Nature's Hormone Factory:

Endocrine Disruptors in the Natural Environment

by Jonathan Tolman


For millions of years plants have been quietly producing chemicals. Through countless generations they have been perfecting a potpourri of chemicals, some benign, some deadly. As the ability to detect, isolate, measure and test chemicals found in nature has progressed, a startling fact has emerged: hundreds of plants appear to produce endocrine disrupters - which affect hormonal balance in animals.

Many of the plants that produce phytoestrogens and other endocrine disrupters are edible. In laboratory tests, more than 43 plants and foods found in the human diet have been shown to be oestrogenically-active. Many phytoestrogen-containing plants are common elements of our diet. Such grains as corn and wheat form a significant part of the human diet. Many legumes have also shown a surprising capacity for phytoestrogen production.

Although much more work has been done on phytoestrogens, some work has been done on plant chemicals which are known to affect the production of sperm. Only a handful have been discovered in the human diet. Of these, the most common is cottonseed oil. Although cottonseed oil is rarely sold as a vegetable oil, it is commonly used in manufactured snack foods.

A great deal of attention has recently been given to the fact that synthetic chemicals have exhibited oestrogenic effects in laboratory studies. The chemicals most prominently cited are PCBs and DDT, both of which have been banned in most developed countries. Compared with phytoestrogens, the concern over synthetic oestrogens may be somewhat overstated. The oestrogenic effects from the phytoestrogens in our diet are an estimated 40 million times greater than those from synthetic chemicals. To date, however, there is no concrete evidence that either pose a risk to human health.

The Silent Spring

For sheep ranchers spring is a busy time. As the ewes in their flock begin to lamb there is a great deal of work to do to ensure the next generation of healthy sheep. But in the 1940s the sheep ranchers of Australia began to notice a peculiar and frightening trend. At first there was a rash of stillborn lambs. Then the ewes became sterile. Each spring there were fewer and fewer lambs. For the ranchers it was literally a silent spring. By the mid forties the sheep ranching industry in Australia was in a state of crisis and faced certain financial ruin unless the cause of the mysterious infertility in the sheep could be found. What could be causing this disastrous sterility. Genetic mutations? Radiation? Poisonous chemicals?

The Australian Department of Agriculture was called in. A cadre of veterinarians and scientists investigated all possible sources of the sterility. By 1946 they had discovered the source of the sheep's sterility - clover.

The ranchers were unaware that the innocuous-looking clover (Trifolium subterraneum) that they had recently begun planting in their fields to feed their sheep, had been producing large quantities of oestrogen-mimicking compounds. It took another decade before scientists finally pinned down the exact chemicals which were causing the sterility, genistein and formononetin. A structural comparison for genistein and formononetin shows why they were causing the sterility. They are surprisingly similar in structure to oestrogen and synthetic oestrogen, diethylstilboestrol, DES. These two compounds mimic the steroidal nucleus of the natural female hormone oestrogen. Although they are in fact rather weak oestrogens, the plants make up for that fact by producing them in comparatively huge quantities: 5% of dry weight in the clover fodder.

The oestrogen-mimicking nature of chemicals found in plants isn't restricted to the clover T. subterraneum. A survey of clover found that 18 different species of the plant produced oestrogen-mimicking substances, or "phytoestrogens," in quantities as high as those found in T. subterraneum.

Moreover, sheep are not the only animals on which the reproductive effects of phytoestrogens have been observed. Oestrogenic effects also have been observed in quail which feed on pastures rich in leguminous species. In years of good rainfall, legumes that are eaten grow luxuriously and are relatively low in phytoestrogens. However, in drought years the levels of phytoestrogens are increased with respect to the weight of the leaves. Consequently, egg laying by female quails is curtailed. There appears to be a self-regulating mechanism whereby the increase in quail population is kept low when food availability is limited. In other words, in order to enhance their survival, the plants reduce the number of quail in the next generation.

As the presence of phytoestrogens in clover and legumes became known, scientists began to wonder if other plants were also producing hormone mimickers which could disrupt the reproduction of animals. And even more importantly, were any of these plants in the human diet?

An answer to this question came from another area half way around the world from where scientists were struggling with the fertility of sheep, in Tibet. A clue emerged from the fact that in the history of Tibet the population has been extremely stable, often for as long as 200 years. During those times the Tibetan diet largely consisted of barley and peas. Could the peas or barley be affecting the fertility of the Tibetans?

When scientists fed mice a diet consisting of 20% peas, litter sizes dropped by 50%. When the mice were fed diets consisting of 30% peas, the mice failed to produce any young at all.

The Original Chemistry Industry

As the ability to detect, isolate, measure and test chemicals found in nature has progressed, a starting fact has emerged: plants are phenomenal chemical manufacturers. For millions of years plants have been quietly producing chemicals. Through countless generations they have been perfecting a potpourri of chemicals for a host of objectives. Some chemicals are used to directly deter animals from eating the leaves or seeds of the plants. Strychnine, a very effective poison, comes from the seed of Strychnos toxifera. Hundreds of plants produce an astounding array of powerful poisons, from exotic tropicals such as curare to the common flower foxglove, all deadly if eaten.

Producing deadly toxins is a rather straightforward method for a plant to avoid becoming lunch. Some plants, however, have chosen more subtle methods of chemical combat, not death but disruption - reproductive disruption.

Over the last few decades scientist have been analysing plants and the chemicals they produce. Most of this research is aimed at finding new and useful drugs. One consequence of all these studies has been that a large number of plants have been studied for their reproductive effects. A survey of literature compile by the University of Illinois at Chicago's Department of Medical Chemistry, showed that 149 chemicals have been isolated from plants and have demonstrated oestrogenic effects in laboratory studies. A total of 173 plants have been shown to have active oestrogenic effects. (See Appendix A)

Although much research has focused on oestrogenic effects, antispermatogenic effects have also been studied. 55 compounds extracted from plants have been shown to inhibit the production of sperm. And 31 plant species have been tested and shown to be anti-spermatogenic. (See Appendix B)

Lab Testing of Phytoestrogens

The reproductive effect of low levels of genistein has been tested in laboratory mice beginning in the 1950s. Mice fed artificially elevated levels of genistein in the diet suffered decreased reproduction compared with control populations. In the control group, 82% of the females produced litters. A second group was fed a diet which consisted of 0.2% genistein. Of this second group, only 59% of the females produced litters. A third group was fed commercial soya bean meal which contained 0.1% genistein, of which 77% of the females gave birth:

Other early studies have also been conducted on the oestrogenic effects of soya bean oil. In these tests, the oestrogenic effect was measured in terms of increased uterine weight in mice. One such study compared a host of different processed oils, both edible and non-edible. The oils were mixed into the feed to constitute 10% of the diet. Of the two soya bean oil samples used, one increased the mean uterine weight 77% and the other 86%. Only two other oils produced more oestrogenic responses in the mice, coconut oil and rice bran oil.

Since these early studies, additional tests have continued to determine the reproductive effect of a number of phytoestrogens. Many of the most recent studies have used new-born mice to assess how these substances effect the development of the reproductive tract. One such study, of another common phytoestrogen - coumestrol, found that: "Neonatal coumestrol treatment is effective in causing a number of morphological alterations in the female reproductive tract."

Other studies have been conducted with other phytoestrogens in foetal and neonatal mice. One such study concluded, "Genistein influences oestrogen-dependent development by modifying both morphological and neuroendocrine endpoints." In other words, phytoestrogens such as genistein and coumestrol, act as endocrine disrupters.

Table 1



Phytoestrogens in Food

Although hundreds of plants appear to produce endocrine disrupters, the concern for human exposure is limited principally to those plants in the human diet. Unfortunately, dozens of edible plants produce phytoestrogens and other endocrine disrupters. In laboratory tests more than 43 plants found in the human diet have been shown to be oestrogenically active (see Table 1)

Many phytoestrogen containing plants are common elements of our diet. Such grains as corn and wheat form a large part of our diet. For example, Many legumes have shown a surprising capacity for phytoestrogen production, it was not unexpected that phytoestrogens would be found in the legumes we eat.

But the soya bean has garnered particular attention. Soya beans are the third largest crop in the United States, Every year nearly 60 million acres of soya beans are planted in the United States alone. Nearly 2 billion bushels of soya beans are harvested every year.

Numerous studies have shown that soya beans contain significant levels of the phytoestrogens,

genistein and daidzein. Dry soya beans on average contain 1,107 milligrams of genistein per kilogram and 846 milligrams of daidzein per kilogram. If one considers only the U.S. crop of soya beans and only the amount found in the seeds, the total quantity of genistein produced each year is roughly 130 million pounds, and an additional 100 million pounds of daidzein.

Although some of the soya beans end up as animal feed, much of the genistein and daidzein is consumed by humans. When most people think of soya bean foods they generally think of tofu or soya sauce. But soya products are not restricted to a few occasionally eaten foods: they are ubiquitous in the modern western diet. Soya flour, soya protein and most importantly, soya bean oil are found in hundreds of products eaten every day.

The Sprouting of the Bean

At the turn of the century, soya beans were a virtually unknown crop in the United States. By 1924, when the U.S. Department of Agriculture started publishing systematic annual statistics 448,000 acres of soya beans were harvested. Back then the yield per acre was also low, a mere 11 bushels per acre. Since then the acreage of soya beans harvested and the yield per acre have soared.

By the end of World War II, acreage had climbed to 10 million acres a year and the yield per acre had nearly doubled. As food processing modernised more and more uses were being found for the soya bean and demand continued to increase. In 1979, soya bean production peaked with 70 million harvested acres producing 2.3 billion bushels. Since then production has continued to hover around 2 billion bushels a year.

The majority of soya beans grown today are crushed for their oil. Soya bean oil is by far the most common of the vegetable oils. Roughly 80% of the vegetable oil consumed in the United States is soya bean oil. Annual per capita consumption of soya bean oil is an estimated 65 pounds per year. For instance, most fried snack foods are manufactured with vegetable oil, predominantly soya bean oil. Virtually all infant formulas are a mixture of dairy and soya solids and proteins.

Antispermatogens Found in Food

Although much more work has been done on phytoestrogens, some work has been done on plant chemicals which are known to effect the production of sperm. Only a handful have been discovered in the human diet (see Table 2).

Of these the most common is cottonseed oil. Although cottonseed oil is rarely sold as a vegetable oil, it is commonly used in manufactured snack foods.

Table 2

Common foods from plants or containing chemicals shown to have antispermatogenic activity in laboratory experiments


Since the 1970s most of the work on the antispermatogenic effects of cottonseed oil has been conducted in China, after Chinese authorities noted decreased fertility in some provinces in the late 1950s and eventually linked it with the use of cottonseed oil in cooking. Since then laboratory and human experiments have been conducted using cottonseed oil. In one study rats fed 0.5 millilitres of cottonseed oil a day produced no viable sperm at the end of the 28-day test period.

Xeno-Oestrogens: Cause for Alarm?

A great deal of attention has recently been given to the fact that synthetic chemicals have exhibited oestrogenic effects in laboratory studies. The chemicals most prominently cited are PCBs and DDT, both of which have been banned in the United States. Compared with the phytoestrogens, the concern over synthetic oestrogens may be somewhat overstated.

Although they are ubiquitous in our diet, phytoestrogens are considered only weakly active. Most of the compounds which have been identified and tested have been found to have a relative potency when compared to synthetic oestrogen of 0.001 to 0.0001. In other words it takes between 1,000 and 10,000 molecules of these phytoestrogens to create the same effect as one synthetic oestrogen molecule. The vast majority of phytoestrogens appear to occur in the family of chemicals known as flavonoids. Total flavonoid consumption in the human diet is estimated at approximately 1 gram per day. Consequently the total daily oestrogenic effect of phytoestrogens could be estimated at roughly 100 micrograms of oestrogen equivalents (See Table 3).

Table 3

Natural v Synthetic Oestrogens

Source of OestrogenOestrogen Equivalent (µg/day)
Birth control Pill18675
Postmenopausal therapy3350
Phytoestrogens in food102

Synthetic oestrogen-mimicking substances, or xeno-oestrogens are considered even less oestrogenically active than the phytoestrogens. Oestrogenically active pesticides such as DDT, Dieldrin, and Endosulfan have been assigned a relative potency of 0.000001. In other words it takes one million xeno-oestrogen molecules to have the same effect as one synthetic oestrogen molecule. Since total intake of these xeno-oestrogens is significantly lower than the naturally occurring phytoestrogens, (around 2.5 micrograms per day) the oestrogen exposure from synthetic chemical has been estimated at 0.0000025 micrograms of oestrogen equivalents per day. In other words, the estimated oestrogenic effects from the phytoestrogens in our diet are 40 million times greater than those from synthetic chemicals, but it is questionable that either affect human health.

It is widely recognised that results obtained in vitro (in a test tube) are not necessarily reproduced in vivo (in the body). For example, oestrdiol, taken by mouth has no activity, it simply passes through the body and is excreted; it is only in an "active" form (e.g. ethynylestradiol - a constituent of the birth control pill), that is can be absorbed by the body. However, a screening test on oestradiol in vitro would show a high degree of activity. This is why we need to be cautious of such tests.

Plant Defficiency Theory

A number of hypotheses have been developed to explain the functioning of plant defensive systems. Two of these hypotheses in particular bear examination. Because defences are costly to produce:

1. Less well-defended individuals have greater fitness than more highly-defended individuals, when enemies are absent.

2. Commitment to defence is decreased when enemies are absent and increased when attacked.

  • A sizeable body of literature has been accumulated which supports both of these hypotheses. For example, varieties of insect resistant soya bean produce a lower yield of seeds and accumulate nitrogen more slowly than insect susceptible varieties in the absence of herbivores including insects.
  • In the case of the second hypothesis, numerous studies have been conducted actually measuring the increased amounts of secondary metabolites. In the species Senecio jacobaea (Toyon), when half of the leaves were removed the plant responded within two day by increasing the amount of total leaf alkaloids and N-oxides in the remaining leaves, 40 to 47%.
  • In another case, when beet plants were infested by beet flies, with 24 days the mortality of the beet flies increased between 29 and 100%. The ability of beets to respond to infestation and in some cases kill all of the infesting insects shows the extreme effectiveness of some secondary plant metabolites and the plant's ability to defend itself. (See Appendix)
  • Risk vs. Risk Analysis for Crop Production

    Current US government policy only rarely regulates naturally occurring chemicals in the food supply. In only two cases has the US Food and Drug Administration (FDA) directly regulated naturally occurring carcinogens. In the first instance, the FDA banned the chemical safrole, which had been used as a natural flavouring for root beer, after it was determined that safrole was highly carcinogenic in high-dose rat studies.

    The second case of federal regulation of a natural chemical is aflatoxin. The FDA has established a standard of 20 parts per billion of aflatoxin for a variety of foods susceptible to aflatoxin poisoning, most commonly peanuts. In setting its aflatoxin standard the FDA engaged in a type of risk versus risk analysis.

    The FDA recognised that it would be virtually impossible to eliminate aflatoxin from the food supply. To do so would have required the massive use of fungicides which may present other equally hazardous long term health effects. The question which faced the agency was what level of aflatoxin contamination should be tolerated to maximise human health. Whether or not the agency came to the correct answer is difficult to determine. However, the fact that the agency considered the health effects of both aflatoxin and prophylactic fungicides increases the chance of establishing a better health standard.

    This type of risk v. risk analysis could benefit other regulatory programs effecting agriculture and the production of food. Current pesticide regulation, for example, does not take into account the potential of pesticides applied on crops to inhibit the formation of the plant's own secondary metabolites.

    More importantly, as biotechnology and the ability to manipulate the genetic code of plants increases, the effects of secondary plant metabolites may increase geometrically.

    A Final Note

    More than fifty years after discovering the fertility problems in sheep in Australia, ranchers continue to struggle with clover disease. Despite decades of research and the attempted introduction of more benign types of clover, clover disease is still responsible for the loss of 1 million lambs every year.

    Had clover disease struck herds of sheep in the middle ages or even in the 17th century the source of this infertility would likely have been blamed on a witch. And no doubt if the infertility continued the shepherds would likely have rounded up some socially unpopular woman with few friends to defend her, held a mock trial, called in religious authorities to absolve their collective conscience and then burned her at the stake.

    Although science has done much to dispel myths and witches, there still exists in humans the urge to blame the socially unpopular for the ills of humanity. In the modern sense it is easy for us to blame a witch, chemical or otherwise. But the responsible and far more laborious task is to determine what actually is causing the problem which is often made even more difficult when we discover, as the sheep ranchers in Australia have, that even decades of science can't always solve the problem.

    The full document is available upon request from ESEF

    About the Author

    Jonathan Tolman is an environmental Policy Analyst at the Competitive Enterprise Institute in Washington DC, which published an earlier version of this paper. His most recent work has been on water pollution and agricultural issues. He is the author of "Federal Agricultural Policy: A Harvest of Environmental Abuse," and "Gaining More Ground: An Analysis of Wetlands Trends in the United States." Prior to working with CEI, he was Associate Producer of the weekly television show, 'TechnoPolitics'. In 1991, Jonathan served as Special Assistant for the President's Council on Competitiveness, focusing on environmental and natural resource regulation. In 1992, he worked for the White House as an environmental analyst in the Office of Policy Development.

    news key contacts links publications