My present lecture will consist of four parts.
Firstly, I want to emphasise the importance of toxicity testing
in our modern society.
Secondly, I am going to criticise animal toxicity tests. The
shortcomings of these tests will be scrutinised from a scientific
standpoint. The LD50 test – that is, a measure of the
dose of a chemical which kills 50 per cent of a population
of rodents – will be taken as an example of a relatively
ineffective and cruel toxicity test. At the same time, I shall
point out that sole reliance on such a test for risk assessment
is decreasing, and in many areas of toxicity testing this
test is in fact already obsolete.
Thirdly, a more refined animal-based approach to toxicity
testing will be described which is today typical in most sectors
of the pharmaceutical and chemical industry.
Fourthly, the basis for non-animal toxicity testing in the
near future is described. This is cell toxicology, with its
possibilities of using batteries of cell culture and other
test-tube tests as models of human reaction to the toxicity
of chemicals.
I shall start from the beginning and remind you that we do
not live in the stone, bronze or iron ages any longer –
we now live in the chemical age. At the same time we live
in the plastics age, the communications age, the information-explosion
age and so on. The chemical age denotes the fact that man
is surrounded by four million chemicals, which penetrate every
segment of his surroundings. Most of these chemicals have
only recently been synthesised, and new ones are pouring out
of industry every day, to the extent of hundreds per month.
And most of them are not released in distinct and easily controlled
doses, as are (or should be) drugs and pesticides. No, the
vast majority of them occur unexpectedly and diffusely, in
food, clothing, furniture, cars, electronic items et cetera
–that is when they are not environmental pollutants,
scattered abroad without any control whatsoever.
The nastiest fact, however, is that most of these chemicals
have been poorly tested for the various effects of their toxicity
on man, such as acute toxicity, chronic toxicity, local irritancy
to the eyes or skin, foetal toxicity, allergic sensitisation
potential and so forth. And many of the chemicals have not
been tested at all, in spite of the production and dispersion
of hundreds of tons of them. This is the so-called “mountain”
of untested or poorly tested chemicals feared by many government
agencies today.
Many laymen and, indeed, many professional toxicologists believe
that this state of affairs is somehow checked by the medical
profession: if some of untested chemicals really are dangerous,
physicians will surely detect this in a typical or unusual
illness. I think that all the physicians in this auditorium
would agree that this “clinical” safety-net is
an illusion. A human being can only respond to injury to cells
and organs in a relatively stereotyped way and limited to
few illnesses, such as cancers, various degrees of inflammation
and so on, connected with each organ. Nearly always, then,
only components of the so-called general panorama of illnesses
will result from chemical toxicity (for example, heart, liver
and kidney problems, headache, cough, vertigo, sleepiness)
– that is, diagnoses are made without obvious reason.
It is very seldom that a new syndrome, specific to the offending
agent only, will result from toxicity. Consequently, there
is no clinical safety-net for untested chemicals as long as
a percentage of the population is ill from unknown causes.
Systematic toxicity testing of chemicals, including human
risk assessment based on the tests, is only about 50 years
old. Up to now, most of this testing has been done by using
animal subjects. At least 25 per cent of all animals used
in biomedical and other research are destined for toxicity
tests. Groups of five to ten animals receive a dose of the
chemical by one method of administration or another, making
up several groups, plus control animals, per experiment. Symptoms
of injury, illness and/or death are then recorded after certain
intervals of time. Often an autopsy is performed, including
histopathological studies. The LD50 test, taken as an example
here, often has an observation period of two weeks. The LD50
test is the first to be performed on an unknown chemical,
followed by a series of other tests on local irritancy, subchronic
and chronic toxicity, teratogenicity, phototoxicity, sensitisation,
fertility and short- or long-term carcinogenicity. The most
common animal toxicity test is the LD50 assay or small modifications
of that test.
A basic animal test, such as the LD50 test, has several shortcomings:
Firstly, the idea behind the test – that is, the use
of an animal as a representative of the human body and mind
for testing purposes – is not correct. We have a species
gap between the various animals due to differences in bodily
functions, making their responses to chemicals differ from
one to another. The toxic dose depends on the reaction of
the most sensitive target in the body. The target is always
the molecules constituting parts of cells, intercellular materials,
extra-cellular transmitters or hormones. But often the effect
will also depend on various toxico-kinetic factors, such as
absorption of the chemical in the intestine or skin, metabolism
of the chemical in the liver, distribution of the chemical
to bodily compartments, storage of the chemical and finally
excretion of it through kidneys, intestines, lungs or skin.
A variation in only one of these factors between the test
animal and the human will make the toxic dose administered
to the animal inapplicable to or unpredictable for man. Thus,
animal experiments are a gamble. Additional tests in other
animal species will not improve the odds – often, confusion
is increased, at the expense of further suffering and costs.
A second flaw in animal testing is that primary toxic events,
such as chemical interference with molecules in cells, cell
organelles and extracellular receptors, are not measured.
Instead, a host of secondary effects from the original insult
obscures the picture – that is, the symptoms of poisoning
such as fall in blood pressure, confusion, convulsions and
so on. Often these secondary effects, which have no quantitative
relation to the original insult, are the basis for the measurement
of toxicity. Animal experiments simply have a very low resolving
power in interpreting toxicity. Routine autopsy may compensate
for 5this to a degree but may also, in most cases, overlook
the functional effects of injury. The main problem with the
measurement of secondary phenomena is that they make test
results still more difficult to extrapolate to man.
Why have such blunt and crude investigative tools been used
in the past and are still being used? The obvious answer seems
to be that nobody has yet come up with better ideas. The real
answer is probably that until recent times nobody has bothered
to try to come up with better ideas. Twenty or thirty years
ago, when industry and regulatory agencies had not yet been
influenced by the modern consciousness of risks in this chemical
age, or by a deeper concern about opinions on animal experiments,
these scientifically crude tests were considered good enough.
However, for decades now, modern chemical and pharmaceutical
industrial and regulatory governmental authorities have not
relied solely on animal tests to measure toxic doses, body
counts and toxic symptoms. Nowadays, the routine toxicity
testing of pharmaceuticals, pesticides, industrial chemicals,
cosmetics et cetera is much more refined. As the first step
towards such a composite toxicology study of a compound, accumulated
historical data on the toxicity of chemical analogues are
used, together with simple physico-chemical data, on the investigated
compound in order to predict human toxicity. This is done
by structure-activity analysis with the help of computer programs.
Guided by such studies, qualified toxicity tests on relatively
few animals are performed. These include toxico-kinetic tests
(blood levels and metabolites of the compound) and tests of
subtle clinical toxicity (blood and urine analysis, including
enzymology and cell counts). A more refined system of histopathology
is applied and related to the other measurements to find out
the cause of primary toxic events. Another characteristic
feature of modern toxicology tests is the tiered-test approach,
in which subsequent tests are individually guided and motivated
by earlier tests. This is very different from the old-time
approach of performing the standard animal tests in parallel,
just to gain time. Instead, the modern approach gains understanding.
Often the primary toxic mechanism is appreciated, which enhances
the effectiveness of human risk prediction enormously. These
routines are indeed a step forward beyond the LD50 test, although
still based on animal testing.
One important problem with the modern, refined use of animals
in toxicological studies is that these studies are very expensive.
Thus, economical reasons are now a limiting factor in extending
such scientifically improved testing – that is, replacing
crude animal tests such as the LD50 or Draize tests in the
mass testing of food additives, industrial chemicals, household
products and so forth. The problem is similar to one in the
technical developments in medicine where economic reasons
prevent the unlimited use of transplantations, heart surgery
et cetera.
To sum up: refined composite animal toxicity experiments have
improved the test situation, but they have not solved the
important problems in that field. Thus, toxicology still relies
on unscientific animal experiments.
In the last ten years, a new method or subdiscipline has evolved
in toxicology, namely, cell toxicology. This method may also
be called test-tube toxicology. Cell toxicology is cell biology
and molecular biology applied to toxicity studies. One main
incentive in this development was the discovery and promotion
of bacterial mutagenicity and carcinogenicity tests by Dr
Bruce Ames in the early 70s. These mechanistically based test-tube
tests have for 15 years been used routinely as adjuncts to
animal carinogenicity tests. Additionally, cultures of specific
cells from diverse organs have been used for 20 years as adjuncts
to animal experiments in the study of toxic mechanisms. As
a third development, ten years ago Christian, Waters, Zucco,
Paganuzzi-Stammati, Walum, Nardone, Halle et al began to realise
that cell cultures could also be used in toxicity testing.
To understand the almost revolutionary effect of our ideas
on how to use cell cultures as replacements for animals in
toxicity testing (and the almost counter-revolutionary reactions
against these ideas), you must be acquainted with the thinking
of the toxicologist of some years ago. His discipline came
from pharmacology, which had been much influenced by the basic
science of physiology. Thus, most toxic effects of chemicals
– apart from the development of cancer and the toxicity
of anti-cancer drugs – were thought to be the results
of interactions between those chemicals and the regulatory
mechanisms of the body, especially the receptor-mediated regulatory
mechanisms. Furthermore, if toxicity in a few cases was attributed
to direct cell injury, the organ-specific functions of cells
were thought to be the target of this toxicity. Any target-organ
toxicity was thus thought to be interference with typical
cells (liver cells in liver injury, brain cells in brain injury
and so on). Hence, the use of cells in the testing of general
toxicity was, according to this thinking, impossible, because:
1) few toxic effects result in cell injury; and 2) if cellular
toxic effects were to be tested, a very large number of organ-specific
cells (liver, kidney, heart, brain, lung, thyroid, et cetera),
must be used in a battery to cover the toxic insults to man,
and this would be costly and unpractical. Mechanistically
based cell tests of general toxicity, analogous with the Ames
test, would also be impossible, since general toxicity is
known by the toxicologist to operate via many different mechanisms.
To summarise results from my own and other cell toxicologists’
research: it seems to be possible to construct batteries of
test-tube tests on various types of general toxicity (LD50,
Draize eye and skin tests and so forth). The core of such
batteries would be cell-line toxicity tests measuring basal
cytotoxicity. Other more costly, animal-dependent, primary
cultures of some key organ-specific cells may be added to
the batter, to cover common types of specific organ toxicity.
Test-tube tests with important macro-molecules in the body
such as proteins and phospholipids, as well as similar tests
with important transmitters, may also be added to the battery
to cover non-cellular toxicity. Such a battery would have
a good chance of predicting toxic blood and tissue levels
of unknown compounds. These results would probably at times
be falsely negative – when a rare toxic effect not covered
by the battery occurs – but would certainly not produce
false positive results. A new toxicology will emerge, performing
risk assessment based on comparisons of blood and tissue concentrations
from typical exposure in man, and resulting from the above-described
battery.
To be able also to predict toxic dosage for humans, and to
predict the possible toxicity of chemicals which are metabolised
by the liver into more toxic compounds, the battery must also
include toxicokinetic tests, and this is under way, although
the necessary in-vitro methods have not yet been fully developed.
Human liver cells and slices are already routinely used by
some pharmaceutical companies as an adjunct in predicting
the metabolism of drugs in man. Absorption tests (intestine,
skin et cetera) are under investigation. Distribution, storage
and excretion tests have not yet been developed.
To summarise: the concept of animal-free toxicity testing
has recently evolved, aspects of which are now being realised
in practical terms by cell toxicologists. Batteries of cellular
and other test-tube tests of toxicity and toxicokinetic events
are taking shape. These batteries will prove a great deal
more scientific and more effective than animal tests. The
reasons for this assertion are as follows. 1) Each component
of a battery is calibrated directly to human conditions for
better prediction. 2) the batteries focus on testing the primary
cellular events. 3) High test sensitivity is ensured by the
high number of cells involved and by the sensitive toxicity
criteria which it is possible to apply to cell cultures. 4)
Mechanistic understanding is very easy to achieve by further
studies on the cellular level.
Return to top
Back to list of speeches
|