The Carcinogeiiicity of Essential Oils, Flavors, and Spices: a Review1

[CANCER RESEARCH 28, 2372-2374, November 1968] The Carcinogeiiicity of Essential Oils, Flavors, and Spices: A Review1

F. Homburger and Eliahu Boger Bio-Research Institute, Cambridge, Massachusetts 02141

Summary safrole is toxic, producing weight loss, testicular atrophy, and bone marrow depletion. It also induces hepatomas. Similar Essential oils, flavors, and spices constitute a large and observations were made by Long et al (13) at a 0.5% dietary heterogeneous group of substances to which humans are level of safrole. In our studies on modifying factors of safrole exposed. The essential oils are defined as a group of odorous carcinogenesis (9, 11), similar doses given to male Carworth principles soluble in alcohol and only to a limited extent in Farms CFN rats on riboflavin, tocopherol, and protein- water. They consist of a mixture of esters, aldehydes, ketones, deficient diets resulted in much smaller hepatic adenomas in and terpenes. Relatively few members of this group of sub comparable periods of time. Casein supplements aggravated stances have been shown to cause cancer in animals. This paper the safrole-induced adenomatosis, resulting in tumors larger summarizes knowledge on the carcinogenic activity of safrole, than those observed on protein-deficient diets. Osborne- citrus oils (cMimonene), turpentine oil (/-pinene), eucalyptus Mendel rats appeared to be more sensitive than Carworth oil (phellandrene), bergamot oil, and red pepper. The possible Farms CFN rats (8). Biotin supplements inhibited hepatocar- role of polycyclic hydrocarbons biosynthesized in plants is cinogenic activity of butter yellow as well as safrole, whereas also discussed. pyridoxine deficiency reduced the carcinogenicity of butter yellow but not that of safrole. Carcinogenic Activity of Safrole At a 0.1% level safrole had no toxic effects. Several deriva tives of safrole have been studied by Long and Jenner (12), Our interest in the carcinogenic activity of essential oils was who found that rats fed with dihydro safrole (Compound la) initiated in the early sixties by the observation of B. Zeitlin have developed tumors of the esophagus and none of the liver. (unpublished results), that safrole (4-allyl-l,2-methylenedioxybenzene, Compound I) causes hepatic adenomas in rats Carcinogenicity of Citrus Oils Homburger et al. (9-11). The testing of citrus oils for carcinogenic and cocarcinogenic activity has also been carried out during the last decade. Roe and Pierce (18) reported that repeated applications of various citrus oils on the skin of mice pretreated with an initiating carcinogen benzpyrene (BP), 7,12 dimethylbenz(a)anthracene (DMBA), or urethan resulted in benign and malignant skin tumors. The initiators in acetone solution were applied to the skin in doses which were not sufficient to cause complete CH2-CH=CH2 CHj-CHg-CHj carcinogenesis. Three weeks later the citrus oil was applied once weekly either in undiluted form or diluted with acetone. Whereas in the control experiments (subcarcinogenic doses of I la BP, DMBA, or urethan) only occasional benign tumors ap Safrole is a component of many essential oils, such as star peared, a relatively large number of benign and malignant anise oil, micranthum oil, camphor oil, and, especially, sassa tumors resulted from the promotion effect of orange, lime, fras oil. It also occurs in oil of mace, nutmeg, Japanese wild lemon, and grapefruit oils. Among these, orange oil and its major constituent, d- ginger, California bay laurel, and cinnamon leaf oil. It has been widely used as a flavoring agent in root beer and in certain limonene, have been the main target of investigation. The same pharmaceutical preparations. author (15) reported that in control experiments in which the We have studied the effect of safrole at two levels and the animals were treated with orange oil only, no carcinogenic ef nutritional factors that modify the hepatic adenomatosis fect could be observed. A year later, however, Roe published induced by this substance. In concentrations of 1% of the diet, another article (16) from which it appears that the terpene fraction of the orange oil containing mainly Â¿/-limonene,when applied on the skin of mice, gives rise to epidermic hyperplasia 1 This investigation was supported by USPHS Research Grant No. and subsequently to tumors. CA-05230 from the National Cancer Institute, and by American Cancer These apparently contradictory results can be explained by Society Grant No. E-469. the fact that Â¿/-limonene (Compound II), an unsaturated

2372 CANCER RESEARCH VOL. 28

Downloaded from cancerres.aacrjournals.org on September 24, 2021. © 1968 American Association for Cancer Research. Essential Oils, Flavors, and Spices hydrocarbon with two double bonds, is an unstable com A more recent paper by Field and Roe (3) deals with the pound, and is transformed gradually, mainly upon exposure to effect of oral administration of citrus oils on the tumor air, into its hydroperoxide. By analogy to studies of some of promotion in the forestomach epithelium of mice. The initia the hydrocarbons which are closely related to d-limonene, it tion was carried out by a single treatment with 50 mg of can be deduced that the species responsible for the carcino DMBA or BP in polyethylene glycol given by stomach tube genic or cocarcinogenic effect is actually the autoxidized form after food had been withheld overnight. of cMimonene. The compound, l-vinyl-cyclohex-3-ene (Com The initiation alone (as a control experiment) resulted in a pound HI), which differs from limonane only by the absence very small number of tumors, but when this was followed by of two methyl groups, was reported as a carcinogen (21), but forty once-a-week treatments of 0.05 ml undiluted lime oil, later it was found that the same compound was devoid of any the tumor incidence in the forestomach was markedly in carcinogenic effect (20) when sufficiently purified and pro creased. Orange oil and ci-limonene also induced a few tumors tected from oxygen. The actual carcinogen was found to be when applied in the same way. the hydroperoxide derivative, 1-hydroperoxy-l-vinyl- cyclohex-3-ene (Compound IV). Other Essential Oils In addition to the oils mentioned above, various others have been tested for cocarcinogenic effect but with less significant results. Turpentine oil and /-pinene, its major constituent, were shown to promote skin tumor development in the rabbit (14) but not in the mouse (1, 19). It is possible that the different results obtained were also due to different compositions of the oils used in these experiments. Eucalyptus oil and phellandrene, one of its constituents,

OOH have also been tested as promoting agents on mouse skin (17). About 10% of the mice treated developed tumors. Oils with less terpenes, such as bergamot oil which consists mainly of alcohols and esters, prove less active as tumor- promoting agents. Bergamot oil was completely inactive when m tested on the skin of mice, although linalool, one of its principal alcohols, elicited a weak tumor-producing response Another chemically related hydroperoxide 1-hydroperoxy- (17). cyclohex-3-ene (Compound V), which differs from Compound IV only by the absence of a vinyl group, also exhibited Flavors and Spices carcinogenicity. Although there are many types of different flavoring agents To prove this assumption, we are now testing the carcino and spices, very little is known of their possible tumorigenic genic and cocarcinogenic effects of a highly purified sample of effect. The case of safrole, which served for many years as a cMimonene, and of a sample of partially autoxidized d- flavoring agent of soft drinks, and is a minor constituent of limonene containing 6% of the hydroperoxide. This sample of several spices, was discussed previously. purified limonene, which displays no other peaks except that The use of lemon oil and similar potential cocarcinogens as of uMimonene itself when assayed by gas chromatography, was food additives should also be considered. Special attention supplied to us by the Givaudan Corporation, Clifton, New should be given to the prevention of their autoxidation which Jersey. apparently produces the biologically active species. Even if the Another support for the assumption that the hydroperoxide active species is present in foodstuffs in very small amounts, is the active form of c/-limonene is found in an observation this does not exclude the possibility of cancer hazards, since reported by I. Zukerman (22). This author has studied the prolonged exposures to minute amounts of carcinogens may inhibitive action of both purified cMimonene and its autoxi result after a long latent period in the development of cancers dized form on various microorganisms by adding them to the in the exposed animals. nutrient medium. He has found that the purified (freshly The formation of liver tumors in rats fed red pepper distilled over sodium) d-limonene has no inhibitive action, but (Capsicum) has been reported by Hoch-Ligeti (7). When rats that it reveals the inhibitive properties when exposed to air were fed 10% chili (red pepper) in their diet, a higher and this increases with the duration of exposure. When incidence of neoplastia changes occurred in their livers than in /-ascorbic acid was added to the nutrient medium, a decrease the control experiment. of inhibition was observed. Roe and Pierce (18) have also studied the induction of Natural Occurrence of Polycyclic Aromatic Hydrocarbons tumors at the urethral orifice of female mice during treatment (PAH) in Plants with orange oil, with or without DMBA pretreatment. In this case the number of tumors induced (all of which were benign In the previous chapters, it was shown that some papillomas) was relatively low. components of certain essential oils, flavors, and spices of

NOVEMBER 1968 2373

Downloaded from cancerres.aacrjournals.org on September 24, 2021. © 1968 American Association for Cancer Research. F. Homburger and Eliahu Boger plant origin play a role in the carcinogenic effect of these 7. Hoch-LigetÃ¬,C.Naturally Occurring Dietary Agents and Their Role foodstuffs when tested in animals. In most of the cases, these in Production of Tumors. Texas Rept. BioL Med. 19: 996-1005, components are only promoting agents and have no effect 1952. unless pretreatment with an initiator takes place. It is, there 8. Homburger, F., Bogdonoff, P. D., and Kelley, T. F. Influence of fore, of great importance to determine whether these initiators Diet on Chronic Oral Toxicity of Safrole and Butter Yellow in (such as BP, etc.) also occur in plants. Rats. Proc. Soc. Exptl. Biol. Med., 119: 1106-1110, 1965. As a result of Wieland and Dane's classical work on the 9. Homburger, F., Kelley, T., Jr., Baker, T. R., and Russfield, A. B. Sex Effect on Hepatic Pathology from Deficient Diet and Safrole synthesis of methylcholanthrene from bile acids, many at in Rats. Arch. Pathol., 73: 118-125, 1962. tempts have been made to detect endogenous formation of 10. Homburger, F., Kelley, T., and Friedler, G. Nutritional Factors carcinogenic PAH in animal tissues, especially in tumors and Modifying Hepatic Adenomatosis Induced by Safrole body fluids of cancerous subjects, but these attempts have so (4-allyH,2-methylenedioxybenzene). Proc. Am. Assoc. Cancer far been unsuccessful (2). In plants, however, the natural Res., 3: 236,1961. occurrence of PAH has recently been demonstrated by Graef 11. Homburger, F., Kelley, T., Jr., Friedler, G., and Russfield, A. B. (4), who has extracted plants, grains, and fruits and examined Toxic and Possible Carcinogenic Effects on 4-allyl-l,2-methylene- the fluorescent fractions of their extracts. Benzpyrene was one dioxybenzene (safrole) in Rats on Deficient Diets. Med. ExptL, 4: of the eight hydrocarbons identified in these extracts. "These 1-11,1961. PAH's show a growth-promoting effect in plants and they 12. Long, E. L., and Jenner, P.M. Esophageal Tumors Produced in Rats by the Feeding of Dihydrosafrole. Federation Proc., 22: 275, 1963. seem to function as hormonal foreign bodies and elicit neo plasms," (5). Since plants grown in artificial media and those 13. Long, E. L., Nelson, A. A., Fitzhugh, O. G., and Hansen, W. H. Liver Tumors Produced in Rats by Feeding Safrole. Arch. Pathol. from distant sites had the same content of hydrocarbons, the 75: 595-604, 1963. substances are believed to be endogenous. Graef also believes 14. MacKenzie, I., and Rous, P. The Experimental Disclosure of Latent that this finding may explain why stomach and intestinal Neoplastic Changes in Tarred Skin. J. Exptl. Med., 73: 391-416, tumors occur more often among vegetarians. 1941. 15. Roe, F. J. C. Oil of Sweet Orange: A Possible Role in Carcino- genesis. Brit. J. Cancer, 13: 92-93, 1959. 16. Roe, F. J. C. Some Newly Discovered Tumour-promoting Sub REFERENCES stances. AbhandL Deut. Akad. Wiss. Berlin Kl. Med., 3: 36-52, 1960. 1. Berenblum, I. The Cocarcinogenic Action of Croton Resin. Cancer 17. Roe, F. J. C., and Field, W. E. H. Chronic Toxicity of Essential Oils Res., 1: 44-48, 1941. and Certain Other Products of Natural Origin. Food Cosmetics 2. Boger, E., Formation of Carcinogenic Polynuclear Hydrocarbons. ToxicoL, 5:311-323, 1965. In: P. Bernfeld (ed.), Biogenesis of Natural Compounds, pp. 1065- 18. Roe, F. J. C., and Pierce, W. E. H. Tumor Promotion by Citrus 1078. New York: Pergamon Press, 1968. Oils: Tumors of the Skin and Urethral Orifice in Mice. J. Nati. 3. Field, W. E. H., and Roe, F. J. C. Tumor Promotion in Mouse Cancer Inst., 24: 1389-1403, 1960. Forestomach Epithelium by Orally Administered Citrus Oil. J. Nat. 19. Shubik, P. Studies on the Promoting Phase in the Stages of Cancer Inst., 35: 771-787, 1965. Carcinogenesis in Mice, Rats, Rabbits and Guinea Pigs. Cancer Res., 4. Graef, W. The Natural Occurrence and Significance of Carcinogenic 10: 13-17, 1950. Polycyclic Aromatic Hydrocarbons. Med. Klin. Munich, 60 (15): 20. Van Duuren, B. L. Carcinogenic Epoxides, Lactones, and Hydro- 561-565, 1965. peroxides, In: G. Hogan (ed.), Mycotoxins in Foodstuffs, pp. 275- 5. Graef, W., and Diehl, H. The Natural Normal Levels of Carcino 285. Cambridge, Mass.: M. I. T. Press, 1965. genic Polycyclic Aromatic Hydrocarbons and the Reasons There 21. Van Duuren, B. L., Nelson, N., Orris, L., Palmes, E. D., and for. Arch. Hyg. Bakteriol., 750(1-2): 49-59, 1966. Schmitt, F. L. Carcinogenicity of Epoxides, Lactones and Peroxy 6. Hoch-Ligeti, C. Production of Liver Tumors by Dietary Means - Compounds. J. Nati. Cancer Inst., 31: 41-55, 1963. Effects of Feeding Chilies to Rats. Acta UniÃ³Intern. Contra 22. Zukerman, I. The Effect of Oxidized d-Limonene on Micro Cancrum, 7:606-611,1951. organisms. Nature, 168: 517, 1951.

2374 CANCER RESEARCH VOL. 28

F. Homburger and Eliahu Boger

Cancer Res 1968;28:2372-2374.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/28/11/2372

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/28/11/2372. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.