FURAN: A FOOD-BORNE FLAVOR CARCINOGEN

Xinchen Zhang, Feng Chen and Mingfu Wang

1 INTRODUCTION

Cancer, the generic name for a group of diseases which can happen at various parts of body, is characterized by uncontrolled rapid reproduction of abnormal cells. As the leading cause of death over the world, WHO estimates continuing growth of deaths from cancer which could be projected to 12 million in 2030.' Although not fully understood, the causes of cancer can be divided into either internal genetic defects or external environmental influences. Studies on the incidence of chronic diseases between the twins and among the migrants suggest that lifestyle and environmental factors account for 90-95% of most chronic illnesses including cancer (Figure I). In this sense, cancer is mostly preventable via avoiding the harmful environmental factors such as tobacco, unhealthy diet, alcohol and infectious organisms. The potential effects of diet on cancer incidence were revealed by observations conducted in countries with different dietary habits and it is noted that carcinogens are either derived from food itself, food additives or produced from processing and cooking. Epidemiological studies and exposure assessment are still needed to help explore these carcinogens' potential on human cancer risk and research on source identification and mitigation approach is of significance.

Furan (C4H4O, MW=68.08) is a kind of heterocyclic organic compound whose single planar aromatic ring is composed of four carbon atoms and one oxygen atom. It exists as colorless liquid but is quite volatile with a boiling point close to room temperature.Commercially, furan can be produced by decarbonylation of furfural under catalyzed steam distillation or oxidation of butadiene and furan produced is in turn used to prepare other organic compounds such as tetrahydrofuran, thiophene and polymers. The presence of furan as a food flavor compound has been well documented, originating from a group of common nutritional components under thermal processing conditions. Given its toxicity and carcinogenicity as revealed by animal studies, growing attention has been drawn to determining furan occurrence in food in order to assess whether human's exposure is within safe dosage. Besides enriching the database of furan occurrence, researchers have been exploring the source and formation of furan in the context of chemical and food models, partially under the expectation of providing instructions on finding realistic strategies to reduce furan content in food products.

2 METABOLISM, TOXICOLOGY AND CARCINOGENESIS

Furan can be quickly absorbed across biological membrane due to its lipophilicity. Based on a research conducted by Burka et al. utilizing [2,5-14C] furan in the male F344 rat, furan was extensively metabolized in the liver and excreted at high efficiency via respiration, feces and urine as carbon dioxide, unchanged furan and multiple metabolites. Cis-2-butene-1,4-dial is an initial oxidation metabolite of furan by the action of cytochrome P450 2E1. Pharmacokinetics studies of furans were performed both in vitro and in vivo. In vitro biotransformation by isolated rat hepatocytes revealed a Km of 0.4µM and a Vmax of 0.018µmol/hr/10 cell while in vivo kinetics determined by a physiologically based pharmacokinetic (PBPK) model were described as a Km of 2.0µM and a Vmax of 27.0µmol/hr/250g F-344 rat. Extrapolation of the Vmax determined in vitro predicted that in vivo very well, showing the validity of modeling the in vivo biotransformation of furan in liver by the in vitro isolated hepatocytes. Further studies aiming to compare the biotransformation of furan in rats, mice and humans reached the conclusion that hepatocytes from all three species metabolized furan rapidly (V of 18-48nmol/hr/10hepatocytes) and with high affinity (Km of 0.4-3.3µM). Interspecies comparison of furan dose indicated that the dose absorbed and cis-2-butene-l,4-dial concentration in human liver were 3- and 10-fold lower than those in rats and mice respectively. Due to blood delivery limitation, stimulation of cytochrome P450 2E1 alone won't result in increase in the quantity of toxic metabolite in the liver, which view is useful when assessing the cancer risk in relation to the population variation of enzyme activity.

The National Toxicology Program published a report on the toxicology and carcinogenesis of furan in 1993. Acute toxicity of furan could lead to death when a level of higher than 40 mg/kg b.w. was administered to F344/N rats orally over 16 days and the toxic effects also involved body weight changes and hepatic historical changes. Based on the 13-week subchronic oral toxicity study, furan at all doses of administration could cause toxic liver lesions to groups of F344/N rats with positive correlation between dosage and severity of lesions. High furan dosage was also potential to result in kidney lesion (30 or 60 mg/kg b.w.) or atrophy of thymus, testis or ovary (60 mg/kg b.w.). Similar lesions to liver were observed for B6C3F1 mice of both sexes. A lower range of dose (0.03-8 mg/kg b.w.) was used in a recent 90-day gavage study on Fischer-344 rats and B6C3F1 mice to further characterize furan's hepatotoxicity. A level of 0.03 mg/kg b.w. was determined as no-observed adverse effect level in rats given the clinical biochemical parameters and hepatic historical changes and the number for mice equaled 0.12mg/kg b.w. Similar low oral dose up to 8 mg/kg b.w. was adopted in another study focusing on furan's subchronic toxicity on male rats' reproductive system and the results indicated a positive correlation between furan ingestion and reproduction impairements. Long term (2-year) evaluation at low furan level treatment (2-15mg/kg b.w.) showed that mice and rats had lower mean body weights and survival rate than controls. Incidence of hepatocellular adenomas or carcinomas, mononuclear cell leukemia, forestomach hyperplasia and other neoplastic or nonneoplastic lesions were found to be significantly increased due to furan administration.

It is clear that furan is capable of targeting organs like lung, liver and kidney and organ necrosis severity is dependent on both the dosage and the expression of cytochrome P450-mixed-function oxidase activity. Research on the mechanism of carcinogenesis is still in progress and the toxicity is believed to be tightly associated with the metabolic fate of furan with cis-2-butene-1,4-dial hypothesized to be the key reactor. Cis-2-butene-1,4-dial could react with glutathione and amino acids, forming metabolites such as its cyclic monoglutathione conjugate, N-acetylcysteine-N-acetyllysine conjugate and cysteinylglycine-glutathione conjugate, which were detected in urine or bile of furan-fed rats. The metabolites associate with liver protein, leading to cell proliferation and chronically cholangiocarcinomas. In contrast, furan does not interact with DNA or form adducts, but cis-2-butene-1,4-dial is reported to form diastereomeric adducts with 2'deoxycytidine, 2'-deoxyguanosine and 2'-deoxyadenosine. The relative reactivity is dependent on pH and the unstable products could be decomposed to secondary products. In consistence, it is cis-2-butene-1,4-dial but not furan that showed concentration-dependent mutagenic activity in Salmonella typhimurium strain TA104 in the Ames assay. Although not able to induce genetic mutations in certain Salmonella typhimurium strains or germ cells of male Drosophila melanogaster, in vitro test in mammalian cells (eg. mouse L5178Y lymphoma cells, Chinese hamster ovary cells) produced some positive result including trifluorothymidine resistance, sister chromatid exchange and chromosomal aberrations. Another aspect of hepatotoxicity proposed by Mugford et al suggested furan-mediated irreversible ATP depletion. Experiments were conducted in freshly isolated F- 344 rat hepatocytes as well as male F-344 rats. State-4 respiration and activity of ATPase were both observed to increase dependent on the concentration of furan and incubation time {in vitro). This uncoupling of mitochondrial oxidative phosphorylation corresponded with cell death and the fact that mitochondrial changes could be prevented by adding cytochrome P450 inhibitor 1-phenylimidazole illustrated the role of cis-2-butene-1,4-dial. Additional consequence of lack of ATP supply was postulated to include activation of endonucleases, leading to breaks of DNA double strands. This process could be partially reversed by adding endonuclease inhibitor or cytochrome P450 inhibitor and DNA-repair happened in living cells within 24hr. Errors during DNA repairing, however, presents as a source of mutation. Lately, experimental work conducted by Chen et al. attempted to test whether furan's carcinogenicity would induce changes in hepatic genes expression associated with cell cycle and apoptosis. As indicated by their results, 4 weeks of treatment with low doses of furan (0.1 and 2 mg/kg) to rats significantly altered mRNA expression profiles related to cell cycle and apoptosis, which was reversible within 2 weeks out of doses administration. Besides, the alteration was shown to be independent of DNA methylation while might be partially accounted by traces of miRNA expression changes. Notably, the conclusion of this research supported the possible role nongenotoxic alteration in gene expression took in furan's carcinogenicity in rodents, which seems to be contradictory to EFSA's viewpoint that "furan-induced carcinogenicity is probably attributable to a genotoxic mechanism", suggesting need of more extensive and comprehensive research in this area.

There is lack of evidence to conclude the health effects of furan on human yet, but studies on rodents did arouse attentions worldwide to potential risk of long-term furan ingestion. The International Agency for Research on Cancer (IARC) classified furan as "possibly carcinogenic to humans" and the European Food Safety Authority (EFSA) also agrees that furan poses risk in food safety. EFSA pointed out that the dosage used in animal experiments to induce tumor might only deviate slightly from possible human exposure to furan. Therefore, precautions are called for till further animal and human studies related to chronic toxicity and exposure analysis provide more reliable risk assessment.

3 OCCURRENCE, ANALYSIS IN FOOD AND EXPOSURE ASSESSMENT

Since the discovery of furan presence in a few thermally-processed food products in the 1960s and 1970s, FDA has been keeping its efforts on monitoring the level of furan in a growing number of foods. It's now believed that furan is much more prevalent in food products than ever thought, with its level ranging from non-detectable to over 120 ppb. The wide presence of furan is especially notable in canned and jarred foods, bread crust, soy sauce, savory snacks, coffee and selected infant formulas and baby foods. Baby foods containing vegetables (especially peas, beans, tomato puree, paprika, and spinach) were detected with much more furan than those containing meat and starch alone or mainly fruit.' Based on the data surveyed on Swiss market, foods that usually contain high level of furan belong to two groups: (1) moist foodstuffs heated in sealed container; (2) baked or roasted food products that are dry and crusty. Accompanying FDA's efforts, EFSA has recommended its member states to submit data on furan concentration analyzed in food samples routinely. The most updated report published in 2011 includes a total of 5050 results belonging to 21 food categories submitted by 20 member states. The samples were analyzed between 2004 and 2010 and the highest furan concentration was observed in coffee, especially "coffee roasted bean" and "coffee roasted ground", whose mean furan contents are 3,660 and 1,936µg/kg respectively. "Baby food" and "soups" are comparatively rich furan-containing non-coffee food products with up to more than 200 ug furan found per kg sample. A summary of furan content in main food categories based on data published by EFSA is present in Table 1.