11, respectively.The number of mice in the GC and SC

Histology
The morphology of hepatocellular neoplasms was classified by cell size, tinctorial properties of the cytoplasm, pattern of growth and other morphological features as previously described (20).Two independent investigators performed the histological assessment in a blind manner.The liver sections from the four groups were randomly shown to the investigators.

Statistics
All data were expressed as means Ϯ SEM.The difference between the two groups was determined by the Mann-Whitney U-test and the χ 2 test.
Significance was set at P Ͻ 0.05.

Body weight and liver weight
The changes in body weight are shown in Figure 2. Body weights increased with time except that the body weight in the C group decreased after 22 weeks.Body weights of the GC group were greater than those of the GL group throughout the experiment (P Ͻ 0.05, each).Body weights of the C group were lower than the SC group at 12, 17, 27 and 32 weeks (P Ͻ 0.05, each).Mice in the GL group had similar body weights to the C group throughout the experiment.Liver weights of the GL group at 12 and 32 weeks were higher than those of the C group at the corresponding weeks (data not shown).Although liver weight/body weight (LW/BW) gradually decreased with time in the SC and GC groups, it gradually increased in the GL and C groups (data not shown).At 32 weeks, LW/BW in the GL group was significantly higher than in the GC group (P Ͻ 0.05, data not shown) and the C group was also higher than the SC group (P Ͻ 0.05; data not shown).

Liver function tests
Serum bilirubin levels in the GL and C groups at 32 weeks were higher than those at 12 weeks (P Ͻ 0.05, respectively; Table I).However, no difference was observed between the GL and C groups.Serum albumin levels gradually decreased in the GL and C groups.However, the albumin decrease in the GL group was smaller than that in the C group.Serum albumin in the GL group was higher than in the C group throughout the experiment.At 22 weeks, serum albumin in the GL group was significantly higher than that in the C group (P Ͻ 0.05).Serum AST in the GL group was lower than in the C group throughout the experiment.Levels of serum AST at 12, 27 and 32 weeks in the GL group were significantly lower than those in the C group (P Ͻ 0.05, respectively).Changes in serum ALT were similar to those of AST.Serum ALT was slightly lower in the GL group, compared with the C group at 32 weeks.Levels of AFP were Ͻ1 ng/ml in all samples (data not shown).There were no changes in serum bilirubin, albumin, AST, ALT, ALP or AFP levels between the GC and SC groups throughout the experiment (Table I).

Tumor incidence
No tumors were observed until 17 weeks in the GL and C groups.At 22 weeks in the C group, a tumor which was a histologically hyperplastic nodule developed.After 27 weeks, adenoma developed in the C group.The numbers of tumors in the C group at 22, 27 and 32 weeks were 0.11 Ϯ 0.11, 0.44 Ϯ 0.34 and 1.64 Ϯ 0.36/liver, respectively.However, the number of tumors in the GL group at 32 weeks was 0.71 Ϯ 0.22/liver, which was significantly lower than in the C group (P Ͻ 0.05; Figure 3A).HCC first developed in the C group at 20 weeks.The numbers of HCC in the GL and C groups at 32 weeks were 0.29 Ϯ 0.13 and 0.82 Ϯ 0.18 /liver, respectively.There was a significant difference between the two groups (P Ͻ 0.05; Figure 3B).Histological examination at 32 weeks showed that 18 tumors in the C group were composed of nine HCCs and nine adenomas, and that 10 tumors in the GL group were composed of four HCCs and six adenomas (data not shown).Macroscopic examination of HCCs in the C group at 32 weeks is shown in Figure 4A.The color of HCCs was whitish and the tumor edges were sharp.The tumors were clearly discriminated from the surrounding tissues (Figure 4B).Cystic dilatation of bile ducts and mild fatty changes were features observed in non-cancerous tissues in this experimental model.Microscopic examination revealed that HCC had a trabecular structure, which is a typical feature of hepatocellular carcinoma (Figure 4C) (20).Although three mice in the C group at 32 weeks exhibited lung metastasis (Figure 4D), no mice in the GL group had lung metastasis.Therefore, the results suggest that this experimental model is useful to assess anti-cancerous drugs.Table II shows the incidence of HCC in the four groups at 32 weeks.The incidence rate of HCC in the GL group was 3/14 (21%), which was significantly Ͻ7/11 (64%) of the C group (P Ͻ 0.05).No tumors were observed in the GC and SC groups throughout the experiment.These data suggest that GL decreased the occurrence of HCC in mice treated with NDEA.

Discussion
HCC is one of the most common cancers worldwide (1).Although many risk factors have been reported (3), viral infection is an important factor for HCC (2,5,6).In the general population of Japan, the carrier rates of HBV and HCV are 1.3 and 1.5%, respectively.However, HBV and HCV carriers account for 20 and 70% of HCC in Japan (6), and the incidence of HCC is ~3-fold higher in chronic hepatitis C than in chronic hepatitis B (5).These data suggest that HCC develops at a fairly high rate in chronic viral liver diseases, especially in HCV carriers.Therefore, chemoprevention of HCC may be as important as therapies for HCC.IFN has been reported to be effective for reducing and eliminating HCV RNA (24).IFN has been therapeutically useful for patients with chronic hepatitis C with viral clearance rates of Ͻ40-50% (25,26).Furthermore, IFN-α (6 mU 3 days a week for 12-24 weeks) significantly decreased the incidence of HCC in cirrhotic patients (27), and responders to IFN therapy exhibited a lower cumulative incidence of HCC (28).Thus, IFN therapy decreases viral replication, resulting in a lower incidence of HCC.In addition, IFN may activate the immune system, or directly reduce cell proliferation which leads to the development of HCC or precancerous cells.This is the first experimental animal study in which GL was used as a chemopreventive agent of HCC.However, the inhibitory mechanism of HCC by GL remains unclear.It has been recently reported that long-term treatment with GL for chronic hepatitis C was effective in lowering the incidence of HCC (29).The 10 year rates of cumulative HCC incidence in patients treated with and without GL were 7 and 12%, and 15 year rates were 12 and 25%, respectively.The relative risk of HCC in patients treated without GL was 2.49-fold higher than patients treated with GL (P Ͻ 0.05, GL-treated versus control group).The authors concluded that the anti-cancerous effect of GL may be associated with its anti-inflammatory effect.Nitrosoamine elevation has been reported in viral hepatitis (17,18).Therefore, GL may decrease nitrosoamine levels in liver tissues.Furthermore, in our preliminary study, the expression rate of proliferating cell nuclear antigen (PCNA) in hepatocytes of the GL group at 7 weeks was 3-fold higher than that in the C group (data not shown).In addition, the apoptotic rate of hepatocytes in the GL group at 12 weeks was similar to that in the C group.However, it increased to 0.41 Ϯ 0.09% in the GL group at 32 weeks, which was 2-fold higher than that in the C group at the corresponding week (data not shown).Therefore, modulation of cell proliferation and apoptosis by GL may be associated with inhibition of HCC in NDEA-treated mice.The inhibitory effect of GL on hepatocarcinogenesis requires further investigation to clarify the mechanism of these findings.

Fig. 1 .
Fig. 1.Experimental schedule of the NDEA model.(A) Experimental schedules of the GL and C groups are shown.(B) Experimental schedules of the GC and SC groups are shown.

Fig. 2 .
Fig. 2. Changes in body weights in the four groups.*P Ͻ 0.05, compared with the GL group; **P Ͻ 0.05, compared with the SC group.

Fig. 4 .
Fig. 4. Gross and microscopic appearances of HCC.(A) HCCs (arrows) are whitish and their margins are clear.(B) HCC (T) and surrounding tissue (N) are shown.Cystic dilatation of bile ducts and mild fatty change are characteristic features of non-cancerous tissues (magnification ϫ40).(C) Trabecular appearance of HCC tissue (magnification ϫ200).(D) Lung metastasis in mice from the C group at 32 weeks (magnification ϫ40).

Table I .
Effect of glycyrrhizin on liver function tests in mice groups at each time point was three.Animals were killed under ether anesthesia, weighed, necropsied and examined for the presence of grossly visible lesions.The livers were removed in total, weighed and sectioned into 1-2 mm strips to examine the presence of tumors.Strips of liver were fixed in 4% paraformaldehyde for 24 h, and then embedded in paraffin.Sectioned paraffin blocks were stained with hematoxylin & eosin.
a Mean Ϯ SEM.b Glycyrrhizin plus NDEA.c Saline plus NDEA.d Glycyrrhizin only.e Saline only.f P Ͻ 0.05, compared with the C group of the corresponding week.60

Table II .
Incidence of hepatocellular carcinoma in mice at 32 weeks a Glycyrrhizin plus NDEA.b Saline plus NDEA.c Glycyrrhizin only.d Saline only.e P Ͻ 0.05, compared with C group.