VOL. XXIX, No.
338
I N D U C T I O N OF MUTATIONS IN MICE BY CHRONIC
GAMMA I R R A D I A T I O N ; INTERIM REPORT*
By T. C. CARTER, MARY F. LYON and
Medical Research Council Radiobiological Research Unit, Atomic Energy Research Establishment,
Harwell, England
(Accepted for publication May, 1955)
THEmedicine
use of nuclear energy in warfare, industry,
and research, involves the exposure of
many human beings to greater or lesser doses of
ionizing radiations. This undoubtedly affects both
their bodies and their germ cells. A good deal is now
known about the magnitude of the somatic hazard,
but almost nothing quantitative about the genetic
hazard.
This fact is widely recognised among geneticists,
who would doubtless, if they were concerned solely
with the scientific aspects of the problem, prefer to
remain silent on a subject about which they are so
ignorant. However, there has been an insistent
demand for quantitative pronouncements, with the
result that a number of calculations have been
published in which attempts have been made to
evaluate the genetic hazard to man. Their value is
doubtful, but they have demonstrated a surprising
measure of agreement among geneticists in three
respects: first, admission of ignorance; second,
recognition that almost any answer can be obtained
from the calculations, according to the assumptions
made and the numerical values fed in; and third,
pessimism. Most geneticists would agree that it is
very undesirable that members of a human popu-
lation should be exposed to more than 25 r until
after the breeding age, though it is doubtful if any
could give really sound reasons for this belief.
From all this it becomes apparent that there is an
urgent need for more factual information and less
theoretical speculation. Fundamentally there are
two questions to be answered: (i) How much genetic
change is induced in man by a given amount of
radiation? (ii) What would be the effect on a human
population of releasing this genetic change into it?
Since this is a human problem, an answer to the
first question cannot be sought by means of direct
experimentation. The relationship between radiation
dosage and genetic damage must, therefore, be
investigated by observing the results of accidental
human exposure, and viewing them against a back-
ground of experimental knowledge obtained from
other organisms, especially mammals. Perhaps the
only occasions on which fairly large numbers of
* Paper read at the British Institute of Radiology Annual
Congress, November 25, 1954.
106
RITA J. S. PHILLIPS
human beings have been exposed to large doses of
ionizing radiations were at Hiroshima and Nagasaki;
and for this reason alone it is most desirable that
there should be an intensive study not only of the
individuals exposed (including those exposed as
embryos), but also of their children conceived after
exposure and of their later descendants. The experi-
mental background is extensive and covers many
plant and animal species. Work on mammals has
been almost entirely on mice, and has been reported
by Charles (1950), Russell (1951), Carter and
Phillips (1952) and Deringer, Heston and Lorenz
(1954). Animal experiments cannot, of course,
provide quantitative answers directly applicable to
human problems; but they can be, and have been,
used to investigate subsidiary questions, e.g. whether
there be any threshold dose-rate for mutagenesis
and whether the linear law relating mutation rate
to accumulated dose holds good at very low doses.
The second main question, that of the effect on a
human population of releasing induced genetic
changes into it, can likewise be answered only when
the answers are known to many subsidiary questions.
What is the level of mutant genes already present
in the population, due to spontaneous mutation in
the past? What selective disadvantage (or advantage)
does a mutant gene confer on a human being,
considered as a potential parent of the next genera-
tion? What disadvantage (or advantage) does it confer
on him when considered as a member of society?
What is the breeding structure of normal human
populations? These and many other relevant ques-
tions can only be answered by a study of man con-
sidered as ecological and genetical material; and
there is undoubtedly room for far more work in these
fields. However, here also a background of experi-
mental population studies is needed, both to bring
to light the unsuspected problems and to stimulate
the research effort; but it can only supplement, and
not supplant, the study of human populations.
The present paper is concerned with one aspect of
experimental radiation mutagenesis in a laboratory
mammal, namely whether the mutation-rate radia-
tion-dose curve is linear for very low accumulated
doses (Fig. 1). In the classical radiation work with
Drosophila the linear law was established for doses

Induction of Mutations in Mice by Chronic Gamma Irradiation: Interim Report
of the order of thousands of rontgens. This was
extended down to acute doses of 25 r by Spencer and
Stern (1948) but Caspari and Stern (1948) and
Uphoff and Stern (1950) obtained results suggesting
that for at least one type of genetic change the
results of a small chronic dose of radiation (about
50 r) can be obscured by other effects. Bonnier and
Liining (1949) on the other hand, found an apparent
departure from linearity in the opposite direction,
and there is a similar trend in Charles' (1950) mouse
data. Recent work by Marcovich (1954), however,
appears to indicate linearity in phage for doses down
to 0-5 r. This question is an important one from the
viewpoint of genetic hazard to man, since man is
MUTATION
RATE
2S
RADIATION
DOSE
FIG. 1.
The effect of non-linearity of the mutation-rate/radiation-
dose law for very low doses. If there is non-linearity of the
type shown, the radiation dose (Dx) required to double the
spontaneous mutation rate (Sx) is much smaller than the
dose (D2) required on the basis of a linear law.
likely to receive only relatively small doses of
radiation through the industrial or research use of
atomic energy and it is therefore important to know
the genetic damage at these low doses. Moreover, it
would probably be unwise to argue from any
organism very dissimilar to man. The Medical
Research Council therefore decided that this
question should be investigated, using mice as the
experimental material. A preliminary report was
given four years ago (Carter and Phillips, 1952):
the present paper is an interim report on subsequent
work.
107
FEBRUARY 1956
Methods and materials
When this work was started, in 1947, at the
Institute of Animal Genetics in Edinburgh, nothing
was known of a similar but considerably larger
experiment then being started for the U.S. Atomic
Energy Commission by W. L. Russell at Oak Ridge.
By a curious process of convergent evolution we
independently came to use almost identical mouse
stocks and procedures for detecting mutations; and
in 1953, by which time each knew about the other's
work, it was decided to bring the British experiment
completely into line with the American one. The
stocks which we had built up were therefore replaced
by Russell's stocks.
FIG. 2.
The multiple-recessive method of mutation detection. A
wild-type male mouse (top left) produces normal sperm
(left) which fertilises an ovum from a female carrying two
doses of a recessive mutant gene (top right); the zygote
(bottom left) therefore carries only one dose and so appears
wild-type. If the sperm also carries the mutant gene, as a
result of mutation, the zygote (bottom right) carries it in
two doses and therefore has the mutant phenotype.
However, though identical mouse stocks and
breeding procedures are now used, the objects of
the experiments are different. Russell's object is to
discover the rate of induction of mutations per
rontgen at seven specific loci; he therefore uses a
high acute dose of X rays, usually 600 r (Russell,
1951). Our object is to test for non-linearity of the
mutation-rate/radiation-dose law at a low accumu-
lated chronic irradiation dose; we therefore give
about 40 r of y radiation, accumulated over one or
VOL. XXIX, No. 338
T. C. Carter, Mary F. Lyon and Rita J. S. Phillips
five weeks in 16 hours per night and five nights per
week. Our measured value of the 40 r mutation rate
is then compared with the value expected on the
basis of a linear law, found by interpolation on
Russell's curve. Our control rate gives, by com-
parison with his, a check that there is no systematic
difference between Oak Ridge and Edinburgh in any
other factor affecting mutation rate.
Our irradiation set-up has been described by
Carter, Lyon and Phillips (1954). Mutations are
detected by the "multiple-recessive" method (Fig. 2):
wild-type male mice, irradiated or control, are mated
with non-irradiated females homozygous for the
seven recessive mutants nonagouti (a), brown (b),
chinchilla {(f11), blue dilution (d), pink-eyed (p), piebald
(s) and short-eared (se). All the Fl mice are wild type,
except those derived from a sperm which had
TABLE I
NUMBER OF MUTATIONS OBTAINED AT EACH OF THE SPECIFIED LOCI
Number of
Dose animals
(r) Series* examined a b
40 PQ 4,986
40 R 5,038
— —
0 PQ 10,795 — —
0 R 7,560
* PQ, using original Edinburgh stocks; R, using offshoots of Russell's stocks.
mutated to a recessive allele at one of the seven loci.
The mutant phenotypes are all easily recognisable,
so the test has very little subjective element.
Russell's (1951) control mutation rate over the
seven loci is about five per 105 Fl mice. His 600 r
mutation rate is about 112 per 105. On the basis of a
linear law, therefore, we must expect about 12
mutations per 105 Fl mice after exposure to 40 r. It
follows that extremely large numbers of mice would
have to be raised in order to distinguish with
statistical significance between the effects of chance
distribution and real non-linearity. However, in
ad hoc research of this kind the object is not so much
to detect the presence of non-linearity as to confirm
the absence of serious non-linearity. We hope it will
not be necessary to raise more than a quarter of a
million mice from irradiated sperm and half a
million controls. Clearly this will need extensive
animal breeding facilities and take a number of years;
for this reason the work is now being transferred to
Harwell.
Results
At Edinburgh the experiment has been on a pilot
1
108
scale; only 28,000 Fl mice have been raised and only
three mutations found among them (Table I). Few
conclusions can be drawn from such scanty data.
Our control mutation rate is clearly of the same
order as Russell's; and our 40 r mutation rate is
compatible with that expected on a linear law. On
the other hand, the present data are also still com-
patible with a much higher rate. Many more data
must be awaited before detailed conclusions can be
drawn about mutation rates. However, it can be
safely said that in our work, as in Russell's, the
s-locus appears to be more mutable than the others;
of our three mutations, two were at this locus.
The nature of the mutations obtained is now under
examination. Genetic tests have been completed on
one of them and this, an allele at the s-locus, proved
to be viable when homozygous.
Number of mutations at locus
c d P s se Total
1 — \
1
-i
2
1 1 J
ACKNOWLEDGMENTS
We are indebted to Miss M. M. Manson and Miss G. I. E.
Mavor for extensive technical assistance.
SUMMARY
Estimation of the genetic hazard of ionizing radiations
to man calls for ecological and genetic study of human
populations and experimental work on other mammals. An
interim report is given of a long-term experiment intended
to check the linearity of the mutation-rate/radiation-dose
law at seven specified loci in the mouse, using low accumu-
lated y-ray doses.
REFERENCES
BONNIER, G., and LUNING, K. G., Hereditas, 1949, xxxv,
163.
CARTER, T. C , LYON, M. F., and PHILLIPS, R. J. S.,
Brit.J. Radiol, 1954, xxxvii, 418.
CARTER, T. C., and PHILLIPS, R. J. S., Biological Hazards
of Atomic Energy, 1952, 73 (Ed. A. Haddow; Oxford,
Clarendon Press).
CASPARI, E. W., and STERN, C , Genetics, 1948, xxxiii, 75.
CHARLES, D. R., Radiology, 1950, lv, 579.
DERINGER, M. K., HESTON, W. E., and LORENZ, E.,
Biological Effects of External X and Gamma Radiation, 1954,
149 (Ed. R. E. Zirkle; New York, McGraw-Hill Co.)
MARCOVICH, H., Nature (London), 1954, clxxiv, 796.
RUSSELL, W. L., Cold Spring Harbor Symposia on
Quantitative Biology, 1951, xvi, 327.
SPENCER, W. P., and STERN, C., Genetics, 1948, xxxiii, 43.
UPHOFF, D. E., and STERN, C , Science, 1950 clx, 609.