Dietary Pesticides (Bruce Ames)

ABSTRACT

The toxicological significance of exposures to synthetic chemicals is examined in the context of exposures to naturally-occurring chemicals.
We calculate that 99.99% (by weight) of the pesticides in the American diet are chemicals that plants produce to defend themselves. Only
52 natural pesticides have been tested in high-dose animal cancer tests, and about half (27) are rodent carcinogens; these 27 are shown to be
present in many common foods. We conclude that natural and synthetic chemicals are equally likely to be positive in animal cancer tests. We
also conclude that at the low doses of most human exposures the comparative hazards of synthetic pesticide residues are insignificant.

Toxicological examination of synthetic chemicals such as pesticides and industrial pollutants, without similar examination of the chemicals in
the natural world to use for comparison, has generated an imbalance in both data and perception about potential hazards to humans (1-6). In
this and two accompanying papers (7,8), we try to redress this imbalance and discuss in detail one major group of natural chemicals in our
diet – nature’s pesticides.

About half of all chemicals (whether natural or synthetic) tested chronically in animal cancer tests at the maximum tolerated dose (MTD)
are carcinogens (7,9-14). The MTD of the test chemical is a near-toxic dose that can cause chronic mitogenesis, often as a result of cell
killing (7). We have argued that mitogenesis increases mutagenesis, and therefore that a high percentage of all chemicals might be expected
to be carcinogenic when tested chronically at the MTD (7). A high proportion of both natural and synthetic test chemicals is positive for
carcinogenicity. Natural chemicals constitute the vast bulk of chemicals in the human diet, and therefore should be used as a reference for
evaluating possible carcinogenic hazards from synthetic chemicals. In recent years, we have compared the possible hazards of various
rodent carcinogens, using the HERP ratio: human exposure/rodent potency (1,6,). It should be emphasized that as the understanding of
carcinogenesis mechanisms improves, these comparisons can be refined but they cannot provide a direct estimate of human hazard. This
paper does not extend the HERP comparisons (1) because our purpose is different and space does not allow a proper analysis.

NATURE’S PESTICIDES: MUTAGENICITY AND CARCINOGENICITY

“Plants are not just food for animals. . . The world is not green. It is colored lectin, tannin, cyanide, caffeine, aflatoxin, and canavanine”

[Janzen (16)].

Dietary Pesticides are 99.99% All Natural. Nature’s pesticides are one important subset of natural chemicals. Plants produce toxins to
protect themselves against fungi, insects, and animal predators (5,16-23). Tens of thousands of these natural pesticides have been
discovered, and every species of plant analyzed contains its own set of perhaps a few dozen toxins. When plants are stressed or damaged,
such as during a pest attack, they may greatly increase their natural pesticide levels, occasionally to levels that can be acutely toxic to
humans. We estimate that Americans eat about 1.5 g of natural pesticides per person per day, which is about 10,000 times more than they
eat of manmade pesticide residues (see below). As referenced in this paper (see 16-21 and Legends to Table 1, 2) there is a very large
literature on natural toxins in plants and their role in plant defenses. The human intake of these toxins varies markedly with diet and would be
higher in vegetarians. Our estimate of 1.5 g of natural pesticides per person per day is based on the content of toxins in the major plant foods
(e.g. 13 g roasted coffee per person per day contains about 765 mg of chlorogenic acid, neochlorogenic acid, caffeic acid, and caffeine; see
(22,23) and Table 2). Phenolics from other plants are estimated to contribute another several hundred mg of toxins. Flavonoids and
glucosinolates account for several hundred mg; potato and tomato toxins may contribute another hundred, and saponins from legumes
another hundred. Grains such as white flour and white rice contribute very little, but whole wheat, brown rice, corn (maize) may contribute
several hundred mg more. The percentage of a plant’s weight that is toxin varies, but a few percent of dry weight is a reasonable estimate:
e.g. 1.5% of alfalfa sprouts is canavanine and 4% of coffee beans is phenolics. However, the percentage in some plant cultivars is lower,
e.g. potatoes and tomatoes.

Concentrations of natural pesticides in plants are usually measured in parts per thousand or million (16-23) rather than parts per billion (ppb),
the usual concentration of synthetic pesticide residues or of water pollutants (1,24). We estimate that humans ingest roughly 5,000 to 10,000
different natural pesticides and their breakdown products (16-23). For example, Table 1 shows 49 natural pesticides (and metabolites) that
are ingested when eating cabbage and indicates how few have been tested for carcinogenicity or clastogenicity. Lima beans contain a
completely different array of 23 natural toxins that, in stressed plants, range in concentration from 0.2 to 33 parts per thousand fresh weight;
none appears to have been tested yet for carcinogenicity or teratogenicity (19). Many Leguminosae contain canavanine, a toxic arginine
analog that, after being eaten by animals, is incorporated into protein in place of arginine. Feeding alfalfa sprouts (1.5 % canavanine dry
weight) or canavanine to monkeys causes a lupus erythematosus-like syndrome (44). Lupus in man is characterized by a defect in the
immune system that is associated with autoimmunity, antinuclear antibodies, chromosome breaks, and various types of pathology. The
toxicity of non-food plants is well known: plants are among the most commonly ingested poisonous substances for children under five.

Surprisingly few plant toxins have been tested for carcinogenicity (10-13,45). Among 1052 chemicals tested in at least one species in chronic
cancer tests, only 52 are naturally occurring plant pesticides (10-13). Among these, about half (27/52) are carcinogenic. Even though only a
tiny proportion of plant toxins in our diet has been tested so far, the 27 natural pesticides that are rodent carcinogens are present in the
following foods: anise, apple, apricot, banana, basil, broccoli, Brussels sprouts, cabbage, cantaloupe, caraway, carrot, cauliflower, celery,
cherries, cinnamon, cloves, cocoa, coffee, collard greens, comfrey herb tea, currants, dill, eggplant, endive, fennel, grapefruit juice, grapes,
guava, honey, honeydew melon, horseradish, kale, lentils, lettuce, mango, mushrooms, mustard, nutmeg, orange juice, parsley, parsnip, peach,
pear, peas, black pepper, pineapple, plum, potato, radish, raspberries, rosemary, sesame seeds, tarragon, tea, tomato, and turnip. Thus, it is
probable that almost every fruit and vegetable in the supermarket contains natural plant pesticides that are rodent carcinogens. The levels of
these 27 rodent carcinogens in the above plants are commonly thousands of times higher than the levels of manmade pesticides. Table 2
shows a variety of natural pesticides that are rodent carcinogens occurring in the ppm range in plant foods.

The catechol-type phenolics such as tannins, and caffeic acid and its esters (chlorogenic and neochlorogenic acids), are more widespread in
plant species than other natural pesticides (e.g., Table 1 and 2). It may be that these phenolics have an antimicrobial role analogous to the
respiratory burst of oxygen radicals from mammalian phagocytic cells. The phenolics oxidize when a plant is wounded, yielding a burst of
mutagenic oxygen radicals, e.g. the browning when an apple is cut.

Caution is necessary in interpreting the implications of the occurrence in the diet of natural pesticides that are rodent carcinogens. It is not
argued here that these dietary exposures are necessarily of much relevance to human cancer. Indeed, a diet rich in fruit and vegetables is
associated with lower cancer rates (86,87). This may be because anticarcinogenic vitamins and antioxidants come from plants (86,87). What
is important in our analysis is that exposures to natural rodent carcinogens may cast doubt on the relevance of far lower levels of exposures
to synthetic rodent carcinogens.

Residues of Manmade Pesticides. A National Research Council report has discussed the regulation of synthetic pesticides that are rodent
carcinogens, but ignored natural pesticides (88). The U.S. Food and Drug Administration (FDA) has assayed food for 200 chemicals
including the manmade pesticide residues thought to be of greatest importance and the residues of some industrial chemicals such as
polychlorinated biphenyls (PCBs) (24). FDA found residues for 105 of these chemicals: the U.S. intake of the sum of these 105 chemicals
averages about 0.09 mg per person per day, which we compare to 1.5 g of natural pesticides (i.e. 99.99% natural). Other analyses of
synthetic pesticide residues are similar (90). About half (0.04 mg) of this daily intake of synthetic pesticides is composed of 4 chemicals (24)
that were not carcinogenic in rodent tests: ethylhexyl diphenyl phosphate, chlorpropham, malathion, and dicloran (10,89). Thus, the intake of
rodent carcinogens from synthetic residues is only about 0.05 mg a day (averaging about 0.06 ppm in plant food) even if one assumes that all
the other residues are carcinogenic in rodents (which is unlikely).

Cooking food. The cooking of food is also a major dietary source of potential rodent carcinogens. Cooking produces about 2 g per person per
day of mostly untested burnt material that contains many rodent carcinogens, e.g. polycyclic hydrocarbons (81,91) heterocyclic amines
(92,93), furfural (22,23), nitrosamines and nitroaromatics (1,94), as well as a plethora of mutagens (91-95). Thus, the number and amounts of
carcinogenic (or total) manmade pesticide residues appear to be minimal compared to the background of naturally-occurring chemicals in the
diet. Roasted coffee, for example, is known to contain 826 volatile chemicals (22); 21 have been tested chronically and 16 are rodent
carcinogens (10-13); caffeic acid, a non-volatile rodent carcinogen, is also present (Table 2). A typical cup of coffee contains at least 10 mg
(40 ppm) of rodent carcinogens (mostly caffeic acid, catechol, furfural, hydroquinone, and hydrogen peroxide)(Table 2). The evidence on
coffee and human health has been recently reviewed, and the evidence to date is not sufficient to show that coffee is a risk factor for cancer
in humans (81,86). The same caution about the implications for humans of rodent carcinogens in the diet that were discussed above for
nature’s pesticides apply to coffee and the products of cooked food.

Clastogenicity and Mutagenicity Studies. Results from in vitro studies also indicate that the natural world should not be ignored and that
positive results are commonly observed in high-dose protocols. For example, Ishidate et al. (26) reviewed experiments on the clastogenicity
(ability to break chromosomes) of 951 chemicals in mammalian cell cultures. Of these 951 chemicals, we identified 72 as natural plant
pesticides: 35 (48%) were positive for clastogenicity in at least one test. This is similar to the results for the remaining chemicals, of which
467/879 (53%) were positive in at least one test.

Of particular interest are the levels at which some of the carcinogenic plant toxins in Table 2 were clastogenic (26) a) Allyl isothiocyanate
was clastogenic at a concentration of 0.0005 ppm, which is about 200,000 times less than the concentration of sinigrin, its glucosinolate, in
cabbage. Allyl isothiocyanate was among the most potent chemicals in the compendium (26), and is also effective at unusually low levels in
transforming (96) and mutating animal cells (30). (See also the discussion of cancer tests in Table 1.) b) Safrole was clastogenic at a
concentration of about 100 ppm, which is 30 times less than the concentration in nutmeg, and roughly equal to the concentration in black
pepper. The rodent carcinogens safrole and estragole, and a number of other related dietary natural pesticides that have not been tested in
animal cancer tests, have been shown to produce DNA adducts in mice (97). c) Caffeic acid was clastogenic at a concentration of 260 and
500 ppm, which is less than its concentration in roasted coffee beans and close to its concentration in apples, lettuce, endive, and potato skin.
Chlorogenic acid, a precursor of caffeic acid, was clastogenic at a concentration of 150 ppm, which is 100 times less than its concentration in
roasted coffee beans and similar to its concentration in apples, pears, plums, peaches, cherries and apricots. Chlorogenic acid and its
metabolite caffeic acid are also mutagens (Table 1). The genotoxic activity of coffee to mammalian cells has been demonstrated (98).

The carcinogenicity and mutagenicity of many plant pesticides have been recently reviewed (45): 5- and 8-methoxypsoralen are light
activated mutagens (17); benzyl acetate and ethyl acrylate mutate mouse lymphoma cells (30). Plant phenolics such as caffeic acid,
chlorogenic acid, and tannins (esters of gallic acid) have been reviewed for their mutagenicity and anti-mutagenicity, clastogenicity, and
carcinogenicity (99).

ACKNOWLEDGMENT

We are indebted to R. Peto, N. B. Manley, T. H. Slone, C. Wehr, R. Beier, L. W. Wattenberg, R. Hall, T. Jukes, G. R. Fenwick, J.
Caldwell, J. Duke, C. VanEtten, D. Freedman, R. Prokopy, and N. Ito. This work was supported by National Cancer Institute Outstanding
Investigator Grant CA39910, by National Institute of Environmental Health Sciences Center Grant ES01896; Contract No.
DE-AC03-76SF00098: Director, Office of Energy Research, Office of Health and Environmental Research, Division of the U.S.
Department of Energy. We dedicate this paper to the memory of William Havender.