I learned – Why we’re wrong about toxic chemicals

Why didn’t he die?

In 2004, Ukranian presidential candidate Viktor Yushchenko ate a meal laced with a molecule called 2,3,7,8-Tetrachlorodibenzodioxin (TCDD), more commonly referred to as “dioxin”.

TCDD has been labeled one of “the most toxic chemicals known to science“.   The public first became aware of TCDD as a contaminant in Agent Orange, an herbicide used during the Vietnam War.  Subsequent testing showed that concentrations as small as five parts per trillion – less than one drop diluted in a million gallons of water – could cause cancer in laboratory rats.   Concentrations of 50 parts per billion (ppb) would cause rapid signs of toxicity and early death.  It has been associated with toxicity in every animal assay studied.

Victor Yuschchenko was exposed to approximately 2 mg of dioxin, giving him a total body load of about 25 ppb.  He developed pancreatitis, nausea, and facial scarring known as chloracne.  His blood contained the second highest concentration of dioxin ever measured in a human.

Yet he survived.

viktor_yuschenkoViktor Yushchenko scarred by chloracne, in 2006.  Photo courtesy Muumi/Wikipedia.

Was Yuschchenko lucky?  Blessed by God?  Or did his would-be assassins make a flawed assumption about how toxicity works?

TCDD has an LD50 – the lethal dose at which 50% of animals are killed – as low as 0.6 ppb, when tested in Hartley guinea pigs.  That is simply off-the-scale toxic, and it led to a near panic about contamination by TCDD.  Concern about TCDD made headlines and helped prompt the evacuation of the polluted Love Canal neighborhood near Niagara Falls, NY in the 1970s.  Protests over trace TCDD produced in waste incineration led to dramatic changes in regulation in the US and Europe.

Less noted, however, was that the lethal dose of TCDD in Golden Syrian hamsters was as high as 5000 ppb, making hamsters 10,000-fold less sensitive than their guinea pig cousins.  While TCDD is clearly still nasty stuff for hamsters, on par with cyanide as a toxin, it is not close to the most toxic material ever known.

And for TCDD, it turns out that humans behaved more like hamsters than guinea pigs.  Yuschchenko returned to win his election, and served five years as Ukraine’s president.

This absurd variability in toxicity is shockingly normal.  An analysis of possible carcinogens that had been tested in both rats and mice showed that chemicals shown to cause cancer in rats had about a 57% chance of causing cancer in mice.  This is only slightly better than a coin flip.

By contrast, a 2006 survey showed that, for 6 known human carcinogens that had been tested on both rats and mice, only 3 were toxic to both.  We may share 90% of our DNA with rats and mice, but toxicity doesn’t correlate anywhere near that well.

Should we be surprised by this?

A significant portion of the DNA that differs between animals arises from differences in diets, including enzymes that aid in the digestion and detoxification of food.  One simple way to quantify this difference is to measure “bioavailability”, the fraction of a molecule that we ingest that actually reaches our bloodstream.  As you might expect for a diet-related measurement, what a dog, rat, or monkey extracts from what they eat shows very little correlation to humans.


A plot of drug bioavailability in animals vs humans.  Each dot represents a separate drug trial in either primates (blue), rodents (green), or dogs (red).  See if you can find a correlation.  Graph from Figure 1 of Are animal models predictive for humans?, doi: 10.1186/1747-5341-4-2.

No wonder toxicity data don’t align.

Hopes to find links between animal and human data get worse from there.  The entire evolutionary history of eating has been a story about how the things-that-get-eaten develop toxins to protect themselves, which then selects the-things-that-eat for resistance. So birds develop enzymes in their livers to digest small berries that are toxic to other species.  Humans develop enzymes to digest broccoli compounds that would kill birds.  The circle of life continues.

These differences in bioavailability and metabolism also impact molecules that were never intended to be significant parts of our diets.  Chocolate contains the molecule theobromine, which is toxic to both dogs and humans.  But humans can eat small amounts of it because we are less prone to its stimulatory effects on the heart than dogs (by about 4X), and because we metabolize it more than twice as quickly.  An order of magnitude drop in sensitivity is the difference between toxic and delicious.

Or take the molecule PFOA, whose toxicity is being heavily studied because of releases by a DuPont manufacturing plant in West Virginia.  PFOA was found decades ago to be toxic to rats, and the molecule has since contaminated water sources all over the globe.  But we have learned nothing about human toxicity from studying rats – rat mortality occurred through interaction of PFOA with a rat liver enzyme that humans simply don’t have.

Will PFOA turn out to be toxic to humans in the amounts we are exposed to?  Without human testing, we no idea.  Will some other animal have a sensitivity to PFOA far higher than rats?  Almost certainly yes.

Trying to understand the human toxicity of a substance by looking at rats or other animals turns out to be a horrible idea.

So what do we do?

Science and technology offer few answers.  Every pharmaceutical compound is tested on animals for efficacy and toxicity before being tried on humans. Yet still less than 10% of molecules deemed safe and effective on animals end up getting approved in humans.  Even when backed by billions of dollars in funding, animal models offer little insight into what will happen to people.

The hardest part is that – outside drug development – there are few incentives to consider toxicity in the context of how biology actually works.  Polluters such as DuPont will lean towards interpreting the data based on the best case animal tests.  Environmentalists promote their cause by highlighting the worst.  And even academics will favor publicizing their most eye-popping animal results, because who wants to discuss a paper talking about a molecule that maybe has a 57% chance of being toxic to humans?

Reality sometimes presents messy, confusing narratives, where no one can tell the good guys from the bad guys, and no one clearly wins.

Well, except maybe Viktor Yushchenko.  It’s now 2016, and he’s still alive, thanks to the failure of his assassins to appreciate the vast variety of biology.


Viktor Yushchenko in 2016, still plenty alive.  Photo from Tasnim News Agency via Wikipedia.

Leave a Reply