Counterintuitive Toxicity

Increasingly, scientists are finding that they can't predict a poison's low-dose effects

For decades, researchers largely assumed that a poison’s effects increase as the dose rises and diminish as it falls. However, scientists are increasingly documenting unexpected effects—sometimes disproportionately adverse, sometimes beneficial—at extremely low doses of radiation and toxic chemicals.

Consider the environmentally ubiquitous plastic-softening agent, di-2-ethylhexyl phthalate (DEHP). A German team recently found that in newborn male rats, the lowest DEHP doses tested suppressed the brain activity of an enzyme critical for male development. This was a surprise because higher DEHP doses stimulated that enzyme’s action.

Anderson J.M. Andrade and his colleagues at Charité University Medical School in Berlin note that the enzyme’s suppressive action would have been missed if they had done what most toxicologists do—project low-dose impacts from high-dose tests. The low dose that suppressed aromatase in the rodents was comparable to exposures occurring in the general human population, Andrade’s team reports in the Oct. 29, 2006 Toxicology.

Other toxic agents have unexpectedly beneficial effects. X-rays and gamma radiation are well-recognized carcinogens. Data collected over decades have shown that exposures to 1 gray (Gy)—the dose from perhaps 100 computerized tomography scans—typically increase an individual’s lifetime risk of cancer by 5 percent. However, a growing body of animal data now indicates that lower radiation exposures can defend against cancer-inducing biological changes. Conceptually, it’s analogous to a vaccine.

“The little dose is turning on some kind of protective mechanisms so that when a big dose comes along, it’s not as damaging,” says radiation biologist J. Leslie Redpath of the University of California, Irvine.

Many such effects have been overlooked because researchers prematurely stopped probing for biological impacts as soon as they identified dosage levels of a poison that appear benign, says toxicologist Edward J. Calabrese of the University of Massachusetts–Amherst. Poisons can have a variety of effects at both high and low doses—whether they trigger release of a hormone, switch a gene on or off, or stimulate cell growth. Indeed, Calabrese told Science News, that he has seen the same low dose of a chemical have beneficial effects on one tissue and detrimental effects on another.

He and others worry that if researchers don’t begin regularly probing the effects of these agents at very low doses, scientists will continue to miss important health impacts—both bad and good—of pollutants, drugs, and other agents.

Anomaly or norm?

Regulatory agencies don’t require scientists to evaluate a poison at exposures below that at which no harm is apparent. This dose is referred to as the NOAEL, for “no observable adverse-effects level”.

Calabrese has campaigned relentlessly over the past 15 years to draw attention to biological effects that occur below a NOAEL. These include nonlinear effects, such as a toxicity that initially decreases as concentration goes down but eventually increases, producing a U-shaped curve. In a related class of nonlinear effects, called hormesis, a compound at high doses has an inhibitory—and generally toxic—effect on some biological process but the opposite effect at certain low doses. Unlike many other toxicologists, Calabrese uses the term hormesis to cover most nonlinear low-dose effects.

Radiation offers one of the best examples of hormesis in its narrower definition. At the Environmental Mutagen Society meeting last September in Vancouver, British Columbia, Redpath reported that cells exposed to no more than 0.1 Gy of radiation were less likely to spawn tumors than were cells receiving either far higher doses or no radiation.

In another study, Brenda E. Rodgers of Texas Tech University in Lubbock gave mice a small dose of radiation by caging them in a Ukrainian forest roughly 1.5 kilometers from where the Chernobyl nuclear accident occurred 18 years earlier. Depending on their location, it took between 10 and 45 days for each mouse to receive a dose of 0.1 Gy.

A day after an animal had reached that dose, it was moved to a nearby lab and quickly bombarded with 1.5 Gy. Blood tests showed that the lab radiation produced only half as many chromosome breaks—an indicator of damage that could lead to cancer—in these animals as it did in mice without the earlier low-dose exposure.

A low dose of radiation can reduce damage even if it comes after a larger dose, says Tanya K. Day of Flinders University in Bedford Park, Australia. In one study, her team gave mice a 1-Gy dose of radiation. Four hours later, some mice received a second, far smaller dose. Rodents getting both doses developed only half as many DNA inversions—a particular type of cellular damage—as did mice getting just the first dose, and often fewer inversions than did mice receiving no radiation at all.

Day reported her findings in June at the International Hormesis Society meeting in Amherst, Mass.

Calabrese’s team reviewed hundreds of toxicology papers that document a biological effect below the NOAEL for chemical poisons. He terms all these effects as instances of hormesis, although only about 5 percent of the articles did. The rest described the results as inexplicable or as evidence of some type of nonlinear toxicity.

To determine how commonly trace exposures trigger unanticipated biological impacts, Calabrese’s team has analyzed databases of biological responses to potentially toxic chemicals, each throughout a broad range of doses. In their most recent study, described in the December 2006 Toxicological Sciences, Calabrese and his colleagues analyzed data showing how cell proliferation in 13 different yeast strains responded to various doses of 2,189 potential anticancer drugs.

Almost 80 percent of the drugs exhibited a NOAEL, the team found. Among these, the group further looked for reports of biological effects triggered by doses even lower than that level. The authors had expected that 25 percent of these drugs, just by chance, would exhibit activity above that seen with no exposure. In fact, 60 percent did.

The effects observed at those low doses were modest, perhaps 60 percent higher or lower than those that occur in the absence of any exposure, Calabrese notes. He acknowledges that such changes might not always have clinical significance.

These findings and earlier analyses by his group, Calabrese says, show that measurable biological effects at low doses appear to be more the norm than an anomaly.

Indeed, even pollutants that don’t have a NOAEL may have nonlinear effects at low doses, notes Bernard Weiss of the University of Rochester (N.Y.) School of Medicine and Dentistry. For example, the drop in a child’s IQ for each 1 microgram per deciliter of lead in the blood is much higher at concentrations below 10 µg/dl than at concentrations above that value (SN: 5/5/01, p. 277: Available to subscribers at Even low lead in kids has a high IQ cost).

So, Weiss concludes, toxicity estimates based on high-dose measurements greatly underestimate low-dose harm.

How does it work?

Scientists have recently begun to discover mechanisms to explain hormesis and other nonlinear dose responses. For instance, Rodgers has been looking at what genes are preferentially turned on or off in the mice exposed to Chernobyl radiation. Compared with unexposed mice, those caged in the Ukraine forest had 600 to 1,200 genes whose activity had been altered.

“We expected to see an increase in the expression of genes involved in DNA repair,” Rodgers says. “What we found instead was an increase in the expression of genes that respond to oxidative stress—such as free radicals.”

Another explanation of hormesis was suggested in 2000 by researchers working with human-cancer cells exposed to epigallocatechin gallate (EGCG), the principal cancer-fighting ingredient in green tea. The team showed that although high-dose exposures of EGCG inhibited cell growth, low doses stimulated cell proliferation. D. James Morré of Purdue University in West Lafayette, Ind., says that his team several years ago found a unique enzyme on cell surfaces that appeared to be “a molecular target for chemical hormesis.”

The group subsequently determined that this enzyme can bind to various substances, in addition to EGCG, and alter their cellular effects. Those responses disappeared when the enzyme was inactive.

Some scientists have suggested additional processes that play a role in hormesis. In an upcoming issue of the International Journal of Low Radiation, Bobby R. Scott of the Lovelace Respiratory Research Institute in Albuquerque, N.M., and his colleagues report that low doses of radiation induce mild oxidative stress in cells, activating a high-efficiency form of DNA repair and stimulating the immune system. This stress also “activates a special apoptosis [cell suicide] process”—one that culls genetically unstable cells, he says.

Scott suggests that these same processes probably work to counteract chemical poisons.


Although most toxicologists today agree that hormesis occurs—a big change from a decade ago—some argue that Calabrese and his team greatly overstate its frequency. A major portion of this controversy hinges on differences in the use of the term hormesis.

“I totally believe that [nonlinear] low-dose responses occur frequently,” says Kristina A. Thayer of the National Institute of Environmental Health Sciences in Research Triangle Park, N.C. “In fact, I have no problem accepting that most of the time they might be stimulatory.”

However, she says that Calabrese equates stimulatory low-dose effects with benefits when there’s no reason to expect that they would necessarily be beneficial. Her research with the plastic-softening agent bisphenol A, a hormone-mimicking agent, illustrates a detrimental effect of low-dose stimulation similar to what Andrade found for DEHP.

Among toxic agents that show positive biologic effects at low doses, Calabrese sees the possibility for better drug design. For example, he says, current treatments for dementia provide tiny doses of drugs that at high doses would be toxic. For instance, he says, “every Alzheimer’s drug on the market today acts via hormetic [low-dose] activities.”

Even though a hormetic treatment may show only a small effect, Calabrese proposes that several treatments might be put together to achieve a therapeutic benefit.

Scott suggests a related therapeutic application of hormesis that uses small doses of radiation to trigger immunological and cell-death processes. However, cancer cells are “reluctant” to undergo programmed death, Scott notes. Because certain compounds—such as resveratrol, a polyphenol in grapes (SN: 11/4/06, 293: L’Chaim: Wine compound lengthens mouse lives)—sensitize cancer cells to radiation, Scott envisions pretreating people with such compounds and following this up with a hormetic dose of radiation. “For lung cancer,” he says, “perhaps just low-dose diagnostic X rays would do.”

Beyond new medical applications, information gleaned from research into low-dose exposures might help fine-tune regulation of chemicals. Scientists may find that many pollutants aren’t as toxic at low doses as has been assumed, Calabrese says.

“You can imagine why industry loves hormesis,” Weiss says. It suggests pollution may not need to be cleaned up as thoroughly as regulations have been asking for.

Calabrese counters, however, that if traces of certain pollutants are not as dangerous as earlier estimates had suggested, why not investigate whether some regulations are unduly strict?

Indeed, proving that some low-dose exposures are “of no regulatory concern could make a qualitative difference in regulations,” observes economist Lester B. Lave of Carnegie Mellon University in Pittsburgh. However, he adds that to justify changing guidelines for regulations, far more research would be needed.

For instance, there has been much discussion suggesting that low doses of chemicals—even pollutants—might rev up immunity in a beneficial way. However, because many people have compromised immune systems, Lave says that before raising the acceptable environmental limits of a pollutant, “I’d want to know if we see a [beneficial] hormetic response in those people, or babies with undeveloped immune systems, or the elderly.” Moreover, he says, effects at low doses tend to be subtle, so “I’d want to see them documented in humans, not just animals”—and to know at precisely what dose they turn detrimental.

Jonathan Borak, a toxicologist at the Yale School of Medicine in New Haven, Conn., agrees that it’s too early for hormesis or other nonlinear low dose–effects data to be “practically relevant” for altering regulatory or health policy.

Although “I believe hormesis is real, it is fundamentally difficult—and expensive—to demonstrate,” Borak says. Looking for relatively small low-dose effects could quadruple the cost of toxicology studies, he estimates, underscoring “practical and economic reasons why today almost nobody looks for them.”

Janet Raloff is the Editor, Digital of Science News Explores, a daily online magazine for middle school students. She started at Science News in 1977 as the environment and policy writer, specializing in toxicology. To her never-ending surprise, her daughter became a toxicologist.

More Stories from Science News on Earth