Chemical and Physical Properties

The representation of the benzene molecule has evolved from the resonance ring structures described by Kekule in 1865 to the structure in which all six carbon-to-carbon bonds are identical (see above representations; NTP, 1986).

Benzene at room temperature is a highly flammable, colorless, transparent liquid having a boiling point of 80° C and a melting point of 5.5° C (NTP, 1986; Merck, 1996). Slightly soluble in water (1.8 g/L), benzene is miscible with a variety of organic solvents and with oils. Benzene has a vapor pressure of 13 mm Hg, a flashpoint of −11° C, and a flammability limit in air of 1.5% to 8.0%. It is generally available in three grades: refined, nitration, and industrial (NTP, 1986). The differences are based on the content of nonaromatics specified (less than 0.15% for refined) and on the distillation range (less than 1° C for refined and nitration grades and less than 2° C for industrial grade).

Production, Use, and Human Exposure

Benzene is used primarily as a solvent in the chemical and pharmaceutical industries, as a starting material and intermediate in the synthesis of numerous chemicals, and in gasoline (NTP, 2004). The major United States source of benzene is petroleum. In 1994, benzene ranked 17th in production volume among chemicals produced in the United States. In 2002, 7.2 million metric tons of benzene were produced in the United States, and over 4 billion liters were imported.

The primary route of benzene exposure is by inhalation (NTP, 2004). Auto exhaust and industrial emissions account for ~20% and cigarette smoke for ~50% of human exposure (Powley and Carlson, 1999). Occupational exposure affects over 200,000 people per year (ATSDR, 1997). The major industries using benzene are those involved in rubber, paint, shoes, lubricants, dyes, detergents, drugs, and pesticides (Bauer et al., 2003).

Metabolism, Absorption, and Distribution

Mice, rats, rabbits, and humans generally metabolize benzene in a similar manner, although quantities of any one metabolite may vary by species as determined using microsome preparations from these different species (Powley and Carlson, 1999).

Metabolism of benzene by the cytochrome enzymes to phenol, catechol, and hydroquinone, as well as enzyme systems that result in ring-opened forms of benzene may all be involved in producing carcinogenic moieties of benzene (USEPA, 1998). Benzene is metabolized to benzene oxide by a cytochrome P450 multifunctional oxygenase system (CYP2E1 and CYP2B1) primarily in the liver; however, other organs may be involved (e.g., lung) (Powley and Carlson, 1999; Ross, 2000). Benzene oxide is then nonenzymatically transformed to phenol or ring-opened muconaldehyde. Muconaldehyde is metabolized to muconic acid. Phenol is further metabolized (via CYP450) to hydroquinone and catechol. Hydroquinone and catechol can then be metabolized in bone marrow by myeloperoxidase to form p-benzoquinone and o-benzoquinone (Powley and Carlson, 1999) (Figure 1).

Figure 1

Pathways in Benzene Metabolism. Glucuronidation and sulfation pathways have been omitted for clarity (Ross, 2000)

The content of peroxidases, which activate phenols to toxic quinones and free radicals, and sulfatases, which remove conjugated sulfate and thus reform free phenol, may be present at different levels in various organ systems and may help explain the different organ carcinogenic properties of benzene (USEPA, 1998). Polymorphisms in enzyme systems involved with benzene metabolism (e.g., epoxide hydrolase) may affect the amount of metabolite formed, and thus, the ultimate toxic and carcinogenic effects of benzene (Bauer et al., 2003).

Detoxification reactions include glutathione conjugation of benzene oxide and sulfate and glucuronide conjugation of phenol, hydroquinone, catechol, and trihydroxy-benzene. Following conjugation, metabolites are excreted in the urine (Powley and Carlson, 1999).

Several of the benzene metabolites may contribute to the toxic and carcinogenic effects of the chemical. Some studies suggest that p-benzoquinone is responsible for the carcinogenic effects of benzene (Irons, 1985; Powley and Carlson, 1999). Catechols and quinone metabolites can be oxidized and react with DNA to form adducts, leading to mutations and cancer (Cavalieri and Rogan, 2004; Gaskell et al., 2005). Benzene metabolites may also form adducts with proteins or interfere with gap-junction intercellular communication, which may also be involved in benzene toxicity (Rivedal and Witz, 2005; Waidyanatha and Rappaport, 2005). Other studies suggest that nitric oxide can react with benzene metabolites in the bone marrow to form nitro metabolites (nitrobenzene, nitrobiphenyl, and other nitrophenol isomers), and these metabolites may contribute to benzene toxicity (Chen et al., 2004).

Disposition and metabolism studies of C¹⁴-labeled benzene have been conducted in male C57Bl/6N mice after a single oral dose of 10 or 200 mg/kg (McMahon and Birnbaum, 1991). In 3-month-old mice, 95% of the labeled 10 mg/kg dose was eliminated in the urine within 48 hours; at 200 mg/kg, 42% was eliminated in the urine and 56% as volatiles in expired air. The major urinary metabolites were hydroquinone glucuronide, phenylsulfate, and muconic acid.

After an oral dose of benzene (0.1 mg/kg) to male Tg.AC mice, male p53^+/− mice, and their respective parent strains (FVB/N and C57BL/6), the major route of excretion was again in the urine (Sanders et al., 2001). The major metabolites recovered in the urine were hydroquinone glucuronide, phenylglucuronide, phenylsulfate, and muconic acid.

After an oral benzene dose of 1, 10, or 200 mg/kg to male F344/N rats and male B6C3F₁ mice, benzene metabolites were found in blood, urine, liver, lung, and bone marrow (Sabourin et al., 1989). The major metabolites in bone marrow were phenylsulfate and prephenylmercapturic acid, muconic acid, hydroquinone monosulfate, and phenylmercapturic acid in rats and hydroquinone, phenylsulfate, and phenylmercapturic acid in mice.

Differences in the rates of benzene metabolism were observed between male and female B6C3F₁ mice after an inhalation exposure to benzene (400 to 2,800 ppm) (Kenyon et al., 1996). Pretreatment of male mice with acetone to induce CYP2E1 enhanced the rate of benzene oxidation. The authors concluded that oxidative metabolism of benzene occurs at a faster rate in males than in females, possibly explaining why more genotoxic damage occurs in males than in females.

Toxicity

Reproductive and Developmental Toxicity

Carcinogenicity

Experimental Animals

Benzene was a multisite carcinogen in 2-year gavage studies in F344/N rats and B6C3F₁ mice (Table 1; NTP, 1986). Benzene also caused treatment-related cancers in gavage (French et al., 2001) and inhalation studies (Recio et al., 2006) in p53^+/− mice (Table 1). Benzene caused skin tumors in Tg.AC mice after dermal administration (Tennant et al., 1995; Spalding et al., 1999; Humble et al., 2005).

Table 1

Chemical-related Neoplasms in NTP Benzene Studies.

Earlier benzene cancer studies in animals have been summarized by the International Programme on Chemical Safety, Environmental Health Criteria 150 (WHO, 1993). These studies also showed that benzene was carcinogenic in rodents after inhalation exposure (0 to 1,000 ppm) and oral gavage exposure (0 to 200 mg/kg) causing neoplasms in many target organs including the forestomach, harderian gland, liver, mammary gland, Zymbal’s gland, and neoplasms within the hematopoietic and lymphoreticular systems (Snyder et al., 1980, 1982, 1984; Goldstein et al., 1982; Maltoni et al., 1983, 1988; NTP, 1986; Cronkite et al., 1989; Huff et al., 1989; Farris et al., 1993).

Microarray analyses of mouse bone marrow tissue before or after benzene exposure suggests that a critical event in toxicity is dysregulation of the p53 pathways resulting in alterations in cell cycle checkpoints, apoptosis, or the DNA repair system; these may be events leading to hematopoietic malignancies (Yoon et al., 2003).

Humans

Benzene is considered a known human carcinogen based on sufficient evidence in humans (IARC, 1982, 1987; NTP, 1986, 2004). Occupational studies provided much of the evidence of benzene’s carcinogenicity in humans, where workers are generally exposed to benzene by inhalation and at much higher levels than the general public (USEPA, 1998). The evidence for the carcinogenic effects of benzene is supported by epidemiology studies, animal data, and mechanistic research. Studies of occupational inhalation exposure to benzene associate exposure with acute nonlymphocytic leukemia, preleukemia, and aplastic anemia. Using a linear dose-response curve, the USEPA (1998) estimated that the inhalation leukemia risk of 1 ppm of benzene was 7.1 × 0⁻³ to 2.5 × 10⁻². This risk estimate is based on benzene exposure in Pliofilm rubber workers at three plants in Ohio (Rinsky et al., 1981, 1987).

The National Cancer Institute and the Chinese Academy of Preventive Medicine reported on cancer outcome in 74,828 benzene exposed workers from 1972 to 1987 at 672 factories in 12 Chinese cities (USEPA, 1998; Hayes et al., 2001). While the workers in this study may have been exposed to chemicals other than benzene, this study provides additional data to show that benzene exposure is associated with hematotoxicity and leukemia.

Benzene effects on bone marrow cells with subsequent development of cancer may be due to the interactive effects of multiple genotoxic benzene metabolites, although epigenetic effects of benzene may also contribute to the cancer process (USEPA, 1998). Smith (1996) outlined proposed mechanisms in benzene-induced leukemia (Figure 2).

Figure 2

A Mechanistic Hypothesis of Benzene-Induced Leukemia (Smith, 1996).

Genetic Toxicity

The genetic toxicity test data for benzene have been reviewed (Dean, 1978; NTP, 1986; IARC, 1987; Waters et al., 1988; WHO, 1993), but no clear consensus on the mechanism of benzene-induced chromosomal damage and carcinogenesis has been reached (Eastmond, 2000; Whysner, 2000). Benzene is not active in bacterial gene mutation assays (Zeiger and Haworth, 1985), mammalian gene mutation assays using mouse lymphoma L5178Y/tk^+/− cells (Myhr et al., 1985), or germ cell mutation assays in male Drosophila melanogaster (Foureman et al., 1994). However, under proper conditions of activation, it has been shown to induce mutations (Tsutsui et al., 1997), DNA damage, sister chromatid exchanges, and aneuploidy in some in vitro mammalian cell systems (Gulati et al., 1989; Tsutsui et al., 1997). IARC (1987) also reported positive results in chromosomal aberration studies with benzene in vitro. Although some sporadic positive responses have been noted in mutagenicity tests in vitro, the mutagenicity of benzene is better studied in vivo because its metabolic pathways are complex and a variety of active metabolites are produced throughout its biotransformation. The literature describing benzene’s ability to induce chromosomal damage in vivo in a variety of test systems is extensive. The earlier studies have been reviewed, as previously cited. More recent investigations have shown that benzene induces micronuclei in peripheral blood, bone marrow, and spleen cells of mice exposed via gavage (Choy et al., 1985; MacGregor et al., 1990; Chen et al., 1994) and in lung cells of mice exposed to a single inhalation dose (Ranaldi et al., 1998). Although bone marrow cell micronuclei in these studies originated predominantly from chromosome breakage, the majority of micronuclei seen in splenocytes, as indicated by fluorescent centromeric probes, arose from aneuploidy events (Chen et al., 1994); micronuclei in lung cells resulted from both chromosomal breakage and aneuploidy events (Ranaldi et al., 1998). Benzene also induced chromosomal aberrations in mouse bone marrow cells (Tice et al., 1980) and lymphocytes (Rithidech et al., 1987) following inhalation exposure, and increased chromosomal breakage was observed in differentiating spermatogonial cells of CD-1 mice given benzene as a single oral dose (Ciranni et al., 1991). In humans, increases in numerical and structural chromosomal damage in lymphocytes of benzene-exposed workers have been well documented (Smith, 1996; Marcon et al., 1999; Zhang et al., 1999; and Kašuba et al., 2000).

Background on Genetically Altered Mice

The CDKN2A genetic locus contains two important tumor suppressor genes located on chromosome 9, 4, and 5 in the human, mouse, and rat, respectively (NCBI, 2005). The locus is unique in that alternate splice variants produce two different tumor suppressor proteins (Sherr and Weber, 2000; Sherr and McCormick, 2002; Lowe and Sherr, 2003). The p16Ink4a and p19Arf variants have exons 2 and 3 in common but use different exons 1 (alpha and beta). Expression of these two splice variants is conserved across mammalian species. Mouse p19^Arf and human p14^ARF polypeptides are approximately 50% identical, and mouse p16^Ink4 and human p16^INK4a proteins are approximately 72% identical (Quelle et al., 1995).

The two proteins translated from the mRNA expressed from CDKN2A are a p16-KDa protein and a p19-KDa protein (or p14-KDa protein in humans) (Serrano et al., 1996). The p16 protein (p16^Ink4a, inhibitor of kinase 4a) is a cell cycle regulatory protein that binds to cyclin dependent kinase 4 or 6 (CDK4/6) and inhibits the catalytic activity of the CDK/cyclin D complex and the phosphorylation of retinoblastoma protein. Since loss of the normal function of p16^Ink4a leads to uncontrolled cell growth, p16 is classified as a tumor suppressor gene (Serrano et al., 1993). The second protein coded, p19^Arf (Arf, alternative reading frame), induces G1 arrest and apoptosis. The 19Arf protein binds to MDM2 and neutralizes MDM2 inhibition of p53 (Sherr and Weber, 2000).

The targeted deletion of exons 2 and 3 of the Cdkn2a gene by a homologous recombination resulted in the elimination of both p16^Ink4a and p19^Arf proteins (Serrano et al., 1996). Homozygous null Cdkn2a^−/− (or Cdkn2a^−/−) mice are viable and fertile (Serrano et al., 1996). On inspection, these animals appear normal until about 2 months of age, but histological analysis of the spleen shows a mild proliferative expansion of the white pulp and the presence of numerous megakaryocytes and lymphoblasts in the red pulp. The p16^−/− mice develop tumors at an average age of 29 weeks. Lymphomas and fibrosarcomas are two common types of tumors seen in this Cdkn2a^−/− mouse. In contrast, the Cdkn2a^+/− mouse does not usually develop any obvious tumors or display compromised health until after 36 weeks (Serrano et al., 1996).

Deletions in the Cdkn2a gene predispose both rodents and humans to cancer at multiple organ sites (Sharpless and DePinho, 1999). The complete loss of Cdkn2a gene(s) function is observed in approximately 10% of small cell lung tumors, 30% of esophageal tumors, 55% of gliomas, 100% of pancreatic tumors, and 20% of head and neck tumors.

Transition from G1 to S phase in the mammalian cell cycle is under complex regulatory control, and one G1-S regulatory pathway involves p16^Ink4a protein. P16^Ink4a inhibits the cdk4/cyclin D1 complex, preventing cdk4 from phosphorylating pRb, and thus ensuring that pRb maintains G1 arrest. Disruption of this pathway, by p16^Ink4a gene mutations, perturbs the cell cycle (Serrano et al., 1993) and in the case of these Cdkn2a genetically altered mice (Serrano et al., 1993), results in more cell proliferation (Figure 3).

Figure 3

The INK4a/ARF Locus. The open reading frames p16^Ink4a (in black) and p19^Arf (in crosshatch) are shown. Each has a unique first exon that then splices to a common second exon, but in alternate reading frames. P16^Ink4a inhibits cdk4/6 activity producing (more...)

Serrano et al. (1996) report that treatment with DMBA and UV light causes an earlier onset of fibrosarcoma and lymphoma in the p16^−/− mouse (at 8 to 10 weeks) and the p16^−/+ mouse (7 to 20 weeks).

Study Rationale

The purpose of this benzene study and the glycidol and phenolphthalein studies (NTP, 2008a,b) was to determine if a mouse with a deletion at the p16 gene locus (CDKN2), a locus that codes for two tumor suppressor genes, would enable the identification of carcinogenic chemicals in a shorter time frame and with fewer animals than the traditional 2-year NTP cancer study. These three chemicals were all multisite carcinogens in the NTP 2-year bioassays (NTP 1986, 1990, 1996). This Report presents the findings from the benzene study.