Altered DNA Mutation Spectrum in Aflatoxin B1-Treated Transgenic Mice That Express the Hepatitis B Virus X Protein

Abstract

Humans chronically infected with hepatitis B virus (HBV) are at further risk of liver cancer upon exposure to dietary aflatoxin B1 (AFB1), a carcinogenic product of the mold Aspergillus flavus. For the present study, we utilized double-transgenic mice (ATX mice) that express the HBV X protein (HBx) and possess a bacteriophage lambda transgene to evaluate the in vivo effect of HBx expression on AFB1-induced DNA mutations. The expression of HBx correlated with a 24% increase in mutation frequency overall and an approximately twofold increase in the incidence of G/C-to-T/A transversion mutations following AFB1 exposure. These results are consistent with a model in which expression of HBx during chronic HBV infection may contribute to the development of hepatocellular carcinoma following exposure to environmental carcinogens.

Chronic infection with hepatitis B virus (HBV) is a primary risk factor for the development of hepatocellular carcinoma (HCC) (2, 49). Additional risk factors include chemical agents, such as aflatoxin and ethanol, which can result in liver damage upon repeated exposure (7, 56, 58). Epidemiological data, as well as experimental studies utilizing transgenic animals, have indicated that chronic HBV infection and exposure to hepatocarcinogens act synergistically in the development of HCC (10, 14, 31, 33, 46, 50, 53, 57). However, the mechanism by which chronic HBV infection sensitizes an individual to carcinogen-induced liver cancer remains unknown.

The 17-kDa HBV X protein (HBx) is a nonstructural protein necessary for the establishment of hepadnaviral infection in woodchucks and presumably in all mammals (8, 59). Depending upon experimental conditions, HBx is reported to inhibit nucleotide excision repair (NER) (3, 17, 21, 32, 41), to influence apoptosis (11, 23, 39, 40, 52, 54, 55), and to induce cell cycle progression (4, 26, 27). By affecting these pathways, HBx could facilitate the mutagenic effects of hepatocarcinogens or promote the survival and proliferation of hepatocytes altered after exposure. Although transgenic mice that express HBx are more sensitive to the carcinogenic effects of the hepatocarcinogen diethylnitrosamine (DEN [10, 33, 50]), they do not demonstrate significant increases in overall mutation accumulation (32, 33). These results suggest that HBx contributes to DEN-mediated carcinogenesis by a mechanism that does not involve a dramatic inhibition of DNA repair. Recently, a study demonstrated that expression of HBx increased mutagenesis in a human liver cell line treated with aflatoxin B1 (AFB1) (17). The effect of HBx on AFB1-induced damage in vivo has not been examined.

Aflatoxins are a family of potent hepatocarcinogens produced by the mold Aspergillus flavus. A common dietary contaminant in many parts of the world, aflatoxins are metabolized in vivo by cytochrome P450 isoenzymes to form mutagenic epoxides (13). The induction of P450 enzyme levels may explain the increased AFB1 sensitivity of mice that overexpress the hepatitis B surface antigen (24). However, the levels of carcinogen-metabolizing enzymes do not appear to be elevated in mice expressing HBx (9), and it remains to be demonstrated whether HBx transgenic mice are more susceptible to the mutagenic effects of AFB1. The purpose of the present study was to evaluate the effect of HBx expression in vivo on the frequency and spectrum of AFB1-induced DNA mutations.

Generation of double-transgenic mice.

Double-transgenic mice were generated that possessed both the HBV X transgene and multiple copies of an integrated bacteriophage lambda (λ) reporter gene. Transgenic ATX mice harbor the X gene (nucleotides 1376 to 1840 of subtype adw₂) under the control of the human α-1-antitrypsin inhibitor regulatory region (29, 50). Hemizygous ATX females (ICR × B6C3) were mated with homozygous λ males (C57BL/6 Big Blue) (25) obtained from Stratagene Corporation, and the male F₁ progeny were used for this study. For simplicity, single transgenic mice harboring only λ are referred to as WT mice, while double-transgenic mice harboring both λ and ATX transgenes are referred to as ATX mice. The expression of HBx in these mice (Fig. 1) is at levels similar to that measured for woodchuck hepatitis virus X protein during chronic infection in woodchucks (10) and does not lead to any increase in liver abnormalities or the accumulation of spontaneous DNA mutations (32).

FIG. 1.

Detection of HBx in transgenic mouse liver tissue. Shown is a representative result after using immunoprecipitation and Western blot hybridization to detect HBx in the 90-day-old mice used in this study. The genotypes of all mice were determined by Southern blot hybridization as previously described (32, 33). HBx expression was clearly demonstrated in all ATX mice (lanes 1, 2, 5, 7, and 10) and absent in all wild-type mice (lanes 3, 4, 6, 8, and 9). The band shown below the 14.3 kDa (Kd) marker occurs at the dye front and represents nonspecific (NS) hybridization. No protein bands were detected for ATX and wild-type mice when nonspecific rabbit serum was substituted for rabbit anti-HBx polyclonal serum during the Western blot hybridization procedure (data not shown). IgG, immunoglobulin G.

AFB1 treatment leads to increased mutation frequency (MF) in ATX/λ mice.

Mice are naturally resistant to the effects of aflatoxins, presumably due to high levels of reduced glutathione, which reacts with and eliminates the epoxide isoform of AFB1 (12, 16, 51). To circumvent this innate resistance, male ATX and WT mice were treated with AFB1 when very young (22). At 7 days of age, male mice were given a single injection (intraperitoneal) of AFB1 (Sigma) (10 mg/kg of body weight) suspended in Tricaprylin (Sigma), as previously described (15, 22). To allow for repair of persistent AFB1-induced DNA adducts, mice were not sacrificed until they had attained 90 days of age. Upon sacrifice, tissues were paraffin embedded and coded hematoxylin- and eosin-stained sections were submitted for histological analysis by M.J.F. The remaining tissue was frozen in liquid nitrogen and stored for subsequent experiments.

The RecoverEase system (Stratagene/Biocrest), which includes no phenol or chemical extractions that might damage DNA, was used to purify liver DNA for use in the Big Blue mutagenesis assay. A functional assay measuring inactivating mutations in the lambda cII gene was used to determine the relative MF, as described previously (20). The MF was calculated as the ratio of mutant PFU to total PFU. All mutant plaques were subsequently isolated and replated under the same conditions to verify the mutant phenotype. Under the conditions used for this study (number of mice per group and number of plaques per animal screened), this assay is able to reliably discern a twofold difference in MF between animal groups.

All mice appeared normal at the time of sacrifice, and no gross or histological abnormalities were noted on microscopic examination of liver tissue. Immunohistochemistry did not reveal abnormal accumulation of either WT or mutant p53 in liver tissue sections isolated from any of the mice used in this study (data not shown). Compared to age- and gender-matched untreated control mice, AFB1-treated mice revealed a significant increase in DNA MF (Table 1; P < 0.002). These results confirm that, under the experimental conditions described, AFB1 causes DNA mutations in these mice.

TABLE 1.

Determination of MF in 90-day-old AFB1-treated transgenic mice

Treatment status	Genotype	Animal	Total PFU (105)a	Mutant PFUb	MF (10−5)	Mean ± SD (10−5)
Treated	ATX	2307	5.28	26	4.93	7.46 ± 1.97
	ATX	2308	4.42	32	7.25
	ATX	2311	3.30	29	8.79
	ATX	2315	4.05	26	6.41
	ATX	2318	3.22	32	9.95
	WT	2309	6.75	25	3.70	6.00 ± 1.38
	WT	2310	6.30	32	5.08
	WT	2313	6.10	55	9.02
	WT	2316	3.11	18	5.78
	WT	2317	2.79	18	6.44
Untreatedc						3.59 ± 0.35

^aCalculated from dilutions incubated at 37°C.

^bIsolated mutants replated and incubated at 24°C to verify phenotype.

^cMF value for untreated WT mice at 90 days of age as reported by Madden et al. (32).

Previous studies have established that HBx can inhibit the ability of cultured cells to repair damaged DNA (3, 17, 21, 32, 41), including damage by AFB1 (17). We predicted that a similar inhibitory effect of HBx in vivo would lead to an increase in the accumulation of DNA mutations in AFB1-treated mice. Comparison of the relative MFs in liver tissue samples obtained from AFB1-treated ATX (n = 5) and wild-type mice (n = 5) revealed a slight elevation (24%) in the mean MF for 90-day-old ATX mice (Table 1); however, this increase was not statistically significant (P value, >0.09). These results demonstrate that the increased MF in AFB1-treated mice is not significantly influenced by the expression of HBx.

Determination of DNA mutation spectrum.

We next considered that HBx might alter the spectrum of DNA mutations. The AFB1 carcinogen is activated by the cytochrome P450 monooxygenase system to an active AFB1-8,9-epoxide, which binds covalently to DNA, RNA, and proteins (18). If not repaired by the NER pathway, the primary DNA adduct, 8,9-dihydro-8-(N⁷-guanyl)9-hydroxy-AFB1 (AFB1-N⁷-Gua), predominately causes transversion (purine to pyrimidine or pyrimidine to purine) mutations (1, 12, 30). As HBx is reported to bind to the UV-damaged DNA binding protein (DDB1 [also known as UV-DDB]) component of NER, mutant λ cII plaques from ATX and WT mice were isolated and sequenced, to determine whether the interaction between HBx and DDB1 might influence the specific type of DNA mutation formed (28, 47, 48). Deletion mutations were found to have occurred at positions 533 to 535 in one ATX mouse and at position 583 in one WT mouse. The majority of mutations (58%) identified in ATX mice were consistent with an AFB1 etiology and were G/C-to-T/A transversions (Table 2). In contrast, only 31% of mutations identified in AFB1-treated WT mice were of this type. Instead, the majority of mutations (62%) found in treated WT mice were G/C-to-A/T transitions, a spectrum similar to that previously identified in livers of untreated animals (32). Indeed, transversion mutations were approximately twice as prevalent in ATX as in WT animals. These results indicate that ATX and WT animals respond differently to AFB1 exposure.

TABLE 2.

cII mutation spectrum in aflatoxin B1-treated mice

Mutation type	No. of specified mutations (% of total)			AFB1-treated ratsd
	ATXa	WTb	Untreatedc
Transitions
G/C→A/T	11 (30)	18 (62)	27 (54)	2 (4)
A/T→G/C	1 (3)	0 (0)	2 (4)	0 (0)
Transversions
G/C→T/A	21 (58)	9 (31)	8 (16)	35 (78)
G/C→C/G	1 (3)	1 (3)	6 (12)	5 (11)
A/T→T/A	1 (3)	0 (0)	1 (2)	1 (2)
A/T→C/G	0 (0)	0 (0)	0 (0)	0 (0)
Othere	1 (3)	1 (3)	6 (12)	2 (4)
Total isolates	36	29	50	50

^aMutant cII isolates from three 90-day-old AFB1-treated ATX animals.

^bMutant cII isolates from three 90-day-old AFB1-treated wild-type animals.

^cControl; mutation spectrum of mutant lacI isolates from untreated mouse liver tissue as reported by Ross and Leavitt (43).

^dMutation spectrum of mutant lacI isolates from adult F344 rats treated with AFB1 as reported by Dycaico et al. (12).

^eIncludes deletion and addition mutations.

Identification of mutation hot spots.

AFB1 can cause more frequent mutations at specific guanine sites, resulting in mutational hot spots (6, 35, 44). The presence of such hot spots in the human p53 gene is believed to contribute to the carcinogenic effects of chronic aflatoxin exposure (5, 19). Our sequence analysis revealed the presence of two possible AFB1 mutational hot spots in the cII gene region of the bacteriophage lambda transgene. Both of these putative hot spots occur within GCG codons (alanine), consistent with previous reports which indicate that AFB1 induces mutations preferentially at 5′-CpG-3′ and 5′-GpG-3′ dinucleotides (34, 37, 38). The first hot spot mutation was found in five of six animals (two ATX and three WT) and occurred at position 426, resulting in C/G-to-T/A transitions or C/G-to-A/T and C/G-to-G/C transversion mutations in the positive strand sequence (Table 3). Of a total of 64 mutant cII isolates sequenced, 6 had mutations at position 426. A more striking hot spot mutation was found at position 549: C/G-to-T/A transitions or C/G-to-A/T transversions in six of six animals (three ATX and three WT). Of the total of 64 mutant cII isolates, 16 were identified with mutations at this location. Although there was no apparent correlation between the expression of HBx and the incidence of mutations at these two hot spot locations, the identification of these sites validates the use of transgenic mice as an appropriate animal model for the study of AFB1 mutagenesis in vivo.

TABLE 3.

cII sequence mutations identified in AFB1-treated ATX and wild-type animals

Positiona	Mutation(s)b (occurrencec)
	ATX	WT
362	G/T (1)	G/T (1)
365	G/A (1)
371	C/T (1)
384	C/A (2)
401	G/A (1)	G/A (2)
411		G/T (1)
426d	C/T (1)	C/T (2)
	C/A (1)	C/A (1)
		C/G (1)
452	C/A (1)
459	G/C (1)
	G/T (2)
460	C/A (1)	G/T (1)
462		G/T (1)
466	G/T (5)
480	T/A (1)
497	C/A (1)
503	G/A (1)
509	C/A (1)
512		G/A (1)
516	G/T (1)
517	G/A (2)
519	G/T (2)
530	G/A (2)
543		G/A (1)
547		G/T (1)
549d	C/T (3)	C/T (10)
	C/A (2)	C/A (1)
552		C/T (1)
569	C/A (1)
619		G/T (1)
630		G/T (1)

^aPosition location based upon the cII sequence reported in reference 45.

^bType of mutation occurring in positive strand.

^cOccurrence of mutation at specified site. Number in parentheses indicates number of isolates bearing that particular mutation from a total of three ATX and three WT animals.

^dMutational hot spot, i.e., identical mutation(s) found in multiple animals.

The role of HBx in carcinogen-induced HCC.

Studies performed with transgenic mice that express the human or woodchuck HBV X protein (HBx or WHx, respectively) previously demonstrated that animals expressing the X protein are more sensitive to the effects of hepatocarcinogens such as DEN (10, 33, 50). These were particularly intriguing observations given the synergistic effect between chronic HBV infection and exposure to dietary aflatoxins in the development of liver cancer (31, 53, 57). In the present study, we demonstrate that HBx expression correlated with a modest 24% increase in MF overall and a dramatic shift in the DNA mutation spectrum following exposure to AFB1. These findings support the hypothesis that HBx is a cofactor in the development of HCC.

AFB1 is a mutagen produced by the mold Aspergillus flavus that demonstrates species-specific susceptibility (12, 16). While humans and rats are highly susceptible to the mutagenic effects of AFB1, mice are extremely resistant due to abundant levels of glutathione-S-transferase activity in the tissues of these animals (1, 51). Our demonstration of a significant increase in DNA mutations in AFB1-treated mice represents one of the few successful uses of transgenic mice to measure the mutagenic effects of AFB1 in vivo and the only successful study to date in which glutathione-depleting agents were not used (1, 22). In our study, a single dose of AFB1 suspended in tricaprylin was administered intraperitoneally at an early age (7 days) (14). Three lines of evidence indicate this treatment was successful. First, there was a significant increase in MF in both ATX and WT mice when compared to that in untreated control animals. Second, sequence analysis of mutant cII isolates showed that both ATX and WT mice experienced an increase in the percentage of G/C-to-T/A transversion mutations, with the increase much more pronounced in ATX animals. Finally, the sequence analysis revealed two potential mutation hot spots within the cII gene of bacteriophage lambda. As studies with bacteria, rats, and humans have demonstrated, the occurrence of G/C-to-T/A transversion mutations and the appearance of mutation hot spots are hallmarks of AFB1 exposure (6, 35, 44). Together, these results demonstrate that transgenic mouse lines can be used to study the detrimental effects of AFB1 and that HBx expression influenced the spectrum of mutations that occurred.

Although the effect of HBx on overall AFB1-induced MF was modest, the increased number of transversion mutations identified in ATX mice and the identification of mutation hot spots may have important physiological consequences. Indeed, inactivating transversion mutations at codon 249 (AGG) of the p53 gene are found in the majority (77%) of human HCCs isolated from geographical regions where dietary aflatoxin exposure is high (5, 19) and in AFB1-treated human cells expressing HBx (17). Although immunohistochemical staining of liver tissue sections isolated from the mice used in this study (both AFB1 treated and untreated) did not reveal WT or mutant p53 accumulation (data not shown), this may reflect a difference between species for AFB1-induced HCC. Inactivating mutations in p53 are rare to nonexistent in liver tumor samples isolated from woodchucks infected with woodchuck hepatitis virus (42). In rats, AFB1 treatment leads to a high frequency of liver tumors containing mutations in codon 12 of the ras gene (35). Additional studies will be needed to measure the incidence of tumor development in these mice and to identify mutations in key regulatory genes following chronic exposure to AFB1, a scenario that more closely resembles the lifelong dietary intake of aflatoxins by humans.

Our original prediction that ATX mice would be more sensitive to AFB1 was based on the finding that HBx binds to DDB1 (3, 28, 47, 48), a component of the NER pathway responsible for repair of AFB1 lesions (30). The idea that HBx binding to DDB1 might interfere with NER is supported by the observation of increased AFB1 mutagenesis in HBx-expressing cells (17). Our observation that HBx in ATX mice in vivo is associated with only a modest increase in MF (24%) relative to that for WT littermates has several possible explanations, including that our results are based on a single injection of AFB1 (instead of repeated dietary exposures seen in humans) and that an effect of HBx may be masked by the relative resistance of mice to AFB1.

The mechanism by which HBx can influence the incidence of transversion mutations in AFB1-treated mice remains unknown. One possible explanation is that HBx may push hepatocytes to undergo DNA synthesis prior to the complete repair of AFB1-induced bulky lesions. Indeed, a mild effect of HBx on hepatocellular proliferation has been measured in these ATX mice and is thought to be responsible for the increased MF in DEN-treated ATX mice (33). This modest effect of HBx on proliferation is similar to that previously reported for the tumor-promoting agent 12-myristate-13-acetate (TPA) (36). This would also explain the altered DNA MF found in ATX mice. A second explanation builds upon the numerous reports that HBx inhibits NER in established cell lines and primary mouse hepatocytes in response to both UV light- and AFB1-induced DNA damage (3, 17, 21, 32, 41). The increased incidence of transversion mutations observed in ATX mice may indicate that HBx alters the response of hepatocytes to a specific type of AFB1-induced DNA adduct.

In summary, we report the successful treatment of mice with the potent liver carcinogen AFB1. Following a single injection with AFB1, significant increases in DNA MF were measured in both ATX and WT mice. The specificity of AFB1 was demonstrated by DNA sequence analysis, which revealed an increase in the number of G/C-to-A/T transversion mutations and the appearance of mutation hot spots in treated mice, both of which are consistent with the effects of AFB1 in other animal species, including humans. Although expression of HBx correlated with a striking increase in the incidence of transversion mutations, the MF in treated ATX mice was only moderately higher (24%) than that measured in treated WT littermates. Additional studies will be necessary to study the deleterious effects of chronic AFB1 exposure in these animals and to elucidate the mechanism by which HBx expression contributes to the occurrence of transversion mutations. Such information will lead to a better understanding of the factors affecting the development and severity of HBV-associated liver cancer and may reveal novel insights into the prevention or treatment of HCC.