Genetic Alterations in Brain Tumors Following 1,3-Butadiene Exposure in B6C3F1 Mice

Abstract

The nervous system of the B6C3F1 mouse has rarely been a target for chemical carcinogenesis in the National Toxicology Program (NTP) bioassays. However, 6 malignant gliomas and 2 neuroblastomas were observed in B6C3F1 mice exposed to 625 ppm 1,3-butadiene (NTP technical reports 288 and 434). These mouse brain tumors were evaluated with regard to the profile of genetic alterations that are observed in human brain tumors. Alterations in the p53 tumor suppressor gene were common. Missense mutations were observed in 3/6 malignant gliomas and 2/2 neuroblastomas and were associated with loss of heterozygosity. Most of the mutations occurred in exons 5–8 of the p53 gene and were G → A transitions, and did not involve CpG sites. Loss of heterozygosity at the Ink4a/Arf gene locus was observed in 5/5 malignant gliomas and 1/1 neuroblastoma, while the PTEN (phosphatase and tensin homologue) gene locus was unaffected by deletions. One of 2 neuroblastomas had a mutation in codon 61 of H-ras, while H-ras mutations were not observed in the malignant gliomas examined. Only 1 brain tumor has been reported from control mice of over 500 NTP studies. This malignant glioma showed no evidence of alterations in the p53 gene or K - and H-ras mutations. It is likely that the specific genetic alterations observed were induced or selected for by 1,3-butadiene treatment that contributed to the development of mouse brain tumors. The observed findings are similar in part to the genetic alterations reported in human brain tumors.

Introduction

Brain tumors in humans are the third most common cause of death among 18- to 35-year olds, and their incidence is increasing (Kleihues and Cavenee, 2000). With the exception of familial cancer syndromes (Li-Fraumeni) involving the nervous system and irradiation, the etiology of brain tumors is still largely unknown (Kleihues and Cavenee, 2000). Epidemiological studies have shown that increased risk of brain tumor development is associated with certain occupations, such as farming, dentistry, fire fighting, metal works, and the rubber industry, but specific exposures or causative environmental agents have not been identified (Thomas et al., 1986; Kaplan et al., 1997). Malignant glioma is the most common primary malignant brain tumor in adults, while neuroblastoma, a tumor of the postganglionic sympathetic nervous system, is the most common extracranial solid tumor of children (DeAngelis, 2001). Malignant gliomas appear to develop by a stepwise accumulation of genetic lesions (Holland, 2001; Schwab et al., 2003). Genetic alterations frequently detected in human gliomas, include inactivation of tumor suppressor genes such as TP53 (Rasheed et al., 1994), retinoblastoma (RB) (He et al., 1995), Ink4a/Arf (He et al., 1995), PTEN (phosphatase and tensin homologue) (Li et al., 1997), activation of growth-promoting pathways e.g., amplification of CDK4 (He et al., 1995), and amplification or mutation of the epidermal growth factor receptor (EGFR) (Ekstrand et al., 1994).

Human neuroblastomas exist as 3 nonoverlapping subtypes exhibiting distinct genetic and morphological features (Lastowska et al., 2001). Chromosome 1p deletion and 17q gain represent early events in transformation, while N-myc gene amplification is a late event contributing to tumor progression (Tonini and Romani, 2003).

Ten of nearly 500 National Toxicology Program (NTP) studies resulted in equivocal neurocarcinogenic effects in the F344 rat brain (Melnick and Huff, 1992; Sills et al., 1999). However, the nervous system of the B6C3F1 mouse has never been a target for chemical carcinogenesis in NTP bioassays (Sills et al., 1999). Only 1 spontaneous malignant glioma has been found in 1100 chamber control male mice in the NTP studies (Haseman et al., 1999).

The chemical 1,3-butadiene is an industrial monomer used in the production of rubber and plastics, and is also generated during the combustion of fossil fuels and is thus present in the environment (Himmelstein et al., 1997). 1,3-butadiene has been detected in main and side stream cigarette smoke (Brunnemann et al., 1990). Inhalation exposure of 1,3-butadiene to Sprague–Dawley rats did not induce brain tumors (Owen et al., 1987). However, the development of malignant gliomas and neuroblastomas in B6C3F1 mice exposed to 1,3-butadiene provided the first opportunity to compare genetic alterations in chemically induced mouse brain tumors with those of humans (NTP, 1984, 1993). Exposure of mice with 1,3-butadiene also caused increased incidences of tumors in multiple organ systems and mutations in tumor suppressor genes and oncogenes were often detected (Hong et al., 2000; Walker and Meng, 2000; Zhuang et al., 2000; Sills et al., 2001; Zhuang et al., 2002).

Here we describe the genetic analysis of brain tumors from 2 NTP studies in which B6C3F1 mice were exposed to 1,3-butadiene (NTP, 1984, 1993). Candidate genes commonly altered in human brain tumors (p53, Ink4a/Arf, PTEN) were analyzed. In addition, genetic alterations in the K-ras and H-ras genes were examined. The results of this study will provide further insights into neurocarcinogenesis, including the potential role of environmental chemical exposure as a causative agent in the development of human brain tumors.

Materials and Methods

Brain Tumors

Male and female B6C3F1 mice, 6–8 week of ages, were exposed to 200, 312, 625, or 1250 ppm 1,3-butadiene by inhalation 6 hour/day, 5 days/week for 13, 26, 40, or 60 weeks (NTP, 1984, 1993). Microscopically, 6 malignant gliomas and 2 neuroblastomas were observed in treated mice (Table 1). A total of 9 brain tumors (including 1 spontaneous tumor from another NTP study) were fixed in 10% neutral buffered for-malin, routinely processed, embedded in paraffin, sectioned to a thickness of 5 μm, and stained with hematoxylin and eosin. Subsequently, 6 unstained serial sections, 10 μm thick, were prepared from each paraffin block containing the brain neoplasms for isolation of DNA for PCR-based assays.

Immunohistochemistry on Paraffin Sections

After deparaffinization, sections were incubated for 20 minutes in methanol 2% H₂O₂, then rinsed 3 times in PBS buffer. For antigen retrieval, sections were treated for 4 minutes in a pressure cooker using 10 mM citrate buffer (pH 6.0) then cooled in PBS for 20 minutes. The sections were incubated with normal goat serum (1/20 in PBS) for 30 minutes in order to reduce nonspecific binding followed by application of the primary antibody (NCL-p53-CM5p, Novocastra Laboratories Ltd, Newcastle, UK, diluted 1/500) at room temperature for 60 minutes. Secondary antibodies with standard incubations were performed using Vectastain Elite Vector Rabbit kit (Ig-6101, Vector Lab Inc, Burlingam, CA, USA). The chromogen used was diaminobenzidin tetrahydrochloride (0.05 g in Tris buffer H₂O₂). Sections were counterstained with Gill’s hematoxylin for 30 seconds, rinsed 3 times in PBS for 5 minutes, dehydrated in ethanol xylene gradients, and mounted using Eukitt solution.

DNA Isolation, Amplification, and Cycle Sequencing

The methods for DNA isolation and sequencing have been described (Hong et al., 2000). Briefly, DNA was isolated from paraffin-embedded sections containing brain neoplasms using a DNAeasy tissue kit (QIAGEN Inc., Valencia, CA). DNA was amplified by the polymerase chain reaction (PCR), using nested primers for K - and H-ras. Touch-down PCR was performed for exons 5 to 8 of the p53 gene, and PCR/sequencing primers and annealing temperature profile of the cycles have been reported previously (Hong et al., 2000). Normal controls for K - and H-ras, or p53genes and no DNA controls were run with all sets of reactions. To identify mutations, samples were sequenced utilizing a cycle sequencing kit (U.S. Biochemical, Cleveland, OH), which incorporates α-³³P-dideoxynucleotide (ddNTP) terminators (A,C,G,T) into the sequencing products. Prior to sequencing, PCR products were purified using a QIAquick Gel Extraction Kit (QIAGEN, Valenicia, CA). The amplification primers also served as sequencing primers. Mutation identification was confirmed with 2 amplification reactions from original DNA.

Analysis for Loss of Heterozygosity (LOH) in Mouse Brain Tumors

Allelotype analysis was performed as previously described (Klein et al., 2000). Brain tumors were scraped from paraffin-embedded sections using a clean razor blade without taking normal brain. Gliomas and neuroblastomas were digested in 50 μl Tail Buffer (50 mM KCl, 10 mM Tris/HCl pH 9.0, 0.45% Nonidet P-40, 0.45% Tween 20) in the presence of 0.1 mg/ml Proteinase K (Roche, Indianapolis, IN) for 5 hours at 55°C at 400 rpm. Proteinase K was inactivated by heating samples for 15 minutes at 85°C. Two μl of DNA were used for LOH analysis performed by PCR. Reactions were standardized as follows: denaturation was performed for 2 minutes at 94°C followed by 30 cycles (40 seconds at 94°C, 50 seconds at 55°C, 30 seconds at 72°C), and 10 minutes at 72°C for final extension in a thermal cycler (GeneAmp PCR 9700, Applied Biosystem, Foster City, CA). Markers used for LOH analysis were D11Mit320 (ca 0.3 cM from p53), D19Mit19 (ca 1.7 cM from PTEN), and D4Mit27 (ca 7.7 cM from Ink4a/Arf ) 〈http://www-genome.wi.mit.edu〉; 〈http://www.informatics.jax.org〉.

Results

Six malignant gliomas and 2 neuroblastomas from B6C3F1 mice exposed to 1,3-butadiene were detected at necropsy (NTP, 1984, 1993). All of the malignant gliomas occurred in the anterior or olfactory lobe of the cerebrum (Figure 1a). Microscopically, gliomas were well demarcated, densely cellular masses and effaced the neuropil of the olfactory lobe. Tumor cells invaded the adjacent parenchyma of the cerebrum (Figure 1b). Malignant gliomas consisted of pleomorphic and poorly differentiated glial cells with giant nuclei (Figure 1c) and were characterized by frequent mitotic figures. Areas of necrosis and blood vessels with proliferative endothelial cells were evident. Neuroblastomas were also detected in the olfactory lobe. The tumors were of high cellular density and invaded the adjacent neuropil. Neoplastic cells had small, round to oval hypo- and hyperchromatic nuclei and scant fibrillar cytoplasm (Figure 1d). True rosettes consisting of round cells surrounding a lumen, or pseudorosettes of cells arranged around cores of cytoplasmic fibrils or blood vessels were occasionally observed in the tumors. The morphologic characteristics of malignant gliomas and neuroblastomas are consistent with those reported for humans (Kleihues and Cavenee, 2000). With immunohistochemical staining, 3 of 5 malignant gliomas and both neuroblastomas exhibited nuclear accumulation of p53 protein (Figure 1e).

The 6 malignant gliomas and 2 neuroblastomas from B6C3F1 mice exposed to 1,3-butadiene were analyzed for genetic alterations in the p53, K -, and H-ras genes. The pro-file of genetic alterations is summarized in Table 2. Missense mutations in p53 exons 5–8 were detected in 3 of 6 malignant gliomas and in both neuroblastomas. Mutations in p53 were identified at codons 132, 170, 192, and 263 and consisted mostly of G → A transitions (5/6) (Figure 2). Furthermore, LOH at the p53 gene locus was observed in 4 of 5 malignant gliomas and both neuroblastomas. All of the tumors exhibited loss of the C3H (H) allele with the exception of 1 tumor with partial loss of the C57 (B) allele (Figure 3). LOH was further evaluated in the vicinity of several genes implicated in the development of human brain tumors including PTEN and Ink4a/Arf. All of the malignant gliomas and neuroblastomas examined displayed loss of the C57 (B) allele at the Ink4a/Arf gene locus. However, no LOH was observed near the PTEN gene locus in any of the tumors examined. One neuroblastoma had a CAA → CGA mutation in codon 61 of the H-ras gene, while no K-ras mutations were identified. In contrast to the chemically induced brain tumors, the spontaneous malignant glioma was negative for mutations in K-ras, H-ras, and exons 5–8 of the p53 gene. Likewise, no p53 protein expression was detectable in the spontaneous brain tumor by immunohistochemical staining.

Discussion

Despite the equivocal increase of brain tumors determined by stastistical analysis, the location and specific genetic alterations observed in our study suggest that brain tumors in mice exposed to 1,3-butadiene may be chemically induced. In addition, the genetic alterations are similar to the major molecular pathways observed in human brain tumors and suggest a possible role of environmental exposure in human neurocarcinogenesis. The predominant G → A transition in the p53 gene in these mouse brain tumors suggests that treatment of 1,3-butadiene may have caused DNA damage as has been shown for other 1,3-butadiene induced tumors, although the mutation profiles are different in several tumors (Wiseman et al., 1994; Zhuang et al., 1997; Hong et al., 2000; Sills et al., 2001; Zhuang et al., 2002). The G → A transition appeared to be consistent with the mutational specificity of alkylating agents, including 1,3-butadiene, that frequently cause G:C → AT transition mutations, suggesting that they may arise from the predicted miscoding of guanine-N7-adducts or from cross linking of DNA strands (Trukhanova et al., 1998; Melnick, 2002). Alternatively, transition mutations could be spontaneous rather than as a consequence of interaction with environmental carcinogens, as suggested in sporadic and familial human astrocytic tumors (Kleihues et al., 1995). In some studies, up to 67% of human secondary glioblastomas had TP53 mutations, and the majority of the mutations were G → A transitions (Watanabe et al., 1996; Fulci et al., 2000). All tumors evaluated for LOH in the vicinity of Ink4a/Arf gene locus lost the C57BL/6 allele. The preference for loss of the C57BL/6 allele of Ink4a/Arf is expected if p16 ^Ink4a has an important tumor suppressing function for the development of such brain tumors. The C3H allele of this gene is known to code for a p16 ^Ink4a protein that has been shown to be defective in its function to inhibit CDK4 due to 2 amino acid changes in exon 2 (Herzog et al., 1999; Zhang et al., 2002). Despite the small sample number, this observation supports the notion that p16 ^Ink4a plays a key role in the development of mouse brain tumors as has been suggested in other studies (Bachoo et al., 2002).

An increased risk of human brain tumor development has been associated with certain occupations (Thomas et al., 1986; Kaplan et al., 1997). However, specific causative environmental exposures with the exception of therapeutic X-irradiation have not been identified. Because of the extremely low rate of spontaneous brain tumors in B6C3F1 mice, the effects of 1,3-butadiene in the present study were considered equivocal, suggesting that the marginal increase in brain tumors may be related to 1,3-butadiene treatment (Sills et al., 1999). The mouse brain tumors examined were present in the anterior or olfactory lobes and the genetic alterations observed in the present study were most likely associated with hematogenous entry of 1,3-butadiene metabolites through the blood–brain barrier. It was not possible to determine whether 1,3-butadiene caused early genetic events in brain tumor induction or simply accelerated the development of spontaneous brain lesions. However, the predominant pattern of p53 transition mutations in mice exposed to 1,3-butadiene for only 13 weeks suggests that p53 mutations may represent an early event.