Abstract
To estimate potential human risk of exposure to a food-derived, genotoxic hepatocarcinogen, 2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline (MeIQx), a 2-year carcinogenicity test was conducted using male F344 rats administered MeIQx-containing diet at doses of 0 (control), 0.001, 1, and 100 ppm. The lowest dose 0.001 ppm was established as equivalent to the daily intake of this carcinogen in humans (0.2 to 2.6 μg/man/day). Significant decreases of survival rate and body weight gain were observed in rats treated with 100 ppm MeIQx. Histopathological examination revealed significant induction of hepatocellular carcinomas, adenomas, and development of glutathione S-transferase placental form–positive foci with MeIQx at 100 ppm. Moreover, the incidences of Zymbal’s glands carcinoma, mammary fibroadenoma, and subcutaneous fibroma were found significantly increased in a 100 ppm MeIQx group. However, no significant induction of altered preneoplastic hepatocellular foci was observed in 0.001 and 1 ppm groups as compared to the controls. 8-Hydroxy-2’-deoxyguanosine levels in the rat liver DNA of the 100 ppm-treated group were not elevated, but MeIQx-DNA adduct formation increased as compared with the 1 ppm case, albeit without significance. No significant induction of any other neoplastic lesions related to the carcinogen administration was found in MeIQx-administered groups except for 100 ppm. These results imply that 1 ppm may be a no-effect level for MeIQx carcinogenesis.
Introduction
Heterocyclic amines (HCAs) detected in cooked or heated meat or fish are known to have mutagenic potential (Sugimura et al., 2004) and show carcinogenicity not only in rodents but also in monkeys (IARC, 1992; Adamson et al., 1995). It can be surmised that carcinogenicity is also exerted in man and this group of agents is designated as possible carcinogens in the IARC carcinogen classification (IARC, 1992). Therefore, for cancer prevention, avoidance of exposure to HCAs is highly recommended (Sugimura et al., 2004).
2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline (MeIQx) is one of the HCAs being first reported as a potent mutagen in fried beef (Kasai et al., 1981). The mutagenicity of MeIQx in Salmonella typhimurium TA98 and TA100 with S9 mix was reported to be the strongest of all HCAs examined (Sugimura et al., 2004). MeIQx is metabolized in vivo to DNA-reactive metabolites, which form DNA adducts considered important for HCA carcinogenesis in various organs (Langouet et al., 2001; Turteltaub et al., 1997; IARC, 1992). Actually, 1.8 to 18 DNA adducts per 1010 nucleotides have been detected in human organs such as the colon (Totsuka et al., 1996). In rats, MeIQx was reported to induce tumors of liver, lung, Zymbal’s gland, clitoral gland, large intestine, oral cavity, mammary gland, and skin. Furthermore, tumorigenicity of MeIQx after its administration at relatively high dose levels (≥100 ppm) was observed regarding mice liver, lung, forestomach, and large intestine, as well as induction of lymphomas and leukemias (Sugimura et al., 2004; IARC, 1992). Chronic administration of MeIQx is reported to result primarily in carcinomas of the liver in rodents (Kato et al., 1988).
It has been generally accepted that genotoxic carcinogens have no threshold in exerting their potential for cancer induction. However, the nonthreshold theory can be challenged for cancer risk assessment in humans. For example, a low-exposure study of diethylnitrosamine questioned whether the effects of carcinogens at high doses can be quantitatively extrapolated to those at low doses (Williams et al., 1993). Recently we have found that DNA adduct formation is elevated linearly in proportion to a rise in the MeIQx dose (Fukushima et al., 2002). However, no increase of formation of 8-hydroxy-2’-deoxyguanosine (8-OHdG), which is the most abundant DNA adduct associated with oxidative stress and resulting in specific types of mutation, as well as induction of glutathione S-transferase placental form (GST-P)–positive foci, considered the rat liver preneoplastic lesions, was observed at (≤10 ppm) of MeIQx (Fukushima, 1999; Fukushima et al., 2002, 2003). Although the number of GST-P-positive foci at 1 ppm was not different from control, it was increased at 10 ppm, albeit without statistical significance. Thus, in the case of exposure to genotoxic carcinogens at low doses, no-effect levels may exist for parameters relevant to carcinogenicity. Therefore, we hypothesize that a no-effect level should exist for cancer induction by MeIQx. Since the
Materials and Methods
Animals
The animals used in the present study were handled in accordance with recommendations of the Guide for the Care and Use of Animals (Institute of Laboratory Animal Resources, 1996). Four-week-old male F344/DuCrj rats were purchased from Charles River Japan, Inc (Hino, Japan) and housed 3 rats to a plastic cage (RT type, Charles River Japan, Inc) with paper chips for bedding in a room maintained under a 12-h (07:00–19:00) light-dark cycle, at a constant temperature of 25°C ± 1° C and a relative humidity of 55% ± 10%. All rats were given pelleted MF diet (Oriental Yeast Co, Tokyo, Japan) and water ad libitum up to 6 weeks of age, and then were used for the experiments. Mortality and general condition were checked daily.
Chemical
The carcinogen, MeIQx (purity, 99.9%), was purchased from the NARD Institute, Ltd (Nishinomiya, Japan).
Experimental Design
Two-hundred-four rats were randomly divided into 4 groups, receiving MeIQx at doses of 0 (group 1, a control), 0.001 (group 2), 1 (group 3), and 100 ppm (group 4) in pelleted MF diet, continuously. The lowest level, 0.001 ppm, of MeIQx was established as equivalent to the daily intake of this carcinogen in humans (0.2 to 2.6 μg/man/day) (IARC, 1992). Pelleted MF diets containing MeIQx were prepared by Oriental Yeast Co. MeIQx concentration in each diet was confirmed by high-performance liquid chromatography (HPLC), and the actual measurement value of MeIQx concentration in food was within ±5% as compared with nominal concentration. The total observation period was 104 weeks. Body weights were measured monthly along with the amount of food consumed on a per-cage basis. At the end of the experiment, all rats were euthanized under anesthesia with diethyl ether for the pathological examination.
Pathological Examination
At necropsy, after macroscopic observation, the livers were removed, weighed, processed for fixation with 10% phosphate-buffered formalin (pH 7.4), routinely embedded in paraffin, sectioned, and stained with hematoxylin and eosin (H&E) for light microscopic examination. All blocks included the left lobe and median lobe as well as nodules observed macroscopically in order to carry out quantitative analysis of altered hepatocel-lular foci (AHF) including basophilic, eosinophilic, clear, vacuolated, and mixed-cell foci according to the criteria reported previously (Harada et al., 1989). Nodules/masses in other organs including the Zymbal’s gland, subcutis, and colon, in which tumors are reported to be induced by MeIQX (Kato et al., 1988; Sugimura et al., 2004; IARC, 1992), were recorded macroscopically and examined histopathologically.
Assessment of GST-P-Positive Foci, 8-OHdG, and MeIQx-DNA Adduct Levels in the Livers
As described previously (Fukushima et al., 2002), liver sections of the left and median lobes were routinely processed for immunohistochemical staining of GST-P (with anti-rabbit GST-P antibody; MBL Co, Ltd, Nagoya, Japan) by the avidin-biotin peroxidase complex (ABC) method using Vectastain Elite ABC kit (Vector Laboratories, Burlingame, CA). Numbers of GST-P-positive foci comprising 2 or more positive hepatocytes were counted under a light microscope. Total areas of sections were measured using a color image processor (IPAP Sumika Technos, Osaka, Japan), and numbers of foci per square centimeter of liver tissue were calculated.
The 8-OHdG formation in liver specimens was measured by a previously described method (Nakae et al., 1995) in macroscopically normal tissue. The number of samples/group assayed was 5. 8-OHdG levels were determined using the calibration from HPLC runs of standard samples containing known amounts of authentic 8-OHdG and dG, and expressed as the number of 8-OHdG per 105 dG.
The levels of MeIQx-DNA adducts in groups 3 and 4 were measured by the 32P-postlabeling method under modified adduct intensification conditions using frozen samples, as previously reported (Totsuka et al., 1996). The number of samples/group assayed was 5. Data are mean ± SD values for 3 samples per group and 3 independent experiments.
Data Evaluation
Statistical analysis of the incidence of histopathological lesions was performed with the Fisher’s exact probability test (program by Shionogi & Co, Ltd, Osaka, Japan). The rest of the data were evaluated with the Dunnett’s t test (program by Shionogi Co, Ltd).
Results
General Observations
Survival curves and rates are presented in Figure 1 and Table 1. Survival in the high-dose group (group 4) deteriorated from week 59 due to tumor development, and the survival rate was only 27% at week 92. Therefore, all survivors in this group were killed at this time point. Mean survival periods for control (group 1), low (group 2), and middle (group 3) dose groups did not differ. Data for rat body weights, liver weights, and intakes of food and MeIQx are summarized in Table 2. Mean body weights of the high-dose group at week 92 were approximately 21% lower than that of the control group (Table 2). At the scheduled sacrifice, mean body weights of rats in the other 3 groups (groups 1–3) did not differ statistically. Relative liver weights in the middle- and high-dose groups were significantly higher than in the control. The first unscheduled necropsy was performed with a rat of group 4 at week 59, and a Zymbal’s gland tumor was found. All remaining F344 rats that survived beyond the 59-week time point were included as the effective numbers. MeIQx intake increased in line with the dose.
Development of GST-P-Positive Foci, AHF, and Liver Tumors
Table 3 summarizes data for GST-P-positive foci and the incidences of liver tumors. The numbers of GST-P-positive foci per unit liver area in group 4 was significantly higher than that of groups 1 to 3. In group 4, larger GST-P-positive foci could be easily recognized macroscopically. Furthermore, hepatocel-lular carcinomas and adenomas were observed in rats given 100 ppm (group 4). The incidence of AHF in group 4 was significantly different from that of groups 1 to 3, independent of whether eosinophilic, clear, vacuolated, or mixed cell foci were counted (Table 4). Namely, the incidences of eosinophilic and vacuolated foci significantly increased, and the incidences of clear and mixed foci decreased.
Rat Liver 8-OHdG Levels
The values for 8-OHdG in the rat liver (mean ± SD) of groups 1, 2, 3, and 4 were 0.92 ± 0.14, 0.81 ± 0.18, 0.94 ± 0.30, and 0.78 ± 0.16, respectively. No significant differences in 8-OHdG formation were observed among the groups.
MeIQx Adduct Formation
The MeIQx-DNA adduct levels in the rat liver DNA (mean ± SD) of groups 3 and 4 were 10.8 ± 10−7 × 0.38 × 10−7 and 37.9 × 10−7 ± 31.1 × 10−7, respectively, with no significant differences between the 2 groups.
Results of Histopathological Examination in Other Organs
Nodules/masses considered to be related to the influence of MeIQx were noted in the Zymbal’s glands of 0, 1, 1, and 15 rats of groups 1–4, respectively, and in the subcutis of 10, 10, 9, and 30 rats. In group 4, most nodules developed on the dorsal thoracic and abdominal skin. The Zymbal’s gland masses were carcinomas, whereas the majority of masses in the sub-cutis were mammary fibroadenomas and fibromas. One colon mass observed in a group 4 rat was an adenoma. The other neoplastic lesion noted was leukemia, which was associated with splenomegaly and liver infiltration by malignant large granular lymphocytes in 3, 3, 5, and 8 rats of groups 1–4, respectively.
Discussion
The lowest level, 0.001 ppm, of MeIQx was established as equivalent to the daily intake of this carcinogen by humans (0.2 to 2.6 μg/man/day) (IARC, 1992). This human dose level is 0.003 to 0.043 μg/kg/day if the body weight of man for calculation is considered to be 60 kg. In this study, the mean MeIQx intake in rats of the lowest dose group was 0.040 μg/kg/day. Since this MeIQx intake level was approximately equivalent to the maximum daily intake in humans, we consider that the lowest dose of 0.001 ppm is appropriate to estimate potential human risk of MeIQx.
The present study demonstrated that low doses of MeIQx lack evident hepatocarcinogenic activity in F344 rats, in clear contrast to the high dose of 100 ppm. A previous report indicated that in case of exposure to genotoxic carcinogens at low doses, different no-effect levels may exist for different parameters relevant to carcinogenicity (Fukushima et al., 2002). Thus, while formation of MeIQx-DNA adducts was increased even at extremely low dose levels (<0.01 ppm), significant generation of 8-OHdG was only observed at 0.01 ppm and significant induction of GST-P-positive foci only at 100 ppm in the livers of rats given MeIQx for 16 weeks (Fukushima et al., 2002). Subsequently, Hoshi et al. showed the no-effect level for in vivo liver mutagenicity of MeIQx in the diet for 16 weeks to be 1 ppm in Big Blue rats (Hoshi et al., 2004). The dose causing in vivo genotoxicity was thus lower than the level for GST-P-positive preneoplastic lesion induction (100 ppm or perhaps 10 ppm) (Fukushima et al., 2002; Hoshi et al., 2004). Moreover, no initiation activity was noted at doses of 0.001–1 ppm when using an in vivo rat medium-term bioassay for detection of initiating activity featuring application of phenobarbital, a well-known hepatopromoter (Fukushima et al., 2003). Therefore, we proposed that no-effect levels should exist for cancer induction by MeIQx (Fukushima et al., 2002). The present study showed no induction of any preneoplastic or neoplastic lesions in any organs including liver of rats treated with MeIQx at 0.001 and 1 ppm.
The present value of number and area GST-P-positive foci of the control group was relatively high and close to the data observed after exposure to MeIQx at 100 ppm after 16 weeks exposure (Fukushima et al., 2002). The incidence of GST-P-positive foci in untreated rats tends to increase depending on observation period (week of age) (Fukushima et al., 2002). Similarly, it was reported that AHF increased with age in untreated aging rats of the F344 strain (Harada et al., 1989). In this study, the number of GST-P foci and the occurrence of eosionophilic type of AHF in the 100 ppm group were significantly higher than those in other groups, in line with the tumor induction. Indeed, it was reported that GST-P-positive foci are available for identification of eosinophilic and clear cell foci in the literature (Narama et al., 2003). Stepwise accumulation of alterations in cancer-related genes leading to malignant neoplasia is considered responsible for carcinogenesis. While it is considered that DNA adduct formation is a good marker for exposure to several carcinogens (Kang et al., 2006), in the present experiment, the results did not show significant difference between 1 and 100 ppm, although the value for MeIQx adducts with 100 ppm had a tendency for increase as compared to the 1 ppm level. The reason why DNA adduct levels were not significantly elevated might be the small number of samples (3 samples/group) examined, because a wide standard deviation, which indicates a higher degree of variability, was observed in the 100 ppm dose group. In the 16-week study, MeIQx-DNA adduct formation level in the livers of rats increased with linear relationship among the various doses (0.01 to 100 ppm) of MeIQx (Fukushima et al., 2002). Furthermore, in the short-term experiment, the difference between 1 and 100 ppm was about 100 times and the data of 100 ppm in the 16-week study did not differ greatly from that of the present experiment. In our study, 4 times higher mean value for DNA adducts was observed in the 100 ppm group as compared to the 1 ppm group. This small difference might imply that the data of 100 ppm indicates the saturation of exposure and the accumulation of adduct formation in the lower dose.
The dose threshold for induction of 8-OHdG in the rat liver DNA was earlier found to become lower in accordance with the length of treatment (Fukushima et al., 2002), so with prolonged exposure to MeIQx, no-effect levels for various parameters relevant to carcinogenicity might be lowered. Unexpectedly, the mean 8-OHdG formation values in all groups were slightly elevated as compared with the level of the untreated group in the 16-week experiment and did not differ among groups. Moreover, in animals of the 100 ppm group, the existence of liver tumors was not dependent on the level of 8-OHdG. However, the 8-OHdG level was reported not to be permanently elevated during the continuous treatment with carcinogens (Kinoshita et al., 2002). Since DNA repair is reported to be enhanced as a cellular response to chronic oxidative stress (Grishko et al., 2005), adaptation during long exposure to MeIQx might have occurred. Moreover, as indicated in the previous 28-week MeIQx treatment study
Previously, strain differences were observed in susceptibility to hepatocarcinogens and the WS/Shi strain of rats was found the most susceptible to their exposure (Murai et al., 2000). In the MeIQx carcinogenicity test using WS/Shi rats, earlier occurrence and higher incidence of liver tumors was observed in the 100 ppm group (data not shown). Furthermore, gender differences in susceptibility to liver carcinogens may also be an important factor in assessing hepatocarcinogenicity. For instance, female rats are less susceptible to hepatocarcinogenicity of MeIQx than males (Kato et al., 1988), so that it is unlikely that MeIQx may induce liver neoplastic lesions with significant incidence at doses less than or equal to 1 ppm in carcinogenicity tests in females. Moreover, incidences of other tumors, including Zymbal’s gland, clitoral gland, and skin lesions, observed in female rats are lower than the incidence of liver tumor in males (Kato et al., 1988).
In the present study, MeIQx at 100 ppm induced various nodules/masses of Zymbal’s gland, subcutis, and colon, as previously reported (Sugimura et al., 2004; IARC, 1992). However, no significant induction of those tumor types was evident in the lower dose groups of MeIQx. Certainly, the evaluation of Zymbal’s gland tumor induction is important for the confirmation of the threshold for MeIQx carcinogenicity, since this tumor was observed at low dose, but not in the control group. However, the incidence in the lower dose groups was within the NTP historical control range (0 to 4%) (National Toxicology Program, 1992). Under our experimental conditions, it appears that the Zymbal’s gland is the most sensitive organ regarding MeIQx carcinogenesis except for liver; nevertheless, in a previous report using 400 ppm of MeIQx, the incidence of Zymbal’s gland tumors was 75%, whereas the incidence of liver tumors reached 100% (Kato et al., 1988). Therefore, from our results, it might be considered that the sensitivity of liver and Zymbal’s gland for MeIQx carcinogenesis is approximately equivalent. It was also found in another 2-year MeIQx carcinogenicity test using the WS/Shi strain that Zymbal’s gland tumors were developed in the control group and the incidence was not significantly increased in lower dose groups (data not shown).
In conclusion, the present study supports our argument that a threshold, at least a practical threshold, exists for carcinogenicity of the genotoxic carcinogen MeIQx. This is of direct significance to cancer risk assessment in humans. However, further studies are necessary to establish the general concept of a threshold for carcinogenicity by genotoxic chemicals, since other geno-toxic carcinogens including 2-amino-1-methyl-6-phenylimidazo [4,5-b]pyridine, diethylnitrosamine, and dimethylnitrosamine have also shown no-effect levels in medium term carcinogenicity tests (Fukushima et al., 2002, 2004, 2005).
Footnotes
Acknowledgments
We thank Ms. M. Hosono, Ms. Y. Iwakura, and Ms. K. Kanzaki at Aburahi Laboratories, Shionogi Research Laboratories, Shionogi & Co., Ltd., and Ms. K. Touma, Ms. C. Imazato, Ms. M. Imanaka, Ms. Y. Onishi, and Ms. M. Dokoh at Osaka City University of Medical School for their expert technical assistance. This research was supported by a grant from the Japan Science and Technology Corporation, included in the Project of Core Research for Evolutional Science and Technology (CREST), and a Grant-in-Aid from the Ministry of Economy, Trade and Industry of Japan.