Abstract
Hexenal is a genotoxic compound to which humans are exposed daily through the consumption of foods and beverages. The present studies were conducted to examine the relationships between the dose-responses of trans-2-hexenal-induced toxicity, DNA adduct formation, and cell proliferation. Male F344 rats were exposed by gavage to single doses of up to 500 mg/kg and killed 1, 2, or 4 days after dosing or were exposed to repeat doses of up to 100 mg/kg once daily for 5 days or 5 days per week for 4 weeks and killed 1 day after the end of the dosing period. Histologically, the primary observations were necroulcerative lesions, inflammation, and hyperplasia in the forestomach and inflammation in the glandular stomach. Hexenal-derived DNA adduct formation and cell proliferation were induced in the forestomach at doses of hexenal that also induced gastric toxicity; DNA adducts were not observed in the glandular stomach. These findings suggest that the toxicity of hexenal was limited to the site of contact (stomach) and that the observed DNA adduct formation and cell proliferation occurred in the setting of severe tissue damage.
Introduction
Humans are exposed daily to the 6 carbon α, β-unsaturated aldehyde, trans-2-hexenal (hexenal), through consumption of food and beverages. Human exposures to hexenal are ~350 μg/kg/day, with 98% derived from natural sources and 2% from artificial flavorings (Stofberg and Grundschober, 1987; Flavor and Extracts Manufacturer’s Association, personal communication).
Few data exist on the toxicity of hexenal (Gaunt et al., 1971; Ping et al., 2003). Following single gavage doses of hexenal, oral LD50 values were 780 or 1,130 mg/kg for male or female Carworth Farm Strain E (CFE) rats and 1,750 or 1,550 mg/kg for male or female Carworth Farm Swiss Webster (CFW) mice (Gaunt et al., 1971). Organ toxicity was not observed during the LD50 study. No clear evidence of toxicity was observed in male and female rats exposed to hexenal through the diet for 13 weeks at doses of up to 257 mg/kg/day in male rats or 304 mg/kg/day in female rats (concentrations of up to 4,000 ppm in feed; Gaunt et al., 1971). However, daily gavage administration of 200 mg/kg hexenal to rabbits for 13 weeks resulted in gastric toxicity, including hemorrhage and ulcers, and a decreased hemoglobin concentration (Gaunt et al., 1971). Cardiotoxicity has been reported in male Institute of Cancer Research (ICR) mice following hexenal exposure (Ping et al., 2003). Decreased left ventricular function, increased cardiomyocyte apoptosis, and increased hexenal-protein adduct formation were observed. Protein adducts were increased both in the heart and in isolated cardiac mitochondria incubated with hexenal. Hexenal has not been thoroughly evaluated for carcinogenic potential, but there is evidence of tumor formation at multiple sites in rodents following intraperitoneal (i.p.) injection of 50 mg/kg, although the study was limited by a small sample size (Nádasi et al., 2005).
Organ toxicity and carcinogenicity following administration of other α, β-unsaturated aldehydes has been demonstrated. Acrolein induced gastric and respiratory toxicity and promoted carcinogenesis of the bladder (Cohen et al., 1992; Costa et al., 1986; Feron et al., 1978; Lijinsky et al., 1987; Lyon et al., 1970; Parent, Caravello, and Hoberman, 1992; Roemer et al., 1993; Sakata et al., 1989). Crotonaldehyde induced gastric, respiratory, and hepatic toxicity and hepatic carcinogenesis (Chung et al., 1986; Wolfe et al., 1987). 2,4-Hexadienal induced gastric toxicity and carcinogenesis (Chan et al., 2003; National Toxicology Program, 2003). Many of the toxic effects of hexenal and other α, β-unsaturated aldehydes were observed at the site of contact; these effects are suggestive of local tissue irritation as an important mode of action. In other studies, toxic or carcinogenic effects have not been observed following exposure to acrolein or croton-aldehyde (Parent et al., 1991; Parent, Caravello, and Long, 1992; Von Tungeln et al., 2002).
There is extensive evidence of hexenal-induced genotoxicity and mutagenicity. Hexenal induces a variety of genotoxic lesions following in vitro exposures, including oxidative DNA damage, exocyclic DNA adducts, single-strand breaks, micronuclei, sister chromatid exchanges, aneuploidy, and unscheduled DNA synthesis (Dittberner et al., 1995; Eder et al., 1993; Eisenbrand et al., 1995; Glaab et al., 2001; Gölzer et al., 1996; Griffin and Segall, 1986; Janzowski et al., 2003). Hexenal was mutagenic in the Ames assay (Eder et al., 1992; Marnett et al., 1985) and in Chinese hamster ovary cells at cytotoxic concentrations (Canonero et al., 1990).
DNA adduct formation is thought to be an important step in the initiation and progression of carcinogenesis, and DNA adducts are biomarkers of both exposure and genotoxicity. Hexenal forms a pair of diastereomeric exocyclic 1,N 2-propan-odeoxyguanosine adducts (Hex-PdG) on reaction with deoxyguanosine (Figure 1; Eder et al., 1993; Eder and Hoffman, 1993) and DNA (Douki and Ames, 1994; Gölzer et al., 1996; Stout et al., 2006). Hex-PdG has been detected, using 32P-postlabeling, in the DNA of treated cells (Gölzer et al., 1996) and tissues (Schuler and Eder, 1999; Schuler et al., 1999) following hexenal exposure. High concentrations of Hex-PdG were reported in F344 rats 2 days after single doses of 200 or 500 mg/kg, especially in the forestomach, liver, esophagus, and kidney, and to a lesser extent, in the duodenum, colon, glandular stomach, lung, and urinary bladder (Schuler and Eder, 1999). Hex-PdG was detectable only in the forestomach and liver and quantifiable only in the esophagus in animals receiving 50 mg/kg.
Hexenal appears to be metabolized primarily by aldehyde dehydrogenase (ALDH), which oxidizes hexenal, and glutathione-S-transferase (GST), which conjugates hexenal to glutathione (GSH; Lame and Segall, 1986; Mitchell and Petersen, 1987, 1989; Boyland and Chasseaud, 1968; Eisenbrand et al., 1995); these reactions are thought to detoxify hexenal (Townsend et al., 2001; Marnett et al., 1985). Reduction by aldose reductase (AR) may also be involved in the detoxification of hexenal. Reduction of hexenal by this enzyme has been demonstrated (Burczynski et al., 2001; Ramana et al., 2001) and may be more efficient following GSH conjugation (Dixit et al., 2000; Ramana, et al., 2001). Since hexenal does not require metabolic activation, the dose at which the detoxification becomes saturated should represent an inflection point for a sublinear dose response. At doses above this point, the response would be much greater per unit dose (La and Swenberg, 1996). Below this dose, metabolic defenses should be sufficient for detoxification. A sublinear dose response was previously reported for Hex-PdG induction in esophagus, forestomach, and liver DNA in F344 rats exposed to single doses of hexenal (Schuler and Eder, 1999).
Currently, hexenal is generally regarded as safe as a flavoring agent, despite its ubiquitous human exposure and genotoxic potential. The present studies were conducted to define the dose-response relationships for hexenal-induced toxicity, DNA binding, and cell proliferation following short-term exposures through conventional and mechanistic toxicology studies to extend molecular dosimetry data to lower, repeat-dose exposures to examine the possible accumulation of Hex-PdG and to place the mechanistic events in the context of organ toxicity.
Materials And Methods
Test Article and Dosing Solutions
Hexenal (CAS 6728-26-3) was obtained from Sigma (St. Louis, Missouri); the reported purity was 98%. The identity of hexenal was confirmed using nuclear magnetic resonance (NMR) spectroscopy. Dosing solutions (2 ml/kg) were prepared in corn oil (Sigma, St. Louis, Missouri). For single-dose and 1-week dosing, dosing solutions were prepared just before the first dose administration. The stability of dosing solutions for 1 week was confirmed by NMR spectroscopy before the initiation of the 4-week study; dosing solutions were prepared once weekly for the 4-week study.
Study Animals
Male F344 rats were obtained from Charles River Laboratories (Raleigh, North Carolina) at 6 weeks of age and allowed to acclimate for 1.5 to 2 weeks before the initiation of dosing. There were no significant differences in body weights between dose groups before the initiation of dosing for either the single-dose or repeat-dose studies. Rats were housed 5 to a cage with a 12-hour day/night cycle and given standard laboratory rat chow (Purina) and tap water ad libitum. Tail markings with a permanent marker were used for individual rat identification. Rats were dosed in the morning and were not fasted before dosing. The studies described below were conducted at the University of North Carolina, Chapel Hill, North Carolina, under the auspices and criteria of the Institutional Animal Care and Use Committee; the Department of Laboratory Animal Medicine is accredited by the Association for the Assessment and Accreditation of Laboratory Animal Care.
Study Design
Single-dose Exposures
The dosing regimen for the single-dose study was designed according to the protocol of Schuler and Eder (1999), with the exception that the present study did not include an 8-hour post-dose killing. Groups of 15 rats were dosed with 0, 50, 200, or 500 mg/kg hexenal by oral gavage. Five animals from each dose group were killed 1, 2, or 4 days after dosing. Body weights were recorded at dosing and just before necropsy. Rats were anesthetized using Nembutal (100 mg/kg i.p.) and exsanguinated via the aorta.
Repeat-dose Exposures
Groups of 5 rats were dosed with 0, 10, 30, or 100 mg/kg hexenal once daily by oral gavage for 5 days or 5 days per week for 4 weeks and killed 1 day after the final dose. Rats were anesthetized using Nembutal (100 mg/kg i.p.) and exsanguinated via the aorta. Body weights were recorded on the first day of dosing, every other day during the dosing period, and just before necropsy.
Pathology
At necropsy, tissues were examined for gross lesions. Sections of stomach, liver, and kidney were fixed in 10% neutral buffered formalin, processed, embedded in paraffin, sectioned, and stained with hematoxylin and eosin. Tissues were examined and images captured microscopically. The remaining forestomach, glandular stomach, and liver were then snap frozen on dry ice or in liquid nitrogen and stored at –80°C until DNA isolation.
Clinical Chemistry and Hematology
Blood was collected from anesthetized animals in microcollection tubes containing potassium ethylenediaminetetracetic acid (EDTA) for hematology. Hematology analysis included measurement of the following parameters: erythrocyte, platelet and leukocyte counts, hematocrit, hemoglobin, mean cell volume, mean cell hemoglobin, and mean cell hemoglobin concentration using a Technicon H-1 hematology analyzer (Bayer, Medfield, Massachusetts). Differential leukocyte counts and cellular morphology were determined microscopically from blood smears stained with a modified Wright-Giemsa stain. Using a Miller disc, new methylene-blue–stained whole blood was used to determine reticulocyte counts. For clinical chemistry analyses, blood samples were collected in microcollection serum separator tubes, allowed to clot, and centrifuged. Serum was analyzed using a Cobas Mira chemistry analyzer (Roche Diagnostic Systems, Inc., Montclair, New Jersey). Serum chemistry analysis included measurement of the following parameters: albumin, total protein, urea nitrogen, creatinine, glucose, cholesterol, triglyceride, total bile acids, alkaline phosphatase, alanine aminotransferase, aspartate aminotransferase, sorbitol dehydrogenase, and 5’-nucleotidase.
Quantitation of Hex-PdG
DNA was isolated essentially as described previously (Morinello et al., 2002) with exceptions. In some cases, a cell lysis solution (Gentra, Minneapolis, Minnesota) was used instead of lysis buffer, and on occasion, the volume of celllysis solution was reduced from 5 ml to 3 ml when <500 mg of forestomach was used. DNA was quantitated by ultraviolet spectrophotometry, using either water or 10 mM tris-1 mM EDTA buffer solution (Sigma, St. Louis, Missouri) as the diluent.
DNA from forestomach (up to 200 μg; single doses of 0, 50, or 200 mg/kg, killed 1, 2, or 4 days after dosing or 0, 10, 30, or 100 mg/kg for 1 or 4 weeks), glandular stomach (500 μg; single doses of 0 or 200 mg/kg, sacrificed 1, 2, or 4 days after dosing or 0 or 100 mg/kg for 1 or 4 weeks), or liver (500 μg; single doses of 0, 200, or 500 mg/kg, sacrificed 1, 2, or 4 days after dosing or 0 or 100 mg/kg for 1 or 4 weeks) of hexenal exposed rats was processed and analyzed by liquid chromatography tandem mass spectrometry (LC/MS/MS) according to a published method (Stout et al., 2006), with some exceptions. Briefly, DNA was enzymatically hydrolyzed in the presence of 61 or 305 femtomol (fmol) [13C4 15N2] Hex-PdG as an internal standard and centrifugally filtered. Next, samples were purified by solid phase extraction, evaporated in a centrifugal lyophilizer, and resus-pended in 20 to 40 μl 10% acetonitrile. For each set of samples processed, a standard curve of 0, 5, 15, and 25 fmol Hex-PdG analyte standard (AS) spiked into 200 μg calf thymus DNA (ctDNA), a method blank, an internal standard blank, and a sample of hexenal-treated ctDNA was included to assess method performance. A complete description of these controls can be found elsewhere (Stout et al., 2006). LC/MS/MS data were acquired using a Finnigan Surveyor autosampler and pump coupled to a Finnigan TSQQuantum triple-quadrupole mass spectrometer. An Aquasil 150 x 2.1 mm (5μ) C18 reversed-phase high performance liquid chromatography (HPLC) column (Fisher Scientific, Raleigh, North Carolina) was used for separation of Hex-PdG from other sample components. The mass spectrometer was equipped with an electrospray source that was operated in the positive ionization mode. For mass analysis, selected reaction monitoring was used to monitor the loss of deoxyribose from Hex-PdG and [13C4 15N2]Hex-PdG. The method limit of quantitation was 0.015 fmol Hex-PdG/μg DNA (200 μg DNA) or 0.006 fmol Hex-PdG/μg DNA (500 μg DNA).
Quantitation of Cell Proliferation
Forestomach and liver-cell proliferation were assessed by proliferating cell nuclear antigen (PCNA) immunohistochemistry, as described previously, with exceptions (La et al., 1996). Unstained 5-μm tissue sections on slides were incubated sequentially with peroxidase blocking agent, anti-PCNA primary antibody (PC 10), polymer labeled secondary antibody (Dako Envision HPR), and 3,3’-diaminobenzidine (DAB) and DAB enhancer solution, all of which were obtained from Dako (Carpinteria, California). Slides were then counterstained with Mayer’s Hematoxylin (Sigma, St. Louis, Missouri). For forestomach, both PCNA-positive cells and the length of the mucosa evaluated were determined manually. Data for the forestomach are expressed as PCNA-positive cells (cells staining brown) per mm mucosa (5 to 10 mm counted). Data for liver are expressed as the percentage of S phase cells (cells with an intense brown color) per total number of nuclei using automated counting. BioQuant Nova Prime Image Analysis software was used to obtain all measurements and for automated counting.
Statistical Analysis
Two approaches were used to assess the significance of pairwise comparisons between dosed and control groups in the analysis of continuous variables. Body-weight data, which have approximately normal distributions, were analyzed using the parametric multiple comparison procedures of Williams (1971, 1972) and Dunnett (1955). Hematology, clinical chemistry, cell proliferation, and cumulative percentage body-weight–change data, which typically have skewed distributions, were analyzed using the nonparametric multiple comparison methods of Shirley (1977) and Dunn (1964). Jonckheere’s test (Jonckheere, 1954) was used to assess the significance of dose-response trends and to determine whether a trend-sensitive test (Williams’or Shirley’s test) was more appropriate for pairwise comparisons than a test that does not assume a monotonic dose-response (Dunnett’s or Dunn’s test). Trend-sensitive tests were used when Jonckheere’s test was significant at p < .01.
Results
Single-dose Exposures
Doses (0, 50, 200, or 500 mg/kg) were the same as those used by Schuler and Eder (1999). One rat given a single dose of 500 mg/kg was removed from the study 2 days after dosing because of excessive weight loss. In some cases, certain tissues were not available from all animals in a group.
Body Weight
Body-weight data are summarized in Table 1. Reductions in final body weight were observed in rats treated with a single dose of 200 or 500 mg/kg and killed 4 days after dosing. Reductions in body-weight change were observed in rats treated with a single dose of 200 mg/kg and killed 4 days after dosing and in rats treated with a single dose of 500 mg/kg and killed 1, 2, or 4 days after dosing. Negative trends in body-weight change were observed 1 or 4 days after dosing and in cumulative percentage body-weight change 4 days after dosing.
Hematology and Clinical Chemistry
No hematological changes attributed to a single administration of up to 500 mg/kg of 2-hexenal were observed at 1, 2, or 4 days postdosing (data not shown). Selected clinical chemistry data demonstrating significant biochemical changes in rats receiving single doses of hexenal are shown in Table 2. After single doses of hexenal, dose-related decreases in albumin and total protein concentrations occurred, with a 30% reduction in the 200 mg/kg group and a 50% reduction in the 500 mg/kg group on days 1 and 2 postdosing. These serum protein effects appeared to be transient, and by day 4 postdosing, the protein decreases ameliorated. Single administration of up to 500 mg/kg of hexenal was associated with increased serum concentrations of triglycerides. The highest triglyceride increase occurred on day 1 post-dosing with an approximate 4-fold increase in the 500 mg/kg animals, with the increase ameliorating with time. Single administration of up to 500 mg/kg of hexenal resulted in decreased serum activities of several enzymes. However, the toxicologic significance of this was not apparent.
Gross Pathology
At necropsy, lesions were observed only in the stomach and increased in both severity and incidence with increases in dose. No significant lesions were observed in any other tissues. After a single dose of 50 mg/kg, there was slight edema in the stomach of 2 animals. At 200 or 500 mg/kg, there was necrosis of the forestomach mucosa with a prominent pseudomembrane of fibrin, degenerative cells, and debris.
Histopathology
Histopathology findings in the forestomach and glandular stomach are summarized in Table 3; selected lesions are presented in Figures 2, 3, and 4. No significant changes were observed in the vehicle control groups.
50 mg/kg
After 1 day, there was minimal multifocal epithelial hyperplasia with mild mucosal edema in the forestomach and cellular infiltrates in the lamina propria at the gastric limiting ridge in one animal. These changes were also observed in another rat but only at the gastric limiting ridge. Similar forestomach changes were observed in some or most rats at 2 or 4 days, respectively.
200 mg/kg
After 1 day, there was severe, acute, locally extensive coagulative necrosis with multifocal erosions and severe sub-mucosal edema in the forestomach of all rats. In general, injury to the gastric limiting ridge was less severe than injury to the forestomach. In 3 rats, the lesions at the gastric limiting ridge were characterized by multifocal erosions, mucosal edema, and hemorrhage. In the other 2 rats, there was necrosis and diffuse ulceration of the gastric limiting ridge. In the glandular stomach of all rats in this group, there was severe, diffuse submucosal to transmural edema with cellular infiltrates. In 1 rat, there were also multifocal mucosal erosions with glandular loss and regeneration.
After 2 days, there was severe, subacute, locally extensive coagulative necrosis with erosions or ulcers and severe edema in the forestomach of all rats. There were also interspersed areas of significant mucosal regeneration and epithelial hyperplasia. The lesions at the gastric limiting ridge were somewhat different between the 3 available sections evaluated. The lesions were characterized by multifocal erosions with or without hemorrhage, mucosal regeneration, and epithelial hyperplasia. In 1 rat, there were multifocal ulcers, erosions, and areas of necrosis with large intraepithelial bullae. A pseudomembrane of fibrin, bacterial colonies, and degenerated neutrophils covered the areas of necrosis. In the lamina propria of the glandular stomach, there were mild cellular infiltrates of mainly neutrophils with or without edema in all rats.
After 4 days, severe mucosal regeneration and epithelial hyperplasia with dysplasia were present in all rats, with submucosal edema, neovascularization, and fibroblast proliferation in most rats. In 2 rats, a large ulcer in the forestomach was covered by a coagulum of fibrin, neutrophils, red blood cells, bacteria, and cellular debris. The lesions at the gastric limiting ridge in 3 rats were similar to that of the forestomach, with the exception of less prominent hyperplasia. The primary response observed in the glandular stomach was mild to severe edema with cellular infiltrates of neutrophils and macrophages, frequently perivascular in location.
500 mg/kg
After 1 day, there was erosion with or without ulceration and necrosis of the forestomach with significant edema, congestion, and infiltrates of neutrophils and macrophages extending in 1 rat to the subserosa. In 1 rat, a section of intact epithelium adjacent to the gastric limiting ridge was hyperplastic with some dysplasia. In the other rats, there was coagulative necrosis of the gastric limiting ridge with edema, hemorrhage, and cellular infiltrates of mainly neutrophils. In these rats, the glandular mucosa adjacent to the gastric limiting ridge was necrotic with or without erosions, edema, and cellular infiltrates. In the remaining 4 rats, there was either diffuse coagulative necrosis or ulceration and erosion of the forestomach mucosa. The ulcers were covered by a coagulum of necrotic debris, fibrin, degenerating neutrophils, bacteria, and red blood cells. The underlying tissue was expanded by mild to severe edema with large numbers of predominantly neutrophils forming prominent cuffs around blood vessels. The epithelium at the gastric limiting ridge was intact with mild epithelial hyperplasia in 2 rats accompanied by edema and cellular infiltrates. In the other 2 rats, the epithelium was necrotic with ulcers and erosions extending into the adjacent glandular stomach. In 1 rat, there were also scattered bacterial colonies and vessels with fibrinoid necrosis in the submucosa. In the glandular stomach of most rats, there was hemorrhage at the tips of the glands with edema and infiltrates of predominantly neutrophils, extending in some transmurally.
After 2 days, there was coagulative necrosis or diffuse loss of the mucosa with ulcers and erosions in the 4 rats available for analysis. The lesions were typically covered by a coagulum of fibrin, hemorrhage, necrotic debris, and bacteria. There was also significant edema of the submucosa with large numbers of infiltrates, that is, primarily neutrophils that were also frequently arranged as perivascular cuffs. In 1 rat, these inflammatory changes extended to the subserosal area. The lesions at the gastric limiting ridge were similar to those of the forestomach in the rats examined. In the glandular stomach of some rats, there was mucosal edema with large numbers of cellular infiltrates and acute hemorrhage. In 1 rat, there was also epithelial necrosis with mineralization and dilation of the lumens. In the submucosa, there was severe diffuse edema, significant infiltrates of predominantly neutrophils, and fibrinoid vascular necrosis. In 1 rat, the inflammatory infiltrate extended into the subserosa.
After 4 days, there was diffuse necrosis of the forestomach mucosa, which extended in some cases into the muscularis mucosa. Associated with the necrosis were severe diffuse edema, significant cellular infiltrates, fibrinoid vascular necrosis, and hemorrhage as well as fibroblast proliferation and mucosal regeneration. In most rats, the lesions at the gastric limiting ridge mirrored those of the forestomach. Variable lesions were observed in the glandular stomach of most rats. The lesions were characterized by multifocal ectatic glands filled with necrotic debris that were lined by regenerating epithelium with or without hemorrhage, submucosal edema, and cellular infiltrates that sometimes formed perivascular cuffs.
Hex-PdG Formation
DNA from the forestomach (200 μg), glandular stomach (500 μg), and liver (500 μg) was examined for Hex-PdG formation. In some cases, less than 200 μg forestomach DNA was available for analysis. Quantitation of Hex-PdG in forestomach DNA of rats exposed to 500 mg/kg was not attempted because of the lack of viable tissue. Hex-PdG was not quantifiable in control forestomach DNA or in forestomach DNA after exposure to 50 mg/kg in most rats (n = 5). Hex-PdG was not quantifiable in control glandular stomach DNA or in glandular stomach DNA after exposure to 200 mg/kg (n = 3). Hex-PdG was not quantifiable in control liver DNA or in liver DNA after exposure to 200 or 500 mg/kg in most rats (n = 3 or 5). Hex-PdG was quantifiable in forestomach DNA of 1 rat exposed to a single dose of 50 mg/kg and sacrificed 2 days after dosing (0.016 fmol/μg DNA; n = 5) and in all rats exposed to a single dose of 200 mg/kg and sacrificed 1 day after dosing (n = 5). Concentrations of Hex-PdG in forestomach DNA are shown in Table 4. Hex-PdG was occasionally observed in liver DNA of rats exposed to a single dose of 200 or 500 mg/kg hexenal and killed 1 day after dosing. The highest concentrations of Hex-PdG in liver DNA were found in the DNA of 1 rat exposed to 200 mg/kg and killed 1 day after dosing (0.006 fmol/μg DNA) and in 2 rats exposed to 500 mg/kg and killed 1 day after dosing (0.002 or 0.004 fmol/μg DNA). Hex-PdG concentrations in these samples were at or slightly below the limit of quantitation but were not duplicated when analyzed more than once.
Repeat-dose Exposures
Doses for the 1-week study were selected based on results of the single-dose study, while doses for the 4-week study were selected based on the results of the 1-week study. Because of a dosing error, 1 rat was killed on the first day of the 4-week study. As a result, there were only 4 control animals at this exposure duration. In some cases, certain tissues were not available from all animals in a group.
Body Weight
Body weight data are summarized in Table 5. Reductions in body-weight change were observed in rats treated with 100 mg/kg for 4 weeks. Negative trends in body-weight change and cumulative percentage body-weight change were observed after treatment for 4 weeks.
Hematology and Clinical Chemistry
No hematological or biochemical changes attributed to administration of up to 100 mg/kg of 2-hexenal were observed following 1 or 4 weeks of daily dosing (data not shown).
Gross Pathology
At necropsy, lesions were observed only in the stomach and increased in both severity and incidence with increases in dose. No significant lesions were observed in any other tissues. Edema in the stomach was observed in 2 rats following exposure to 30 mg/kg for 4 weeks and in most rats following treatment to 100 mg/kg for 5 days or 4 weeks. Gross lesions in animals exposed to repeat doses of 100 mg/kg were less severe than lesions resulting from a single dose of 200 mg/kg and were characterized primarily by edema but with ulcerations in some cases.
Histopathology
Histopathology findings in the forestomach and glandular stomach are summarized in Table 6; selected lesions are presented in Figure 5. No significant changes were observed in the vehicle control groups. Histopathologic lesions of the stomach following treatment with daily oral doses for 5 days or 5 days/ week for 4 weeks were confined to the forestomach. After 5 days, there were no significant histopathologic changes at 10 mg/kg. Exposure to 30 mg/kg induced minimal, focal, or multifocal epithelial hyperplasia in 3 of the 5 rats. Following exposure to 100 mg/kg, there was moderate-to-severe diffuse epithelial hyperplasia (epithelization) with dysplasia in all rats and ortho-keratotic hyperkeratosis in 3 rats. There were also chronic-active ulcers with submucosal edema and fibroblast proliferation in 2 rats. Exposure to 10 mg/kg for 4 weeks resulted in minimal, multifocal epithelial hyperplasia. Following exposure to 30 mg/kg, there was mild-to-moderate, multifocal-to-diffuse epithelial hyperplasia. Exposure to 100 mg/kg resulted in either moderate-to-severe diffuse epithelial hyperplasia in 4 rats or severe diffuse hyperplasia (epithelization) with dysplasia in 1 rat. Damage to the gastric limiting ridge in exposed rats was not different from damage to the forestomach in these animals.
Hex-PdG Formation
DNA from the forestomach (200 μg), glandular stomach (500 μg), and liver (500 μg) was examined for Hex-PdG formation. In some cases, less than 200 μg of forestomach DNA was available for analysis. Hex-PdG was not quantifiable in control forestomach DNA (1 week: n = 4; 4 weeks: n = 5) or forestomach DNA of rats after exposure to 10 or 30 mg/kg for 1 or 4 weeks (n = 5). Hex-PdG was not quantifiable in control glandular stomach DNA or glandular stomach DNA after exposure to 100 mg/kg for 1 or 4 weeks (n = 3). Hex-PdG was not quantifiable in control liver DNA or in liver DNA after exposure to 100 mg/kg for 1 (n = 5) or 4 (n = 3) weeks. Hex-PdG was quantifiable in forestomach DNA of most rats exposed to 100 mg/kg of hexenal for 1 or 4 weeks (n = 5). Concentrations of Hex-PdG in forestomach DNA are shown in Table 5, and representative LC/MS/MS chromatograms from rats treated with 0 or 100 mg/kg for 4 weeks are shown in Figure 6.
Cell Proliferation
Cell-proliferation data of the liver and forestomach of rats treated with hexenal for 1 or 4 weeks are summarized in Tables 7 and 8. Positive trends in forestomach cell proliferation were observed following treatment for 1 or 4 weeks. Increases were significant after exposure to 30 mg/kg for 4 weeks (p < .05) or 100 mg/kg hexenal for 1 (p < .05) or 4 (p < .01) weeks. Cell proliferation was not significantly increased over controls in the liver after exposure to 100 mg/kg hexenal for 1 or 4 weeks.
Discussion
Hexenal is a genotoxic compound to which humans are exposed daily through the consumption of foods and beverages, in which it is present as both a natural and an artificial ingredient. The purpose of these studies was to examine the relationships between the dose-responses of toxicity, DNA binding, and cell proliferation following in vivo exposure to hexenal, thus anchoring cellular and molecular mechanistic endpoints to organ toxicity. In addition, these studies provide data on the hazards that occur following exposure to high-bolus doses of hexenal.
Decreased body weights or body-weight gains were observed following single doses of 200 or 500 mg/kg, or repeat doses of 100 mg/kg. On gross or microscopic evaluation of tissues, there were no significant lesions observed in the livers of rats at any dose or exposure duration. Damage to the glandular stomach was minor compared to the forestomach but clearly present and primarily characterized by inflammation at single doses of 200 or 500 mg/kg. In the forestomach, damage was minimal at 50 mg/kg, while necroulcerative lesions accompanied by inflammation were predominant at single doses of 200 or 500 mg/kg. This would suggest that metabolic defenses were saturated between 50 and 200 mg/kg and that an inflection point for a sublinear dose response would occur between these doses. Forestomach hyperplasia was the predominant lesion observed following repeat dosing. Hyperplasia was more extensive across dose groups at 4 weeks compared to 1 week, but was most apparent at 100 mg/kg for both exposure durations, indicating that damage occurred at lower doses with a longer period of exposure. The progression of lesions of the forestomach from those observed following single-dose exposures to those observed following repeat-dose exposures indicates that an adaptive proliferative response was activated to repair damage. The induction of gastric toxicity by hexenal was most likely the cause of the observed body-weight effects. The finding of gastric toxicity was consistent with that observed in rabbits (Gaunt et al., 1971).
Clinical chemistry analysis provided biochemical evidence in support of the histopathologic findings. The decreases in total protein concentration were consistent with the decrease in serum albumin. A decrease in serum albumin is a common form of dysproteinemia (Kaneko, 1989) and has been related to protein loss, nutritional deficiencies, or altered albumin metabolism. Because the change in serum proteins occurred acutely postdosing and since there was gross or histopathological evidence of gastric injury but no biochemical or histopathological evidence of liver injury, the transient alteration in albumin concentration was most likely caused by the acute toxic insult to the gastrointestinal tract. The mechanism for the acute but transient increase in serum triglyceride concentration was unknown. Since the triglyceride effect was also acute and apparently transient, similar to the serum protein effects, it could be that the change in the triglyceride concentration was also related to an acute toxic insult to the gastrointestinal tract. The toxicological implication of the decreased serum enzyme activities in this study is unknown.
Analysis of Hex-PdG formation was undertaken to compare Hex-PdG measurements by LC/MS/MS to those that were measured by 32P-postlabeling after single-dose exposures to place Hex-PdG formation in the context of organ toxicity, which would aid in determining if genotoxicity would potentially occur at doses relevant to human exposures or at high doses secondary to organ toxicity and to extend molecular dosimetry data to lower, repeat-dose exposures to examine the possible accumulation of Hex-PdG. Despite the replication of the single-dose exposure parameters, the results reported herein were strikingly different from those obtained by 32P-postlabeling, in which adduct concentrations were relatively high, particularly in the esophagus, forestomach, and liver (Schuler and Eder, 1999); surprisingly, the extensive toxicity to the forestomach observed during the present study was not reported (Schuler and Eder, 1999).
The purity of hexenal used in the Schuler and Eder (1999) study was 99%, while the purity of hexenal used in the current study was 98%. Corn oil was used as a vehicle in both studies. Even in the unlikely event that such a small difference in purity could account for the difference in toxicity, Hex-PdG concentrations should still be similar. In rats killed 2 days after dosing with 200 mg/kg, 1.3 and 0.45 fmol Hex-PdG/μg DNA were reported in the forestomach and liver (using the approximate conversion factor of 70 nmol dGuo/100 μg DNA). Two days after treatment with 500 mg/kg, these concentrations were 9.0 and 5.0 fmol/μg DNA. In the present study, Hex-PdG was quantifiable only in the forestomach of rats killed 1 day after a single dose of 200 mg/kg or 1 day after exposure to 100 mg/kg for 1 or 4 weeks but at concentrations 25-fold to 50-fold lower than previously reported (0.02 to 0.04 fmol/μg DNA). These concentrations of Hex-PdG were approaching the method limit of quantitation; Hex-PdG was quantifiable in most but not all of these samples. Hex-PdG was not detected 2 days or 4 days after single doses or at lower concentrations of single or repeat doses. Hex-PdG measurements in forestomach DNA were not examined in rats exposed to 500 mg/kg hexenal, because of complete or near-complete loss of the forestomach mucosa. Doses that did not induce damage to the mucosa or induce cell proliferation did not result in Hex-PdG formation in forestomach DNA.
Detection of Hex-PdG in these tissues may have been obscured by tissue necrosis or cell proliferation. It is also possible that cells containing Hex-PdG undergo apoptosis, preventing the measurement of Hex-PdG in these cells. Hex-PdG was not quantifiable in glandular stomach DNA under any exposure scenario. Hex-PdG was occasionally quantifiable in liver DNA at concentrations (≤0.006 fmol/μg DNA) 2 to 3 orders of magnitude lower than those previously reported on treatment with single doses that induced severe gastric toxicity (1 day after 200 or 500 mg/kg); these signals may have been spurious. In any case, Hex-PdG formation in liver DNA was not definitively shown. Hex-PdG was not quantifiable in the liver following repeat-dose exposures.
The time course of Hex-PdG formation was also different between the 2 studies. In the previous study, the highest concentrations of Hex-PdG were found 2 days after dosing, while in the present study, formation of Hex-PdG was only found 1 day after dosing. The discrepancy in the Hex-PdG concentrations between 32P-postlabeling and LC/MS/MS is possibly caused by inherent differences in the methods used. 32P-postlabeling, while sensitive, is not specific, does not use an internal standard, and can be prone to false positive results because of poor resolution and lack of structural characterization during analysis. It is possible that an unresolved mixture of adducts was detected. The advantages of LC/MS/MS over 32P-postlabeling have been described elsewhere (Koc and Swenberg, 2002; Singh and Farmer, 2006).
In primary rat hepatocytes, [14C2]hexenal bound with decreasing relative affinity to DNA, RNA, and protein (Eisenbrand et al., 1995). However, cell-protein content is much higher than DNA content, so the majority of the hexenal was bound to protein. As a result, the extent of DNA binding at doses that do not extensively bind protein would likely be low because of the excess of protein available for binding, while doses that induced extensive protein binding would likely also overwhelm metabolism and result in toxicity. In the present study, animals were not fasted before dosing, which occurred in the morning following nocturnal feeding. Thus, glutathione levels would not be artificially reduced; since α, β-unsaturated aldehydes are direct-acting and have been shown to interact with glutathione, depletion might result in increased toxicity and DNA binding. The previous study does not indicate when dosing occurred or if animals were fasted.
The extent of cell proliferation was quantitated using PCNA immunohistochemistry, which has been demonstrated to be an effective means of examining cell proliferation retrospectively (Dietrich et al., 1994). An increase in cell proliferation of the forestomach was anticipated because of the initial necroulcerative damage and subsequent hyperplasia observed microscopically, and it likely occurred secondarily to tissue damage. The trends for forestomach cell proliferation were positive at both 1 and 4 weeks. Increases were significant after dosing with 30 mg/kg for 4 weeks and with 100 mg/kg for 1 week or 4 weeks. Since the liver is also able to proliferate following cytotoxic damage, the lack of significant proliferation supports the absence of toxicity in the liver. Rapid cell turnover reduces the time available for DNA repair and is required to convert DNA adducts to mutations. Cell proliferation is also believed to promote tumor formation because of selective expansion of initiated cells. Furthermore, cell proliferation is a frequent sequela to injury and provides a relatively objective and quantitative measure of toxic response.
While hexenal is toxic to the glandular stomach, the extent is low compared to the forestomach and not accompanied by Hex-PdG formation. The glandular epithelium in rats is similar in structure to the stomach of humans, where the production of mucous provides protection against damage. The function of the forestomach in rats is primarily for storage of food, which increases the duration of exposure. Thus, it appears that the phenomenon of toxicity and DNA binding in the rat forestomach following hexenal treatment is caused by direct contact for relatively long periods of time, making its relevance to human exposure less clear. In contrast to the gastric toxicity observed, no histopathological, biochemical, or molecular evidence of liver injury was detected after hexenal exposure. Since hexenal is thought to exert its deleterious effects directly, detoxification by ALDH, GST/GSH, and possibly by AR or other pathways or removal resulting from binding to tissue and dietary constituents such as proteins are possible explanations for the lack of hepatic toxicity.
The National Toxicology Program (NTP) conducted studies of the toxicity and carcinogenicity of 2,4-hexadienal (hexadienal; Chan et al., 2003; National Toxicology Program, 2003). Like hexenal, hexadienal is found in foods as a natural and artificial component and is generally regarded as safe. Subacute (16-day) dosing of up to 240 mg hexadienal/kg by gavage resulted in necroulcerative lesions of the forestomach at high doses and forestomach hyperplasia at lower doses in both F344N rats and B6C3F1 mice. There was no histologic or biochemical evidence of liver injury. There was an increase in absolute liver weight in female rats and mice and liver weight and body weight in female rats at 240 mg/kg. Subchronic (90-day) dosing of up to 120 mg/kg resulted primarily in forestomach hyperplasia. Again, there was no evidence of liver injury except for increased liver weights; relative liver weights were increased at all doses in females, while both absolute and relative liver weights were increased in males and females at 60 mg/kg. Treatment with doses of up to 90 mg/kg in mice and 120 mg/kg in rats for 2 years resulted in increases in forestomach hyperplasia and forestomach papillomas and carcinomas in male and female rats and mice, which the NTP interpreted as clear evidence of carcinogenicity. The relevance of this tumor type to humans is unclear because of the structural differences between the human and rodent stomach. The authors suggest that tumors in this study may have been a result of local irritation, direct or indirect DNA damage, or both. These results are in agreement with the results presented in the present report, both in regard to the spectrum of lesions observed in the forestomach and to the lack of significant liver or glandular stomach toxicity.
The present report details gastric toxicity at high doses following exposure of F344 rats to hexenal by oral gavage. The primary findings of this study were that the deleterious effects of hexenal occurred in the stomach, primarily in the forestomach, but not the liver and that gastric toxicity accompanied Hex-PdG formation, which was found at much lower concentrations than previously reported. The occurrence of both DNA binding and cell proliferation is typically favorable for the initiation and promotion of tumors; however, following gavage exposure to hexenal, these effects on the forestomach occur only at doses that also induce forestomach toxicity. The observation that the deleterious effects observed in this study occurred only at the site of exposure provides support for hexenal’s acting directly, while both the observed histopathology and the relationships among the dose responses provide support for hexenal’s acting primarily through an irritation mode of action, with damage followed by regeneration.
Humans are exposed to a daily dose of ~350 μg hexenal/kg/ day. Human exposure is ~300-fold lower than the 100 mg/kg dose that induced Hex-PdG formation and toxicity to the forestomach at 1 or 4 weeks in rats and ~30-fold lower than the 10 mg/kg dose that induced minimal toxicity to the forestomach but did not induce detectable Hex-PdG in rats. In addition, gavage dosing delivered bolus doses of hexenal directly to the stomach, while human exposures are spread over the day as part of dietary intake. This difference in dosing has been shown to greatly exaggerate toxicity for direct-acting chemicals (La and Swenberg, 1996) because of saturation of metabolic detoxification. Based on the observations that gastric toxicity occurred at relatively high bolus doses, that concentrations of Hex-PdG would be much lower than typical endogenous DNA adduct concentrations (even if formation in the rat forestomach following gavage exposure is linear and similar to the human stomach after dietary exposure concentrations), and that cell proliferation was likely a regenerative response to gastric toxicity, the observed effects on the stomach are unlikely at levels of human exposure. To more thoroughly characterize the dose response of toxicity following hexenal exposure and the potential for hexenal to induce tumors, subchronic and chronic toxicity studies would be needed. Inclusion of the dietary route of administration would elucidate whether the toxicity of gavage doses is caused by the delivery of a bolus dose of a direct-acting toxicant. These studies should follow the dose-selection rationale used by the NTP for study on 2,4-hexadienal, in which doses for 2-year bioassay did not produce overt gastric toxicity in short-term studies (Chan et al., 2003; National Toxicology Program, 2003).
Footnotes
Acknowledgments
We thank Ralph Wilson and Sandy Ward for skilled technical assistance with clinical chemistry and hematology, Norris Flagler for preparation of figures for publication, and Drs. Louise M. Ball and Po Chan for critical review of the manuscript. The mass spectrometry facility at the University of North Carolina at Chapel Hill is partly funded by NIH Grant P30-ES10126. Additional support was provided by NIH Grants T32-CA72319, T32-ES07126, and ES11746, the Flavor and Extract Manufacturers Association, and the National Institute of Environmental Health Sciences, Division of Intramural Research.