The Mutagenicity of Pesticides
Concepts and Evaluation

Samuel S. Epstein and Marvin S. Legator

with a foreword by Joshua Lederberg
The MIT Press
Cambridge, Massachusetts, an

d London, England

 
Copyright © 1 971 by
The Massachusetts Institute of Technology

This book was designed by

the MIT Press Design Department.

It was set in Univers 685/689

Printed on 60 Ib. P&S Offset

by Halliday Lithograph Corp.

and bound by The Colonial Press Inc.
in the United States of America

All rights reserved. No part of this book may be reproduced in any
form or by any means, electronic or mechanical, including photo-
copying, recording, or by any information storage and retrieval
system, without permission in writing from the publisher.

ISBN (hard cover) 0 262 05008 O

Library of Congress catalog card number: 70-130274
Foreword

x

FOREWORD

Socially responsible scientists should be concerned
about the potential hazards of chemically induced
mutation for atleast three reasons. The most important
is also the most remote in the scale of time: the

human nature that defines our posterity is energized
by our cultural tradition; butitis also bounded by the
integrity of the genetic information of which each
generation is the vessel.

Second, genetic impairments already account fora
very large part of our existing burden of disease and
premature death. If we give proper weight to the
genetic component of many common diseases which
have amore complex etiology than the textbook
examples of Mendelian defects, we can calculate that
at least 25 percent of our health burden Is of genetic
origin. This figure is a very conservative estimate in
view of the genetic component of such griefs as
schizophrenia, diabetes, atherosclerosis, mental
retardation, early senility, and many congenital
malformations. In fact, the genetic factor in disease Is
bound to increase to an even larger proportion, for as
we deal with infectious disease and other environ-
mental insults, the genetic legacy of the species will
compete only with traumatic accidents as the major
factor in health.
 

Finally, experimental evidence of mutagenic
capability should be a danger signal that a compound
may also be capable of other, somatic hazards through
its action on DNA and other cellular constituents. For
example. the demonstration of chromosome break-
age in cultured cells exposed to cyclohexylamine was
the (administratively neglected) forerunner of the
eventual inculpation of the parent compound,
cyclamate, as a cancer hazard.°

Our existing genetic load is a summation of three
kinds of process: the historical accumulation of
recessive gene mutations reappearing from time to
time as their heterozygous carriers chance to mate;
the immediate manifestation of dominant mutations
and chromosome anomalies, which are only rarely
propagated; and the paradoxical segregational load,
where deleterious recessives had been stabilized
within the population through some present, or more
often historical, advantage of the heterozygotes.

Any assessment of the social and personal costs of
mutation must take account both of absolute and of
relative measures. (And of course we must use the
same perspective in weighing the social and personal
benefits claimed for a given environmental additive. )
A 10 percent increase in the existing, “spontaneous”

mutation rate is, in effect, the standard that has been
adopted as the ‘maximum acceptable” level of

public exposure to radiation by responsible regulatory
bodies.* This level can be defended on the argument
that we neglect to take a number of measures that
could probably improve the mutation index to acom-
parable degree. It can be attacked by reciting the
absolute level of eventual biological injury that might
come from public exposure at such a level, were this
in fact to occur from the proliferation of nuclear power
plants and unregulated weapons tests.

A rational approach to the assessment of chemical
hazards likewise calls for a detailed quantitative
examination of risk, if only to expose the policy
assumptions to the same degree of public under-
standing and debate as pertains to atomic energy. In
both sets of situations, we are plagued by serious
uncertainties about the numerical estimates for
induced mutation in man, perhaps much worse for
any chemical than for radiation. Given the problem of
the number of new, suspicious compounds now
pervading the environment, we face a formidable task
in putting our genetic house in order.

For many purposes, we could dispose with a
quantitative analysis, simply because most of our
xi FOREWORD
assays are So insensitive that a compound that scores
positively as a mutagen must have a portentous effect
in any but the most peculiar circumstances. Never-
theless, the labeling of compounds as “mutagenic” or
“nonmutagenic’”’ may, particularly if we pursue the
development of assays for sensitivity rather than
selectivity, be regarded as simplistic, both in quanti-
tative and in qualitative terms. However, when there
are murky doubts about an issue as important as
mutational damage, they must be resolved in favor of
the species.

The sobriety with which we face the task of setting
up rational criteria for decision must be robust
enough to withstand some inevitable ridicule. The
very first compound to be reported! in the published
literature as mutagenic was a//y/ isothiocyanate, or
mustard oil, well known as a natural constituent of
horseradish and other widely used condiments. (This
should not be confused with mustard gas, research
on which was classified during World War II, and
which was later revealed to be a mutagen of un-
common potency.) Whether allyl isothiocyanate is
mutagenic in mammals is not known, and it is too
soon to condemn a foodstuff having such a venerable
tradition. This is a concession to that tradition, rather
xii FOREWORD

than to any factual assurance that mustard- eaters
have a mutation rate demonstrably unaltered by their
diet. Further research may allay or confirm these
suspicions; if they are confirmed, it would not be the
firsttime that a common dietary article was found to
be harmful to some consumers despite centuries of
common use (cf. the role of wheat gluten, only
recently understood in the etiology of celiac disease.)

This kind of diffidence in the face of custom and
uncertainty does notextend to efforts to promulgate a
compound like mustard oil for uses lacking an
ancient precedent. It has, for example, been advocated
as a way of denaturing hobby glues to deter sniffing,
an admirable social purpose if it reached the roots of
the problem. The FDA's approval of mustard oil as an
obviously safe additive is the entire argument against
unwonted side effects. This must be balanced against
along-buried observation of skin tumors in mice
painted with mustard oil)? which is also another
example of the predictive value of mutagenicity for
carcinogenicity. Theoretically, isothiocyanates can be
expected to function as alkylating agents.

Mustard oil is not a pesticide (except, perhaps, as a
natural one evolved by cruciferous plants); why
mention it here ? Mainly because, little as we know
about it, more is known about mustard oil than about
most pesticides. Furthermore, most of our concerns
about pesticides are prototypes of our general con-
cerns about environmental additives. We should
admit, at the outset, that pesticides must be differ-
entiated from food additives in the dilemmas that they
raise about the cost-benefit equation. (Some critics
believe, however, that pesticides are overpromoted
to the point where economic benefits would be left
intact or even improved if they were used more
selectively; and the environmental load could then be
cut by a factor of 10 or 100.)

It will encourage a fuller exposure of the magnitude
and incidence of economic benefits of potentially
hazardous chemicals if we do suggest some quan-
titative standards of acceptable mutagenicity. |
believe that the present standards for population
exposure to radiation should and will (atleast de
facto) be made more stringent, to about 1 percent of
the spontaneous rate, and that this is a reasonable
standard for the maximum tolerable mutagenic effect
of any environmental chemical (better, for them in
aggregate).

Accepting, for present argument, the formal argu-
ments of the UN advisory group,® | translate this

standard into arate of about one recessive mutation
perl ,000 gametes (1077 per nominal locus) per
generation of typical exposure. Dominant mutations
and chromosome aberrations may deserve even more
stringent scrutiny, in view of the immediacy of their
personal and social cost. The corresponding standard
of 50 per million induced, viable chromosome
anomalies and 2 per million dominant mutations
entails araw social cost of over $100 million. Itis
probably at the margin of ultimate detectability even
by extrapolation from experimental assays: so that
atleast a billion-dollar argument needs to be put up
by the defendant of any additive that gives a positive
experimental score in such tests.

The costs of recessive mutations are much more
difficult to estimate, being quite sensitive to the pro-
portion of the mutationa/to the segregationa!/ load. At
equilibrium, a 1 percent increase in the mutation rate
will generate an estimated economic loss of about $1
billion per year (measured in the 1970 economy of
the United States) but taking at least ten generations
to approach full impact.

An extremely conservative estimate would then put
the near-term annual tax connected to this standard
at about $100 million for the recessives. These
xiii FOREWORD
calculations give no weight to such costs (or savings)
as may attend the gradual deterioration of intelligence
and other complex functions as a consequence of
cumulative genetic damage. Nor do they puta value
on heartache.

These estimates are surely subject to an uncertainty
of a factor of tenorso. They predicate the value of a
human life as between $50,000 and $1 million per
capita, depending on the age at which disability or
death occurs and the level of custodial care entailed
by it, as well as loss of economic productivity. They
assign no value to early prenatal losses, though some
would regard these as beneficial for the aim of zero
population growth. This kind of cost-accounting is
morally insufferable, but we must find some de facto
standard of value in making hard decisions. If lives are
valued at much more than a ‘million per bod,” there is
little evidence of this from the pragmatic behavior of
the community or of most individuals in the choices
they make in their daily lives.? However, these
choices are made in a hindered market where the cost
of safety is aside issue, more often obscured than
intelligently ventilated.

On the other hand, a pesticide manufacturer would
fire his director of public relations if he were to
xtv FOREWORD

advertise that he calculated a life atless than his own
annual income.

Ahealth cost (from the ‘acceptable’ standard) of
$200 million per year is a grievous burden in absolute
terms, but is immediately lost in an overall budget of
over $100 billion. (Of this, $60 billion is direct health
care: the indirect economic costs of disease, injury,
and premature senescence are open-ended.) This is
to say that a level of risk that approaches the intoler-
able, once we are well aware of it, may be impossible
to verify by direct measurements of disease diffused
throughout the population! In exceptional circum-
stances, an effect like the peculiar malformations
induced by thalidomide comes to the surface, and
then achieves a visibility and notoriety all out of pro-
portion to other agents. /f the ma/formation induced
by thalidomide were a mental retardation of 70
percent of the /.Q., instead of a highly characteristic
and unusual deformation of the limbs, in an equal
number of subjects, we would be unaware of it to
this day.

All this is to say that we must look to extrapolations
from laboratory measures for the on/y reliable indica-
tion of mutagenicity in the human population !

Someday, it will be argued that the standard risk
 

should be elevated in a particular case, for example,
were there to be a demonstrable net social benefit of,
say, $1 billion per year from the use of an agent that
elevated the mutation rate by 5 percent. The argument
should not be rejected out of hand. Forexample, |
believe that an acceleration of health research by
$1 billion per year would improve the genetic and the
overall health climate so as to more than outweigh
the penalties of more mutations. If there were a
harmonious redistribution of the resource benefit, we
could foresee an advantageous tradeoff. The problem
is to produce that harmony, to ensure that the people
who bear the risk and eventually pay the price will
also reap the benefits. Perhaps we will inventa taxon
pesticides, earmarked for compensatory research. This
makes sense only If we have exhausted alternative
sources of income for such restorative purposes.
Should pesticides be particularly suspect as muta-
gens ? Or would their pesticida! activity bear only an
accidental relationship to genetic damage? One
should answer this in relation to how much we know
about the mechanism of action of the compound, and
particularly the basis of its specificity. The fundamental
cellular processes of all organisms are remarkably
similar, the more so the closer we go to the genetic

foundations. The DNA of the bacterium, the insect,
the weed plant, the rodent pest, and of man has
precisely the same architecture. For this reason, the
most suspicious agents should be the disinfectants,
the compounds that act directly on cells, and indeed
most of these probably do acton the DNAof the
microbe as the target. Other pesticides may also act
on DNA, but owe their specificity to details of pene-
tration, or of metabolism in the pest which will deliver
the final toxic molecule to the cellular target. Finally,
most pesticides probably act as enzyme inhibitors,
but may have mutagenic effects as (1) incidental
side effects of their own structure: (2) further
metabolism in the mammalian system to genetically
active products, or (3) as side effects of the inhibition
of cellular enzymes. At the very least, pesticidal action
does not disqualify a compound as a mutagen, and

in some cases this may be very closely related to its
intended mode of action. Only empirical studies, of
the kind outlined in this monograph, can give a
conclusive answer.

These approaches are ingenious and already well
enough calibrated to be a sensible basis of regulatory
policy—at minimum the routine screening of every
proposed new compound through several tests.

XV FOREWORD
 

However, this is only a start.

| do not believe that any routine arbitrary procedure
will cover all of the potential hazards involved in
extrapolating from laboratory data to man. Many new
discoveries, some of fundamental importance to
biological theory, will be made from unanticipated
findings of harmful biological effects—just as the
discovery of genetic damage by x-rays was a flash of
creative genius when first surmised by H. J. Muller.
In most cases it will be metabolic products of a pesti-
cide that will cause trouble, not the initial compound,
and it is imperative that we have a clear picture of their
biochemical pathways in man. With such data, an
investigator might. for example, realize that some bio-
chemical deviants in the human population will be
uniquely sensitive; or that the pesticide will interact
with some drug or other environmental additive: or
that some stage in fetal life may be uniquely sensitive.
In any case, the system of evaluation must display all
that is known about a new compound so as to assure
the most creative thinking by pluralistic critics. This
recommendation runs counter to existing policy
which regards safety-evaluation data ona pesticide |
as the private property of the sponsor. a

This purpose is probably unachievable, despite the
xvi FOREWORD

most socially minded intentions, in the context of
sponsor-directed research. | would advocate, instead,
that the responsibility for safety-testing bein the
hands of disinterested third parties, funded by fees
taxed to the sponsors. This would also allow a fair
allocation of costs to the coattail riders who will jump
into competition now only after a pioneer has paid
the initial costs of an evaluation out of his own
pocket. It would also afford some leeway in selecting
which tests are most likely to elicit damning news, a
discretion that obviously cannot be vested in
interested sponsors.

When chemical mutagenesis was a matter of
speculation 30 years ago, many geneticists believed
that its realization was indispensable to understanding
the chemistry of the genetic material. History has
revealed the opposite—other biophysical studies
have given us the main clues about the structure of
DNA, and we are just beginning to understand the
complexity of the cellular processes that result in
mutation in the light of that knowledge. We still do
not know the local chemical change involved in
radiation damage to DNA, while mutagenesis by
base-substitution has been analyzed in so much
detail that itis the textbook example of molecular

 
 

pharmacology.® But for many other chemical muta-
gens there is much more to be learned, perhaps
mostly in the metabolism and repair or misrepair of the
DNA once it has suffered its primary damage.

This kind of insight already has great promise for
more powerful assays of mutagenic potentiality in
human cells, for example by direct measurement of the
extent of DNA repair in cells subject to experimental
or environmental insult, along the lines of Cleaver’s
studies on repair-deficient genotypes of man.?

Geneticists must not now overlook the other side of
the coin—the enormous value of reliable measures to
decrease the spontaneous mutation rate. There has
been very little followup of the pioneering work on
anti-mutagenic chemicals by Novick and Szilard two
decades ago.®

Joshua Lederberg

References
1. Auerbach, Charlotte and Robson, J. M.

1944. Production of mutations by ally! isothiocyanate.

Nature 154: 81.

2. Cleaver, J. E.
1969. Xeroderma Pigmentosum: A Human Disease in

Which an Initial Stage of DNA Repair is Defective.
Proc. N.A.S., 63:428-435.

3. Some of the history of this incident is recounted by
R. Egeberg, et al. 1970, Report of the Medical Ad-
visory Group on Cyclamates, UA.VM.A., 277:
1358-1361. See also Mintz, Morton, Rise and Fall of
Cyclamates, Washington Post, Oct. 26,1969 (re-
printed in the Congressional Record, Oct. 27, 1969,
S-13256-8).

4. Federal Radiation Council.

May 13,1960. Radiation Protection Standards Report
No. 1. See also Federal Register, 4402, May 18, 1960.
5. Goldstein, Avram, Aronow, Lewis, and Kalman, S.M.
1968. Principles of Drug Action. Harper and Row, N.Y.

6. Novick, A., and Szilard, L.
1952. Antimutagens. Nature 7170: 962-927.

7. Rusch, H. P.. Bosch, Dorothy, and Boutwell, R. K.
1955. The influence of irritants on mitotic activity and

tumor formation in mouse epidermis. Acta Un. Int.
Canc., 77: 699-703.

8. Scientific Committee on the Effects of Atomic
Radiation.

1966. Report to U.N. General Assembly, Official
Records, 21st session. Suppl. 14 (A/6314).

9. Starr, C.

1969. Social benefit versus technological risk.
Science 165: 1232-1238.

xvii FOREWORD